├── .gitignore ├── LICENSE ├── README.md ├── acleto ├── __init__.py ├── al4nlp │ ├── __init__.py │ ├── active_learner.py │ ├── active_selector.py │ ├── constructors │ │ ├── __init__.py │ │ ├── construct_active_learner.py │ │ └── construct_wrapper.py │ ├── model_wrappers │ │ ├── __init__.py │ │ ├── flair │ │ │ ├── __init__.py │ │ │ ├── flair_trainer.py │ │ │ └── wrapper_flair.py │ │ ├── pytorch │ │ │ ├── __init__.py │ │ │ ├── wrapper_cls_pytorch.py │ │ │ ├── wrapper_encoder_pytorch.py │ │ │ └── wrapper_ner_pytorch.py │ │ └── transformers │ │ │ ├── __init__.py │ │ │ ├── trainer_for_pseudo_labeled.py │ │ │ ├── transformers_base_wrapper.py │ │ │ ├── wrapper_ats.py │ │ │ ├── wrapper_cls.py │ │ │ ├── wrapper_encoder.py │ │ │ ├── wrapper_ner.py │ │ │ ├── wrapper_nmt.py │ │ │ └── wrapper_seq2seq.py │ ├── models │ │ ├── __init__.py │ │ ├── bilstm_crf.py │ │ ├── fnet.py │ │ ├── ner_bilstm.py │ │ ├── spectral_normalization │ │ │ ├── __init__.py │ │ │ ├── spectral_norm_conv_inplace.py │ │ │ └── spectral_norm_fc.py │ │ └── text_classification_cnn.py │ ├── pool_subsampling_strategies │ │ ├── __init__.py │ │ ├── naive_subsampling.py │ │ ├── random_subsampling.py │ │ └── ups_subsampling.py │ ├── query_strategies │ │ ├── __init__.py │ │ ├── actune.py │ │ ├── akim.py │ │ ├── akim_batchbald.py │ │ ├── al_strategy.py │ │ ├── al_strategy_utils.py │ │ ├── alps.py │ │ ├── badge.py │ │ ├── bait.py │ │ ├── bald.py │ │ ├── batchbald.py │ │ ├── bleuvar.py │ │ ├── breaking_ties.py │ │ ├── cal.py │ │ ├── cluster_margin.py │ │ ├── coreset.py │ │ ├── egl.py │ │ ├── embeddings_km.py │ │ ├── ensp.py │ │ ├── entropy.py │ │ ├── lc.py │ │ ├── mahalanobis_sampling.py │ │ ├── mnlp.py │ │ ├── ngram_sampling.py │ │ ├── nsp.py │ │ ├── random_sampling.py │ │ ├── strategy_utils │ │ │ ├── __init__.py │ │ │ ├── batchbald │ │ │ │ ├── README.md │ │ │ │ ├── __init__.py │ │ │ │ ├── batchbald.py │ │ │ │ ├── compute_entropy.py │ │ │ │ ├── consistent_dropout.py │ │ │ │ ├── joint_entropy.py │ │ │ │ └── utils.py │ │ │ ├── gmm.py │ │ │ ├── mahalanobis.py │ │ │ ├── mahalanobis_v2.py │ │ │ └── ue │ │ │ │ ├── __init__.py │ │ │ │ ├── alpaca_calibrator.py │ │ │ │ ├── mahalanobis_distance.py │ │ │ │ ├── spectral_norm.py │ │ │ │ ├── text_classifier.py │ │ │ │ ├── ue_estimator_hybrid.py │ │ │ │ ├── utils_heads.py │ │ │ │ └── utils_inference.py │ │ └── strategy_wrappers │ │ │ ├── __init__.py │ │ │ ├── modal_sampling.py │ │ │ └── small_text_sampling.py │ └── utils │ │ ├── __init__.py │ │ ├── augmentation.py │ │ ├── cache_all_necessary_files.py │ │ ├── calculate_rouges.py │ │ ├── cluster_margin.py │ │ ├── cluster_utils.py │ │ ├── data │ │ ├── __init__.py │ │ ├── load_arbitrary_dataset.py │ │ ├── load_data.py │ │ ├── load_from_json_or_csv.py │ │ ├── load_from_url.py │ │ ├── load_huggingface_dataset.py │ │ └── preprocessing.py │ │ ├── embeddings.py │ │ ├── general.py │ │ ├── get_embeddings.py │ │ ├── get_gradient_lengths.py │ │ ├── get_train_constants.py │ │ ├── init_model_and_tokenizer.py │ │ ├── main_decorator.py │ │ ├── model_modifications.py │ │ ├── packages │ │ ├── __init__.py │ │ └── bart_score.py │ │ ├── restore_queries.py │ │ ├── seq2seq_metrics.py │ │ ├── summarization_metrics.py │ │ ├── time_decorator.py │ │ ├── token_classification.py │ │ ├── transformers_dataset.py │ │ └── utils_tasks.py ├── al_benchmark │ ├── __init__.py │ ├── configs │ │ ├── al_ats.yaml │ │ ├── al_ats_pegasus.yaml │ │ ├── al_ats_pl.yaml │ │ ├── al_ats_pl_pegasus.yaml │ │ ├── al_ats_prophetnet.yaml │ │ ├── al_cls.yaml │ │ ├── al_cls_asm.yaml │ │ ├── al_cls_asm_cnn.yaml │ │ ├── al_cls_cnn.yaml │ │ ├── al_cls_plasm.yaml │ │ ├── al_cls_plasm_cnn.yaml │ │ ├── al_cls_ru.yaml │ │ ├── al_ner.yaml │ │ ├── al_ner_asm.yaml │ │ ├── al_ner_bilstm.yaml │ │ ├── al_ner_bilstm_crf_flair.yaml │ │ ├── al_ner_da.yaml │ │ ├── al_ner_distilbert-bilstm_crf_flair.yaml │ │ ├── al_ner_plasm.yaml │ │ ├── al_ner_plasm_bilstm_flair.yaml │ │ ├── al_ner_ru.yaml │ │ ├── al_strategy │ │ │ ├── ddu.yaml │ │ │ ├── lc.yaml │ │ │ ├── logits_lc.yaml │ │ │ ├── mahalanobis.yaml │ │ │ ├── mnlp.yaml │ │ │ ├── nuq.yaml │ │ │ └── random.yaml │ │ ├── auto_ner.yaml │ │ ├── config.yaml │ │ ├── exps_al.yaml │ │ ├── exps_ens.yaml │ │ ├── flask_cfg.yaml │ │ ├── full_data_ats.yaml │ │ ├── full_data_ats_pegasus.yaml │ │ ├── full_data_cls.yaml │ │ ├── full_data_cls_cnn.yaml │ │ ├── full_data_ner.yaml │ │ ├── full_data_ner_bilstm.yaml │ │ ├── full_data_ner_bilstm_crf_flair.yaml │ │ └── ue │ │ │ ├── beam.yaml │ │ │ ├── ensemble.yaml │ │ │ ├── mc.yaml │ │ │ └── sampling.yaml │ ├── data │ │ ├── abssum_debate │ │ │ ├── dataset.json │ │ │ └── train.json │ │ ├── bbc_news │ │ │ ├── README.TXT │ │ │ ├── business │ │ │ │ ├── 001.txt │ │ │ │ ├── 002.txt │ │ │ │ ├── 003.txt │ │ │ │ ├── 004.txt │ │ │ │ ├── 005.txt │ │ │ │ ├── 006.txt │ │ │ │ ├── 007.txt │ │ │ │ ├── 008.txt │ │ │ │ ├── 009.txt │ │ │ │ ├── 010.txt │ │ │ │ ├── 011.txt │ │ │ │ ├── 012.txt │ │ │ │ ├── 013.txt │ │ │ │ ├── 014.txt │ │ │ │ ├── 015.txt │ │ │ │ ├── 016.txt │ │ │ │ ├── 017.txt │ │ │ │ ├── 018.txt │ │ │ │ ├── 019.txt │ │ │ │ ├── 020.txt │ │ │ │ ├── 021.txt │ │ │ │ ├── 022.txt │ │ │ │ ├── 023.txt │ │ │ │ ├── 024.txt │ │ │ │ ├── 025.txt │ │ │ │ ├── 026.txt │ │ │ │ ├── 027.txt │ │ │ │ ├── 028.txt │ │ │ │ ├── 029.txt │ │ │ │ ├── 030.txt │ │ │ │ ├── 031.txt │ │ │ │ ├── 032.txt │ │ │ │ ├── 033.txt │ │ │ │ ├── 034.txt │ │ │ │ ├── 035.txt │ │ │ │ ├── 036.txt │ │ │ │ ├── 037.txt │ │ │ │ ├── 038.txt │ │ │ │ ├── 039.txt │ │ │ │ ├── 040.txt │ │ │ │ ├── 041.txt │ │ │ │ ├── 042.txt │ │ │ │ ├── 043.txt │ │ │ │ ├── 044.txt │ │ │ │ ├── 045.txt │ │ │ │ ├── 046.txt │ │ │ │ ├── 047.txt │ │ │ │ ├── 048.txt │ │ │ │ ├── 049.txt │ │ │ │ ├── 050.txt │ │ │ │ ├── 051.txt │ │ │ │ ├── 052.txt │ │ │ │ ├── 053.txt │ │ │ │ ├── 054.txt │ │ │ │ ├── 055.txt │ │ │ │ ├── 056.txt │ │ │ │ ├── 057.txt │ │ │ │ ├── 058.txt │ │ │ │ ├── 059.txt │ │ │ │ ├── 060.txt │ │ │ │ ├── 061.txt │ │ │ │ ├── 062.txt │ │ │ │ ├── 063.txt │ │ │ │ ├── 064.txt │ │ │ │ ├── 065.txt │ │ │ │ ├── 066.txt │ │ │ │ ├── 067.txt │ │ │ │ ├── 068.txt │ │ │ │ ├── 069.txt │ │ │ │ ├── 070.txt │ │ │ │ ├── 071.txt │ │ │ │ ├── 072.txt │ │ │ │ ├── 073.txt │ │ │ │ ├── 074.txt │ │ │ │ ├── 075.txt │ │ │ │ ├── 076.txt │ │ │ │ ├── 077.txt │ │ │ │ ├── 078.txt │ │ │ │ ├── 079.txt │ │ │ │ ├── 080.txt │ │ │ │ ├── 081.txt │ │ │ │ ├── 082.txt │ │ │ │ ├── 083.txt │ │ │ │ ├── 084.txt │ │ │ │ ├── 085.txt │ │ │ │ ├── 086.txt │ │ │ │ ├── 087.txt │ │ │ │ ├── 088.txt │ │ │ │ ├── 089.txt │ │ │ │ ├── 090.txt │ │ │ │ ├── 091.txt │ │ │ │ ├── 092.txt │ │ │ │ ├── 093.txt │ │ │ │ ├── 094.txt │ │ │ │ ├── 095.txt │ │ │ │ ├── 096.txt │ │ │ │ ├── 097.txt │ │ │ │ ├── 098.txt │ │ │ │ ├── 099.txt │ │ │ │ ├── 100.txt │ │ │ │ ├── 101.txt │ │ │ │ ├── 102.txt │ │ │ │ ├── 103.txt │ │ │ │ ├── 104.txt │ │ │ │ ├── 105.txt │ │ │ │ ├── 106.txt │ │ │ │ ├── 107.txt │ │ │ │ ├── 108.txt │ │ │ │ ├── 109.txt │ │ │ │ ├── 110.txt │ │ │ │ ├── 111.txt │ │ │ │ ├── 112.txt │ │ │ │ ├── 113.txt │ │ │ │ ├── 114.txt │ │ │ │ ├── 115.txt │ │ │ │ ├── 116.txt │ │ │ │ ├── 117.txt │ │ │ │ ├── 118.txt │ │ │ │ ├── 119.txt │ │ │ │ ├── 120.txt │ │ │ │ ├── 121.txt │ │ │ │ ├── 122.txt │ │ │ │ ├── 123.txt │ │ │ │ ├── 124.txt │ │ │ │ ├── 125.txt │ │ │ │ ├── 126.txt │ │ │ │ ├── 127.txt │ │ │ │ ├── 128.txt │ │ │ │ ├── 129.txt │ │ │ │ ├── 130.txt │ │ │ │ ├── 131.txt │ │ │ │ ├── 132.txt │ │ │ │ ├── 133.txt │ │ │ │ ├── 134.txt │ │ │ │ ├── 135.txt │ │ │ │ ├── 136.txt │ │ │ │ ├── 137.txt │ │ │ │ ├── 138.txt │ │ │ │ ├── 139.txt │ │ │ │ ├── 140.txt │ │ │ │ ├── 141.txt │ │ │ │ ├── 142.txt │ │ │ │ ├── 143.txt │ │ │ │ ├── 144.txt │ │ │ │ ├── 145.txt │ │ │ │ ├── 146.txt │ │ │ │ ├── 147.txt │ │ │ │ ├── 148.txt │ │ │ │ ├── 149.txt │ │ │ │ ├── 150.txt │ │ │ │ ├── 151.txt │ │ │ │ ├── 152.txt │ │ │ │ ├── 153.txt │ │ │ │ ├── 154.txt │ │ │ │ ├── 155.txt │ │ │ │ ├── 156.txt │ │ │ │ ├── 157.txt │ │ │ │ ├── 158.txt │ │ │ │ ├── 159.txt │ │ │ │ ├── 160.txt │ │ │ │ ├── 161.txt │ │ │ │ ├── 162.txt │ │ │ │ ├── 163.txt │ │ │ │ ├── 164.txt │ │ │ │ ├── 165.txt │ │ │ │ ├── 166.txt │ │ │ │ ├── 167.txt │ │ │ │ ├── 168.txt │ │ │ │ ├── 169.txt │ │ │ │ ├── 170.txt │ │ │ │ ├── 171.txt │ │ │ │ ├── 172.txt │ │ │ │ ├── 173.txt │ │ │ │ ├── 174.txt │ │ │ │ ├── 175.txt │ │ │ │ ├── 176.txt │ │ │ │ ├── 177.txt │ │ │ │ ├── 178.txt │ │ │ │ ├── 179.txt │ │ │ │ ├── 180.txt │ │ │ │ ├── 181.txt │ │ │ │ ├── 182.txt │ │ │ │ ├── 183.txt │ │ │ │ ├── 184.txt │ │ │ │ ├── 185.txt │ │ │ │ ├── 186.txt │ │ │ │ ├── 187.txt │ │ │ │ ├── 188.txt │ │ │ │ ├── 189.txt │ │ │ │ ├── 190.txt │ │ │ │ ├── 191.txt │ │ │ │ ├── 192.txt │ │ │ │ ├── 193.txt │ │ │ │ ├── 194.txt │ │ │ │ ├── 195.txt │ │ │ │ ├── 196.txt │ │ │ │ ├── 197.txt │ │ │ │ ├── 198.txt │ │ │ │ ├── 199.txt │ │ │ │ ├── 200.txt │ │ │ │ ├── 201.txt │ │ │ │ ├── 202.txt │ │ │ │ ├── 203.txt │ │ │ │ ├── 204.txt │ │ │ │ ├── 205.txt │ │ │ │ ├── 206.txt │ │ │ │ ├── 207.txt │ │ │ │ ├── 208.txt │ │ │ │ ├── 209.txt │ │ │ │ ├── 210.txt │ │ │ │ ├── 211.txt │ │ │ │ ├── 212.txt │ │ │ │ ├── 213.txt │ │ │ │ ├── 214.txt │ │ │ │ ├── 215.txt │ │ │ │ ├── 216.txt │ │ │ │ ├── 217.txt │ │ │ │ ├── 218.txt │ │ │ │ ├── 219.txt │ │ │ │ ├── 220.txt │ │ │ │ ├── 221.txt │ │ │ │ ├── 222.txt │ │ │ │ ├── 223.txt │ │ │ │ ├── 224.txt │ │ │ │ ├── 225.txt │ │ │ │ ├── 226.txt │ │ │ │ ├── 227.txt │ │ │ │ ├── 228.txt │ │ │ │ ├── 229.txt │ │ │ │ ├── 230.txt │ │ │ │ ├── 231.txt │ │ │ │ ├── 232.txt │ │ │ │ ├── 233.txt │ │ │ │ ├── 234.txt │ │ │ │ ├── 235.txt │ │ │ │ ├── 236.txt │ │ │ │ ├── 237.txt │ │ │ │ ├── 238.txt │ │ │ │ ├── 239.txt │ │ │ │ ├── 240.txt │ │ │ │ ├── 241.txt │ │ │ │ ├── 242.txt │ │ │ │ ├── 243.txt │ │ │ │ ├── 244.txt │ │ │ │ ├── 245.txt │ │ │ │ ├── 246.txt │ │ │ │ ├── 247.txt │ │ │ │ ├── 248.txt │ │ │ │ ├── 249.txt │ │ │ │ ├── 250.txt │ │ │ │ ├── 251.txt │ │ │ │ ├── 252.txt │ │ │ │ ├── 253.txt │ │ │ │ ├── 254.txt │ │ │ │ ├── 255.txt │ │ │ │ ├── 256.txt │ │ │ │ ├── 257.txt │ │ │ │ ├── 258.txt │ │ │ │ ├── 259.txt │ │ │ │ ├── 260.txt │ │ │ │ ├── 261.txt │ │ │ │ ├── 262.txt │ │ │ │ ├── 263.txt │ │ │ │ ├── 264.txt │ │ │ │ ├── 265.txt │ │ │ │ ├── 266.txt │ │ │ │ ├── 267.txt │ │ │ │ ├── 268.txt │ │ │ │ ├── 269.txt │ │ │ │ ├── 270.txt │ │ │ │ ├── 271.txt │ │ │ │ ├── 272.txt │ │ │ │ ├── 273.txt │ │ │ │ ├── 274.txt │ │ │ │ ├── 275.txt │ │ │ │ ├── 276.txt │ │ │ │ ├── 277.txt │ │ │ │ ├── 278.txt │ │ │ │ ├── 279.txt │ │ │ │ ├── 280.txt │ │ │ │ ├── 281.txt │ │ │ │ ├── 282.txt │ │ │ │ ├── 283.txt │ │ │ │ ├── 284.txt │ │ │ │ ├── 285.txt │ │ │ │ ├── 286.txt │ │ │ │ ├── 287.txt │ │ │ │ ├── 288.txt │ │ │ │ ├── 289.txt │ │ │ │ ├── 290.txt │ │ │ │ ├── 291.txt │ │ │ │ ├── 292.txt │ │ │ │ ├── 293.txt │ │ │ │ ├── 294.txt │ │ │ │ ├── 295.txt │ │ │ │ ├── 296.txt │ │ │ │ ├── 297.txt │ │ │ │ ├── 298.txt │ │ │ │ ├── 299.txt │ │ │ │ ├── 300.txt │ │ │ │ ├── 301.txt │ │ │ │ ├── 302.txt │ │ │ │ ├── 303.txt │ │ │ │ ├── 304.txt │ │ │ │ ├── 305.txt │ │ │ │ ├── 306.txt │ │ │ │ ├── 307.txt │ │ │ │ ├── 308.txt │ │ │ │ ├── 309.txt │ │ │ │ ├── 310.txt │ │ │ │ ├── 311.txt │ │ │ │ ├── 312.txt │ │ │ │ ├── 313.txt │ │ │ │ ├── 314.txt │ │ │ │ ├── 315.txt │ │ │ │ ├── 316.txt │ │ │ │ ├── 317.txt │ │ │ │ ├── 318.txt │ │ │ │ ├── 319.txt │ │ │ │ ├── 320.txt │ │ │ │ ├── 321.txt │ │ │ │ ├── 322.txt │ │ │ │ ├── 323.txt │ │ │ │ ├── 324.txt │ │ │ │ ├── 325.txt │ │ │ │ ├── 326.txt │ │ │ │ ├── 327.txt │ │ │ │ ├── 328.txt │ │ │ │ ├── 329.txt │ │ │ │ ├── 330.txt │ │ │ │ ├── 331.txt │ │ │ │ ├── 332.txt │ │ │ │ ├── 333.txt │ │ │ │ ├── 334.txt │ │ │ │ ├── 335.txt │ │ │ │ ├── 336.txt │ │ │ │ ├── 337.txt │ │ │ │ ├── 338.txt │ │ │ │ ├── 339.txt │ │ │ │ ├── 340.txt │ │ │ │ ├── 341.txt │ │ │ │ ├── 342.txt │ │ │ │ ├── 343.txt │ │ │ │ ├── 344.txt │ │ │ │ ├── 345.txt │ │ │ │ ├── 346.txt │ │ │ │ ├── 347.txt │ │ │ │ ├── 348.txt │ │ │ │ ├── 349.txt │ │ │ │ ├── 350.txt │ │ │ │ ├── 351.txt │ │ │ │ ├── 352.txt │ │ │ │ ├── 353.txt │ │ │ │ ├── 354.txt │ │ │ │ ├── 355.txt │ │ │ │ ├── 356.txt │ │ │ │ ├── 357.txt │ │ │ │ ├── 358.txt │ │ │ │ ├── 359.txt │ │ │ │ ├── 360.txt │ │ │ │ ├── 361.txt │ │ │ │ ├── 362.txt │ │ │ │ ├── 363.txt │ │ │ │ ├── 364.txt │ │ │ │ ├── 365.txt │ │ │ │ ├── 366.txt │ │ │ │ ├── 367.txt │ │ │ │ ├── 368.txt │ │ │ │ ├── 369.txt │ │ │ │ ├── 370.txt │ │ │ │ ├── 371.txt │ │ │ │ ├── 372.txt │ │ │ │ ├── 373.txt │ │ │ │ ├── 374.txt │ │ │ │ ├── 375.txt │ │ │ │ ├── 376.txt │ │ │ │ ├── 377.txt │ │ │ │ ├── 378.txt │ │ │ │ ├── 379.txt │ │ │ │ ├── 380.txt │ │ │ │ ├── 381.txt │ │ │ │ ├── 382.txt │ │ │ │ ├── 383.txt │ │ │ │ ├── 384.txt │ │ │ │ ├── 385.txt │ │ │ │ ├── 386.txt │ │ │ │ ├── 387.txt │ │ │ │ ├── 388.txt │ │ │ │ ├── 389.txt │ │ │ │ ├── 390.txt │ │ │ │ ├── 391.txt │ │ │ │ ├── 392.txt │ │ │ │ ├── 393.txt │ │ │ │ ├── 394.txt │ │ │ │ ├── 395.txt │ │ │ │ ├── 396.txt │ │ │ │ ├── 397.txt │ │ │ │ ├── 398.txt │ │ │ │ ├── 399.txt │ │ │ │ ├── 400.txt │ │ │ │ ├── 401.txt │ │ │ │ ├── 402.txt │ │ │ │ ├── 403.txt │ │ │ │ ├── 404.txt │ │ │ │ ├── 405.txt │ │ │ │ ├── 406.txt │ │ │ │ ├── 407.txt │ │ │ │ ├── 408.txt │ │ │ │ ├── 409.txt │ │ │ │ ├── 410.txt │ │ │ │ ├── 411.txt │ │ │ │ ├── 412.txt │ │ │ │ ├── 413.txt │ │ │ │ ├── 414.txt │ │ │ │ ├── 415.txt │ │ │ │ ├── 416.txt │ │ │ │ ├── 417.txt │ │ │ │ ├── 418.txt │ │ │ │ ├── 419.txt │ │ │ │ ├── 420.txt │ │ │ │ ├── 421.txt │ │ │ │ ├── 422.txt │ │ │ │ ├── 423.txt │ │ │ │ ├── 424.txt │ │ │ │ ├── 425.txt │ │ │ │ ├── 426.txt │ │ │ │ ├── 427.txt │ │ │ │ ├── 428.txt │ │ │ │ ├── 429.txt │ │ │ │ ├── 430.txt │ │ │ │ ├── 431.txt │ │ │ │ ├── 432.txt │ │ │ │ ├── 433.txt │ │ │ │ ├── 434.txt │ │ │ │ ├── 435.txt │ │ │ │ ├── 436.txt │ │ │ │ ├── 437.txt │ │ │ │ ├── 438.txt │ │ │ │ ├── 439.txt │ │ │ │ ├── 440.txt │ │ │ │ ├── 441.txt │ │ │ │ ├── 442.txt │ │ │ │ ├── 443.txt │ │ │ │ ├── 444.txt │ │ │ │ ├── 445.txt │ │ │ │ ├── 446.txt │ │ │ │ ├── 447.txt │ │ │ │ ├── 448.txt │ │ │ │ ├── 449.txt │ │ │ │ ├── 450.txt │ │ │ │ ├── 451.txt │ │ │ │ ├── 452.txt │ │ │ │ ├── 453.txt │ │ │ │ ├── 454.txt │ │ │ │ ├── 455.txt │ │ │ │ ├── 456.txt │ │ │ │ ├── 457.txt │ │ │ │ ├── 458.txt │ │ │ │ ├── 459.txt │ │ │ │ ├── 460.txt │ │ │ │ ├── 461.txt │ │ │ │ ├── 462.txt │ │ │ │ ├── 463.txt │ │ │ │ ├── 464.txt │ │ │ │ ├── 465.txt │ │ │ │ ├── 466.txt │ │ │ │ ├── 467.txt │ │ │ │ ├── 468.txt │ │ │ │ ├── 469.txt │ │ │ │ ├── 470.txt │ │ │ │ ├── 471.txt │ │ │ │ ├── 472.txt │ │ │ │ ├── 473.txt │ │ │ │ ├── 474.txt │ │ │ │ ├── 475.txt │ │ │ │ ├── 476.txt │ │ │ │ ├── 477.txt │ │ │ │ ├── 478.txt │ │ │ │ ├── 479.txt │ │ │ │ ├── 480.txt │ │ │ │ ├── 481.txt │ │ │ │ ├── 482.txt │ │ │ │ ├── 483.txt │ │ │ │ ├── 484.txt │ │ │ │ ├── 485.txt │ │ │ │ ├── 486.txt │ │ │ │ ├── 487.txt │ │ │ │ ├── 488.txt │ │ │ │ ├── 489.txt │ │ │ │ ├── 490.txt │ │ │ │ ├── 491.txt │ │ │ │ ├── 492.txt │ │ │ │ ├── 493.txt │ │ │ │ ├── 494.txt │ │ │ │ ├── 495.txt │ │ │ │ ├── 496.txt │ │ │ │ ├── 497.txt │ │ │ │ ├── 498.txt │ │ │ │ ├── 499.txt │ │ │ │ ├── 500.txt │ │ │ │ ├── 501.txt │ │ │ │ ├── 502.txt │ │ │ │ ├── 503.txt │ │ │ │ ├── 504.txt │ │ │ │ ├── 505.txt │ │ │ │ ├── 506.txt │ │ │ │ ├── 507.txt │ │ │ │ ├── 508.txt │ │ │ │ ├── 509.txt │ │ │ │ └── 510.txt │ │ │ ├── entertainment │ │ │ │ ├── 001.txt │ │ │ │ ├── 002.txt │ │ │ │ ├── 003.txt │ │ │ │ ├── 004.txt │ │ │ │ ├── 005.txt │ │ │ │ ├── 006.txt │ │ │ │ ├── 007.txt │ │ │ │ ├── 008.txt │ │ │ │ ├── 009.txt │ │ │ │ ├── 010.txt │ │ │ │ ├── 011.txt │ │ │ │ ├── 012.txt │ │ │ │ ├── 013.txt │ │ │ │ ├── 014.txt │ │ │ │ ├── 015.txt │ │ │ │ ├── 016.txt │ │ │ │ ├── 017.txt │ │ │ │ ├── 018.txt │ │ │ │ ├── 019.txt │ │ │ │ ├── 020.txt │ │ │ │ ├── 021.txt │ │ │ │ ├── 022.txt │ │ │ │ ├── 023.txt │ │ │ │ ├── 024.txt │ │ │ │ ├── 025.txt │ │ │ │ ├── 026.txt │ │ │ │ ├── 027.txt │ │ │ │ ├── 028.txt │ │ │ │ ├── 029.txt │ │ │ │ ├── 030.txt │ │ │ │ ├── 031.txt │ │ │ │ ├── 032.txt │ │ │ │ ├── 033.txt │ │ │ │ ├── 034.txt │ │ │ │ ├── 035.txt │ │ │ │ ├── 036.txt │ │ │ │ ├── 037.txt │ │ │ │ ├── 038.txt │ │ │ │ ├── 039.txt │ │ │ │ ├── 040.txt │ │ │ │ ├── 041.txt │ │ │ │ ├── 042.txt │ │ │ │ ├── 043.txt │ │ │ │ ├── 044.txt │ │ │ │ ├── 045.txt │ │ │ │ ├── 046.txt │ │ │ │ ├── 047.txt │ │ │ │ ├── 048.txt │ │ │ │ ├── 049.txt │ │ │ │ ├── 050.txt │ │ │ │ ├── 051.txt │ │ │ │ ├── 052.txt │ │ │ │ ├── 053.txt │ │ │ │ ├── 054.txt │ │ │ │ ├── 055.txt │ │ │ │ ├── 056.txt │ │ │ │ ├── 057.txt │ │ │ │ ├── 058.txt │ │ │ │ ├── 059.txt │ │ │ │ ├── 060.txt │ │ │ │ ├── 061.txt │ │ │ │ ├── 062.txt │ │ │ │ ├── 063.txt │ │ │ │ ├── 064.txt │ │ │ │ ├── 065.txt │ │ │ │ ├── 066.txt │ │ │ │ ├── 067.txt │ │ │ │ ├── 068.txt │ │ │ │ ├── 069.txt │ │ │ │ ├── 070.txt │ │ │ │ ├── 071.txt │ │ │ │ ├── 072.txt │ │ │ │ ├── 073.txt │ │ │ │ ├── 074.txt │ │ │ │ ├── 075.txt │ │ │ │ ├── 076.txt │ │ │ │ ├── 077.txt │ │ │ │ ├── 078.txt │ │ │ │ ├── 079.txt │ │ │ │ ├── 080.txt │ │ │ │ ├── 081.txt │ │ │ │ ├── 082.txt │ │ │ │ ├── 083.txt │ │ │ │ ├── 084.txt │ │ │ │ ├── 085.txt │ │ │ │ ├── 086.txt │ │ │ │ ├── 087.txt │ │ │ │ ├── 088.txt │ │ │ │ ├── 089.txt │ │ │ │ ├── 090.txt │ │ │ │ ├── 091.txt │ │ │ │ ├── 092.txt │ │ │ │ ├── 093.txt │ │ │ │ ├── 094.txt │ │ │ │ ├── 095.txt │ │ │ │ ├── 096.txt │ │ │ │ ├── 097.txt │ │ │ │ ├── 098.txt │ │ │ │ ├── 099.txt │ │ │ │ ├── 100.txt │ │ │ │ ├── 101.txt │ │ │ │ ├── 102.txt │ │ │ │ ├── 103.txt │ │ │ │ ├── 104.txt │ │ │ │ ├── 105.txt │ │ │ │ ├── 106.txt │ │ │ │ ├── 107.txt │ │ │ │ ├── 108.txt │ │ │ │ ├── 109.txt │ │ │ │ ├── 110.txt │ │ │ │ ├── 111.txt │ │ │ │ ├── 112.txt │ │ │ │ ├── 113.txt │ │ │ │ ├── 114.txt │ │ │ │ ├── 115.txt │ │ │ │ ├── 116.txt │ │ │ │ ├── 117.txt │ │ │ │ ├── 118.txt │ │ │ │ ├── 119.txt │ │ │ │ ├── 120.txt │ │ │ │ ├── 121.txt │ │ │ │ ├── 122.txt │ │ │ │ ├── 123.txt │ │ │ │ ├── 124.txt │ │ │ │ ├── 125.txt │ │ │ │ ├── 126.txt │ │ │ │ ├── 127.txt │ │ │ │ ├── 128.txt │ │ │ │ ├── 129.txt │ │ │ │ ├── 130.txt │ │ │ │ ├── 131.txt │ │ │ │ ├── 132.txt │ │ │ │ ├── 133.txt │ │ │ │ ├── 134.txt │ │ │ │ ├── 135.txt │ │ │ │ ├── 136.txt │ │ │ │ ├── 137.txt │ │ │ │ ├── 138.txt │ │ │ │ ├── 139.txt │ │ │ │ ├── 140.txt │ │ │ │ ├── 141.txt │ │ │ │ ├── 142.txt │ │ │ │ ├── 143.txt │ │ │ │ ├── 144.txt │ │ │ │ ├── 145.txt │ │ │ │ ├── 146.txt │ │ │ │ ├── 147.txt │ │ │ │ ├── 148.txt │ │ │ │ ├── 149.txt │ │ │ │ ├── 150.txt │ │ │ │ ├── 151.txt │ │ │ │ ├── 152.txt │ │ │ │ ├── 153.txt │ │ │ │ ├── 154.txt │ │ │ │ ├── 155.txt │ │ │ │ ├── 156.txt │ │ │ │ ├── 157.txt │ │ │ │ ├── 158.txt │ │ │ │ ├── 159.txt │ │ │ │ ├── 160.txt │ │ │ │ ├── 161.txt │ │ │ │ ├── 162.txt │ │ │ │ ├── 163.txt │ │ │ │ ├── 164.txt │ │ │ │ ├── 165.txt │ │ │ │ ├── 166.txt │ │ │ │ ├── 167.txt │ │ │ │ ├── 168.txt │ │ │ │ ├── 169.txt │ │ │ │ ├── 170.txt │ │ │ │ ├── 171.txt │ │ │ │ ├── 172.txt │ │ │ │ ├── 173.txt │ │ │ │ ├── 174.txt │ │ │ │ ├── 175.txt │ │ │ │ ├── 176.txt │ │ │ │ ├── 177.txt │ │ │ │ ├── 178.txt │ │ │ │ ├── 179.txt │ │ │ │ ├── 180.txt │ │ │ │ ├── 181.txt │ │ │ │ ├── 182.txt │ │ │ │ ├── 183.txt │ │ │ │ ├── 184.txt │ │ │ │ ├── 185.txt │ │ │ │ ├── 186.txt │ │ │ │ ├── 187.txt │ │ │ │ ├── 188.txt │ │ │ │ ├── 189.txt │ │ │ │ ├── 190.txt │ │ │ │ ├── 191.txt │ │ │ │ ├── 192.txt │ │ │ │ ├── 193.txt │ │ │ │ ├── 194.txt │ │ │ │ ├── 195.txt │ │ │ │ ├── 196.txt │ │ │ │ ├── 197.txt │ │ │ │ ├── 198.txt │ │ │ │ ├── 199.txt │ │ │ │ ├── 200.txt │ │ │ │ ├── 201.txt │ │ │ │ ├── 202.txt │ │ │ │ ├── 203.txt │ │ │ │ ├── 204.txt │ │ │ │ ├── 205.txt │ │ │ │ ├── 206.txt │ │ │ │ ├── 207.txt │ │ │ │ ├── 208.txt │ │ │ │ ├── 209.txt │ │ │ │ ├── 210.txt │ │ │ │ ├── 211.txt │ │ │ │ ├── 212.txt │ │ │ │ ├── 213.txt │ │ │ │ ├── 214.txt │ │ │ │ ├── 215.txt │ │ │ │ ├── 216.txt │ │ │ │ ├── 217.txt │ │ │ │ ├── 218.txt │ │ │ │ ├── 219.txt │ │ │ │ ├── 220.txt │ │ │ │ ├── 221.txt │ │ │ │ ├── 222.txt │ │ │ │ ├── 223.txt │ │ │ │ ├── 224.txt │ │ │ │ ├── 225.txt │ │ │ │ ├── 226.txt │ │ │ │ ├── 227.txt │ │ │ │ ├── 228.txt │ │ │ │ ├── 229.txt │ │ │ │ ├── 230.txt │ │ │ │ ├── 231.txt │ │ │ │ ├── 232.txt │ │ │ │ ├── 233.txt │ │ │ │ ├── 234.txt │ │ │ │ ├── 235.txt │ │ │ │ ├── 236.txt │ │ │ │ ├── 237.txt │ │ │ │ ├── 238.txt │ │ │ │ ├── 239.txt │ │ │ │ ├── 240.txt │ │ │ │ ├── 241.txt │ │ │ │ ├── 242.txt │ │ │ │ ├── 243.txt │ │ │ │ ├── 244.txt │ │ │ │ ├── 245.txt │ │ │ │ ├── 246.txt │ │ │ │ ├── 247.txt │ │ │ │ ├── 248.txt │ │ │ │ ├── 249.txt │ │ │ │ ├── 250.txt │ │ │ │ ├── 251.txt │ │ │ │ ├── 252.txt │ │ │ │ ├── 253.txt │ │ │ │ ├── 254.txt │ │ │ │ ├── 255.txt │ │ │ │ ├── 256.txt │ │ │ │ ├── 257.txt │ │ │ │ ├── 258.txt │ │ │ │ ├── 259.txt │ │ │ │ ├── 260.txt │ │ │ │ ├── 261.txt │ │ │ │ ├── 262.txt │ │ │ │ ├── 263.txt │ │ │ │ ├── 264.txt │ │ │ │ ├── 265.txt │ │ │ │ ├── 266.txt │ │ │ │ ├── 267.txt │ │ │ │ ├── 268.txt │ │ │ │ ├── 269.txt │ │ │ │ ├── 270.txt │ │ │ │ ├── 271.txt │ │ │ │ ├── 272.txt │ │ │ │ ├── 273.txt │ │ │ │ ├── 274.txt │ │ │ │ ├── 275.txt │ │ │ │ ├── 276.txt │ │ │ │ ├── 277.txt │ │ │ │ ├── 278.txt │ │ │ │ ├── 279.txt │ │ │ │ ├── 280.txt │ │ │ │ ├── 281.txt │ │ │ │ ├── 282.txt │ │ │ │ ├── 283.txt │ │ │ │ ├── 284.txt │ │ │ │ ├── 285.txt │ │ │ │ ├── 286.txt │ │ │ │ ├── 287.txt │ │ │ │ ├── 288.txt │ │ │ │ ├── 289.txt │ │ │ │ ├── 290.txt │ │ │ │ ├── 291.txt │ │ │ │ ├── 292.txt │ │ │ │ ├── 293.txt │ │ │ │ ├── 294.txt │ │ │ │ ├── 295.txt │ │ │ │ ├── 296.txt │ │ │ │ ├── 297.txt │ │ │ │ ├── 298.txt │ │ │ │ ├── 299.txt │ │ │ │ ├── 300.txt │ │ │ │ ├── 301.txt │ │ │ │ ├── 302.txt │ │ │ │ ├── 303.txt │ │ │ │ ├── 304.txt │ │ │ │ ├── 305.txt │ │ │ │ ├── 306.txt │ │ │ │ ├── 307.txt │ │ │ │ ├── 308.txt │ │ │ │ ├── 309.txt │ │ │ │ ├── 310.txt │ │ │ │ ├── 311.txt │ │ │ │ ├── 312.txt │ │ │ │ ├── 313.txt │ │ │ │ ├── 314.txt │ │ │ │ ├── 315.txt │ │ │ │ ├── 316.txt │ │ │ │ ├── 317.txt │ │ │ │ ├── 318.txt │ │ │ │ ├── 319.txt │ │ │ │ ├── 320.txt │ │ │ │ ├── 321.txt │ │ │ │ ├── 322.txt │ │ │ │ ├── 323.txt │ │ │ │ ├── 324.txt │ │ │ │ ├── 325.txt │ │ │ │ ├── 326.txt │ │ │ │ ├── 327.txt │ │ │ │ ├── 328.txt │ │ │ │ ├── 329.txt │ │ │ │ ├── 330.txt │ │ │ │ ├── 331.txt │ │ │ │ ├── 332.txt │ │ │ │ ├── 333.txt │ │ │ │ ├── 334.txt │ │ │ │ ├── 335.txt │ │ │ │ ├── 336.txt │ │ │ │ ├── 337.txt │ │ │ │ ├── 338.txt │ │ │ │ ├── 339.txt │ │ │ │ ├── 340.txt │ │ │ │ ├── 341.txt │ │ │ │ ├── 342.txt │ │ │ │ ├── 343.txt │ │ │ │ ├── 344.txt │ │ │ │ ├── 345.txt │ │ │ │ ├── 346.txt │ │ │ │ ├── 347.txt │ │ │ │ ├── 348.txt │ │ │ │ ├── 349.txt │ │ │ │ ├── 350.txt │ │ │ │ ├── 351.txt │ │ │ │ ├── 352.txt │ │ │ │ ├── 353.txt │ │ │ │ ├── 354.txt │ │ │ │ ├── 355.txt │ │ │ │ ├── 356.txt │ │ │ │ ├── 357.txt │ │ │ │ ├── 358.txt │ │ │ │ ├── 359.txt │ │ │ │ ├── 360.txt │ │ │ │ ├── 361.txt │ │ │ │ ├── 362.txt │ │ │ │ ├── 363.txt │ │ │ │ ├── 364.txt │ │ │ │ ├── 365.txt │ │ │ │ ├── 366.txt │ │ │ │ ├── 367.txt │ │ │ │ ├── 368.txt │ │ │ │ ├── 369.txt │ │ │ │ ├── 370.txt │ │ │ │ ├── 371.txt │ │ │ │ ├── 372.txt │ │ │ │ ├── 373.txt │ │ │ │ ├── 374.txt │ │ │ │ ├── 375.txt │ │ │ │ ├── 376.txt │ │ │ │ ├── 377.txt │ │ │ │ ├── 378.txt │ │ │ │ ├── 379.txt │ │ │ │ ├── 380.txt │ │ │ │ ├── 381.txt │ │ │ │ ├── 382.txt │ │ │ │ ├── 383.txt │ │ │ │ ├── 384.txt │ │ │ │ ├── 385.txt │ │ │ │ └── 386.txt │ │ │ ├── politics │ │ │ │ ├── 001.txt │ │ │ │ ├── 002.txt │ │ │ │ ├── 003.txt │ │ │ │ ├── 004.txt │ │ │ │ ├── 005.txt │ │ │ │ ├── 006.txt │ │ │ │ ├── 007.txt │ │ │ │ ├── 008.txt │ │ │ │ ├── 009.txt │ │ │ │ ├── 010.txt │ │ │ │ ├── 011.txt │ │ │ │ ├── 012.txt │ │ │ │ ├── 013.txt │ │ │ │ ├── 014.txt │ │ │ │ ├── 015.txt │ │ │ │ ├── 016.txt │ │ │ │ ├── 017.txt │ │ │ │ ├── 018.txt │ │ │ │ ├── 019.txt │ │ │ │ ├── 020.txt │ │ │ │ ├── 021.txt │ │ │ │ ├── 022.txt │ │ │ │ ├── 023.txt │ │ │ │ ├── 024.txt │ │ │ │ ├── 025.txt │ │ │ │ ├── 026.txt │ │ │ │ ├── 027.txt │ │ │ │ ├── 028.txt │ │ │ │ ├── 029.txt │ │ │ │ ├── 030.txt │ │ │ │ ├── 031.txt │ │ │ │ ├── 032.txt │ │ │ │ ├── 033.txt │ │ │ │ ├── 034.txt │ │ │ │ ├── 035.txt │ │ │ │ ├── 036.txt │ │ │ │ ├── 037.txt │ │ │ │ ├── 038.txt │ │ │ │ ├── 039.txt │ │ │ │ ├── 040.txt │ │ │ │ ├── 041.txt │ │ │ │ ├── 042.txt │ │ │ │ ├── 043.txt │ │ │ │ ├── 044.txt │ │ │ │ ├── 045.txt │ │ │ │ ├── 046.txt │ │ │ │ ├── 047.txt │ │ │ │ ├── 048.txt │ │ │ │ ├── 049.txt │ │ │ │ ├── 050.txt │ │ │ │ ├── 051.txt │ │ │ │ ├── 052.txt │ │ │ │ ├── 053.txt │ │ │ │ ├── 054.txt │ │ │ │ ├── 055.txt │ │ │ │ ├── 056.txt │ │ │ │ ├── 057.txt │ │ │ │ ├── 058.txt │ │ │ │ ├── 059.txt │ │ │ │ ├── 060.txt │ │ │ │ ├── 061.txt │ │ │ │ ├── 062.txt │ │ │ │ ├── 063.txt │ │ │ │ ├── 064.txt │ │ │ │ ├── 065.txt │ │ │ │ ├── 066.txt │ │ │ │ ├── 067.txt │ │ │ │ ├── 068.txt │ │ │ │ ├── 069.txt │ │ │ │ ├── 070.txt │ │ │ │ ├── 071.txt │ │ │ │ ├── 072.txt │ │ │ │ ├── 073.txt │ │ │ │ ├── 074.txt │ │ │ │ ├── 075.txt │ │ │ │ ├── 076.txt │ │ │ │ ├── 077.txt │ │ │ │ ├── 078.txt │ │ │ │ ├── 079.txt │ │ │ │ ├── 080.txt │ │ │ │ ├── 081.txt │ │ │ │ ├── 082.txt │ │ │ │ ├── 083.txt │ │ │ │ ├── 084.txt │ │ │ │ ├── 085.txt │ │ │ │ ├── 086.txt │ │ │ │ ├── 087.txt │ │ │ │ ├── 088.txt │ │ │ │ ├── 089.txt │ │ │ │ ├── 090.txt │ │ │ │ ├── 091.txt │ │ │ │ ├── 092.txt │ │ │ │ ├── 093.txt │ │ │ │ ├── 094.txt │ │ │ │ ├── 095.txt │ │ │ │ ├── 096.txt │ │ │ │ ├── 097.txt │ │ │ │ ├── 098.txt │ │ │ │ ├── 099.txt │ │ │ │ ├── 100.txt │ │ │ │ ├── 101.txt │ │ │ │ ├── 102.txt │ │ │ │ ├── 103.txt │ │ │ │ ├── 104.txt │ │ │ │ ├── 105.txt │ │ │ │ ├── 106.txt │ │ │ │ ├── 107.txt │ │ │ │ ├── 108.txt │ │ │ │ ├── 109.txt │ │ │ │ ├── 110.txt │ │ │ │ ├── 111.txt │ │ │ │ ├── 112.txt │ │ │ │ ├── 113.txt │ │ │ │ ├── 114.txt │ │ │ │ ├── 115.txt │ │ │ │ ├── 116.txt │ │ │ │ ├── 117.txt │ │ │ │ ├── 118.txt │ │ │ │ ├── 119.txt │ │ │ │ ├── 120.txt │ │ │ │ ├── 121.txt │ │ │ │ ├── 122.txt │ │ │ │ ├── 123.txt │ │ │ │ ├── 124.txt │ │ │ │ ├── 125.txt │ │ │ │ ├── 126.txt │ │ │ │ ├── 127.txt │ │ │ │ ├── 128.txt │ │ │ │ ├── 129.txt │ │ │ │ ├── 130.txt │ │ │ │ ├── 131.txt │ │ │ │ ├── 132.txt │ │ │ │ ├── 133.txt │ │ │ │ ├── 134.txt │ │ │ │ ├── 135.txt │ │ │ │ ├── 136.txt │ │ │ │ ├── 137.txt │ │ │ │ ├── 138.txt │ │ │ │ ├── 139.txt │ │ │ │ ├── 140.txt │ │ │ │ ├── 141.txt │ │ │ │ ├── 142.txt │ │ │ │ ├── 143.txt │ │ │ │ ├── 144.txt │ │ │ │ ├── 145.txt │ │ │ │ ├── 146.txt │ │ │ │ ├── 147.txt │ │ │ │ ├── 148.txt │ │ │ │ ├── 149.txt │ │ │ │ ├── 150.txt │ │ │ │ ├── 151.txt │ │ │ │ ├── 152.txt │ │ │ │ ├── 153.txt │ │ │ │ ├── 154.txt │ │ │ │ ├── 155.txt │ │ │ │ ├── 156.txt │ │ │ │ ├── 157.txt │ │ │ │ ├── 158.txt │ │ │ │ ├── 159.txt │ │ │ │ ├── 160.txt │ │ │ │ ├── 161.txt │ │ │ │ ├── 162.txt │ │ │ │ ├── 163.txt │ │ │ │ ├── 164.txt │ │ │ │ ├── 165.txt │ │ │ │ ├── 166.txt │ │ │ │ ├── 167.txt │ │ │ │ ├── 168.txt │ │ │ │ ├── 169.txt │ │ │ │ ├── 170.txt │ │ │ │ ├── 171.txt │ │ │ │ ├── 172.txt │ │ │ │ ├── 173.txt │ │ │ │ ├── 174.txt │ │ │ │ ├── 175.txt │ │ │ │ ├── 176.txt │ │ │ │ ├── 177.txt │ │ │ │ ├── 178.txt │ │ │ │ ├── 179.txt │ │ │ │ ├── 180.txt │ │ │ │ ├── 181.txt │ │ │ │ ├── 182.txt │ │ │ │ ├── 183.txt │ │ │ │ ├── 184.txt │ │ │ │ ├── 185.txt │ │ │ │ ├── 186.txt │ │ │ │ ├── 187.txt │ │ │ │ ├── 188.txt │ │ │ │ ├── 189.txt │ │ │ │ ├── 190.txt │ │ │ │ ├── 191.txt │ │ │ │ ├── 192.txt │ │ │ │ ├── 193.txt │ │ │ │ ├── 194.txt │ │ │ │ ├── 195.txt │ │ │ │ ├── 196.txt │ │ │ │ ├── 197.txt │ │ │ │ ├── 198.txt │ │ │ │ ├── 199.txt │ │ │ │ ├── 200.txt │ │ │ │ ├── 201.txt │ │ │ │ ├── 202.txt │ │ │ │ ├── 203.txt │ │ │ │ ├── 204.txt │ │ │ │ ├── 205.txt │ │ │ │ ├── 206.txt │ │ │ │ ├── 207.txt │ │ │ │ ├── 208.txt │ │ │ │ ├── 209.txt │ │ │ │ ├── 210.txt │ │ │ │ ├── 211.txt │ │ │ │ ├── 212.txt │ │ │ │ ├── 213.txt │ │ │ │ ├── 214.txt │ │ │ │ ├── 215.txt │ │ │ │ ├── 216.txt │ │ │ │ ├── 217.txt │ │ │ │ ├── 218.txt │ │ │ │ ├── 219.txt │ │ │ │ ├── 220.txt │ │ │ │ ├── 221.txt │ │ │ │ ├── 222.txt │ │ │ │ ├── 223.txt │ │ │ │ ├── 224.txt │ │ │ │ ├── 225.txt │ │ │ │ ├── 226.txt │ │ │ │ ├── 227.txt │ │ │ │ ├── 228.txt │ │ │ │ ├── 229.txt │ │ │ │ ├── 230.txt │ │ │ │ ├── 231.txt │ │ │ │ ├── 232.txt │ │ │ │ ├── 233.txt │ │ │ │ ├── 234.txt │ │ │ │ ├── 235.txt │ │ │ │ ├── 236.txt │ │ │ │ ├── 237.txt │ │ │ │ ├── 238.txt │ │ │ │ ├── 239.txt │ │ │ │ ├── 240.txt │ │ │ │ ├── 241.txt │ │ │ │ ├── 242.txt │ │ │ │ ├── 243.txt │ │ │ │ ├── 244.txt │ │ │ │ ├── 245.txt │ │ │ │ ├── 246.txt │ │ │ │ ├── 247.txt │ │ │ │ ├── 248.txt │ │ │ │ ├── 249.txt │ │ │ │ ├── 250.txt │ │ │ │ ├── 251.txt │ │ │ │ ├── 252.txt │ │ │ │ ├── 253.txt │ │ │ │ ├── 254.txt │ │ │ │ ├── 255.txt │ │ │ │ ├── 256.txt │ │ │ │ ├── 257.txt │ │ │ │ ├── 258.txt │ │ │ │ ├── 259.txt │ │ │ │ ├── 260.txt │ │ │ │ ├── 261.txt │ │ │ │ ├── 262.txt │ │ │ │ ├── 263.txt │ │ │ │ ├── 264.txt │ │ │ │ ├── 265.txt │ │ │ │ ├── 266.txt │ │ │ │ ├── 267.txt │ │ │ │ ├── 268.txt │ │ │ │ ├── 269.txt │ │ │ │ ├── 270.txt │ │ │ │ ├── 271.txt │ │ │ │ ├── 272.txt │ │ │ │ ├── 273.txt │ │ │ │ ├── 274.txt │ │ │ │ ├── 275.txt │ │ │ │ ├── 276.txt │ │ │ │ ├── 277.txt │ │ │ │ ├── 278.txt │ │ │ │ ├── 279.txt │ │ │ │ ├── 280.txt │ │ │ │ ├── 281.txt │ │ │ │ ├── 282.txt │ │ │ │ ├── 283.txt │ │ │ │ ├── 284.txt │ │ │ │ ├── 285.txt │ │ │ │ ├── 286.txt │ │ │ │ ├── 287.txt │ │ │ │ ├── 288.txt │ │ │ │ ├── 289.txt │ │ │ │ ├── 290.txt │ │ │ │ ├── 291.txt │ │ │ │ ├── 292.txt │ │ │ │ ├── 293.txt │ │ │ │ ├── 294.txt │ │ │ │ ├── 295.txt │ │ │ │ ├── 296.txt │ │ │ │ ├── 297.txt │ │ │ │ ├── 298.txt │ │ │ │ ├── 299.txt │ │ │ │ ├── 300.txt │ │ │ │ ├── 301.txt │ │ │ │ ├── 302.txt │ │ │ │ ├── 303.txt │ │ │ │ ├── 304.txt │ │ │ │ ├── 305.txt │ │ │ │ ├── 306.txt │ │ │ │ ├── 307.txt │ │ │ │ ├── 308.txt │ │ │ │ ├── 309.txt │ │ │ │ ├── 310.txt │ │ │ │ ├── 311.txt │ │ │ │ ├── 312.txt │ │ │ │ ├── 313.txt │ │ │ │ ├── 314.txt │ │ │ │ ├── 315.txt │ │ │ │ ├── 316.txt │ │ │ │ ├── 317.txt │ │ │ │ ├── 318.txt │ │ │ │ ├── 319.txt │ │ │ │ ├── 320.txt │ │ │ │ ├── 321.txt │ │ │ │ ├── 322.txt │ │ │ │ ├── 323.txt │ │ │ │ ├── 324.txt │ │ │ │ ├── 325.txt │ │ │ │ ├── 326.txt │ │ │ │ ├── 327.txt │ │ │ │ ├── 328.txt │ │ │ │ ├── 329.txt │ │ │ │ ├── 330.txt │ │ │ │ ├── 331.txt │ │ │ │ ├── 332.txt │ │ │ │ ├── 333.txt │ │ │ │ ├── 334.txt │ │ │ │ ├── 335.txt │ │ │ │ ├── 336.txt │ │ │ │ ├── 337.txt │ │ │ │ ├── 338.txt │ │ │ │ ├── 339.txt │ │ │ │ ├── 340.txt │ │ │ │ ├── 341.txt │ │ │ │ ├── 342.txt │ │ │ │ ├── 343.txt │ │ │ │ ├── 344.txt │ │ │ │ ├── 345.txt │ │ │ │ ├── 346.txt │ │ │ │ ├── 347.txt │ │ │ │ ├── 348.txt │ │ │ │ ├── 349.txt │ │ │ │ ├── 350.txt │ │ │ │ ├── 351.txt │ │ │ │ ├── 352.txt │ │ │ │ ├── 353.txt │ │ │ │ ├── 354.txt │ │ │ │ ├── 355.txt │ │ │ │ ├── 356.txt │ │ │ │ ├── 357.txt │ │ │ │ ├── 358.txt │ │ │ │ ├── 359.txt │ │ │ │ ├── 360.txt │ │ │ │ ├── 361.txt │ │ │ │ ├── 362.txt │ │ │ │ ├── 363.txt │ │ │ │ ├── 364.txt │ │ │ │ ├── 365.txt │ │ │ │ ├── 366.txt │ │ │ │ ├── 367.txt │ │ │ │ ├── 368.txt │ │ │ │ ├── 369.txt │ │ │ │ ├── 370.txt │ │ │ │ ├── 371.txt │ │ │ │ ├── 372.txt │ │ │ │ ├── 373.txt │ │ │ │ ├── 374.txt │ │ │ │ ├── 375.txt │ │ │ │ ├── 376.txt │ │ │ │ ├── 377.txt │ │ │ │ ├── 378.txt │ │ │ │ ├── 379.txt │ │ │ │ ├── 380.txt │ │ │ │ ├── 381.txt │ │ │ │ ├── 382.txt │ │ │ │ ├── 383.txt │ │ │ │ ├── 384.txt │ │ │ │ ├── 385.txt │ │ │ │ ├── 386.txt │ │ │ │ ├── 387.txt │ │ │ │ ├── 388.txt │ │ │ │ ├── 389.txt │ │ │ │ ├── 390.txt │ │ │ │ ├── 391.txt │ │ │ │ ├── 392.txt │ │ │ │ ├── 393.txt │ │ │ │ ├── 394.txt │ │ │ │ ├── 395.txt │ │ │ │ ├── 396.txt │ │ │ │ ├── 397.txt │ │ │ │ ├── 398.txt │ │ │ │ ├── 399.txt │ │ │ │ ├── 400.txt │ │ │ │ ├── 401.txt │ │ │ │ ├── 402.txt │ │ │ │ ├── 403.txt │ │ │ │ ├── 404.txt │ │ │ │ ├── 405.txt │ │ │ │ ├── 406.txt │ │ │ │ ├── 407.txt │ │ │ │ ├── 408.txt │ │ │ │ ├── 409.txt │ │ │ │ ├── 410.txt │ │ │ │ ├── 411.txt │ │ │ │ ├── 412.txt │ │ │ │ ├── 413.txt │ │ │ │ ├── 414.txt │ │ │ │ ├── 415.txt │ │ │ │ ├── 416.txt │ │ │ │ └── 417.txt │ │ │ ├── sport │ │ │ │ ├── 001.txt │ │ │ │ ├── 002.txt │ │ │ │ ├── 003.txt │ │ │ │ ├── 004.txt │ │ │ │ ├── 005.txt │ │ │ │ ├── 006.txt │ │ │ │ ├── 007.txt │ │ │ │ ├── 008.txt │ │ │ │ ├── 009.txt │ │ │ │ ├── 010.txt │ │ │ │ ├── 011.txt │ │ │ │ ├── 012.txt │ │ │ │ ├── 013.txt │ │ │ │ ├── 014.txt │ │ │ │ ├── 015.txt │ │ │ │ ├── 016.txt │ │ │ │ ├── 017.txt │ │ │ │ ├── 018.txt │ │ │ │ ├── 019.txt │ │ │ │ ├── 020.txt │ │ │ │ ├── 021.txt │ │ │ │ ├── 022.txt │ │ │ │ ├── 023.txt │ │ │ │ ├── 024.txt │ │ │ │ ├── 025.txt │ │ │ │ ├── 026.txt │ │ │ │ ├── 027.txt │ │ │ │ ├── 028.txt │ │ │ │ ├── 029.txt │ │ │ │ ├── 030.txt │ │ │ │ ├── 031.txt │ │ │ │ ├── 032.txt │ │ │ │ ├── 033.txt │ │ │ │ ├── 034.txt │ │ │ │ ├── 035.txt │ │ │ │ ├── 036.txt │ │ │ │ ├── 037.txt │ │ │ │ ├── 038.txt │ │ │ │ ├── 039.txt │ │ │ │ ├── 040.txt │ │ │ │ ├── 041.txt │ │ │ │ ├── 042.txt │ │ │ │ ├── 043.txt │ │ │ │ ├── 044.txt │ │ │ │ ├── 045.txt │ │ │ │ ├── 046.txt │ │ │ │ ├── 047.txt │ │ │ │ ├── 048.txt │ │ │ │ ├── 049.txt │ │ │ │ ├── 050.txt │ │ │ │ ├── 051.txt │ │ │ │ ├── 052.txt │ │ │ │ ├── 053.txt │ │ │ │ ├── 054.txt │ │ │ │ ├── 055.txt │ │ │ │ ├── 056.txt │ │ │ │ ├── 057.txt │ │ │ │ ├── 058.txt │ │ │ │ ├── 059.txt │ │ │ │ ├── 060.txt │ │ │ │ ├── 061.txt │ │ │ │ ├── 062.txt │ │ │ │ ├── 063.txt │ │ │ │ ├── 064.txt │ │ │ │ ├── 065.txt │ │ │ │ ├── 066.txt │ │ │ │ ├── 067.txt │ │ │ │ ├── 068.txt │ │ │ │ ├── 069.txt │ │ │ │ ├── 070.txt │ │ │ │ ├── 071.txt │ │ │ │ ├── 072.txt │ │ │ │ ├── 073.txt │ │ │ │ ├── 074.txt │ │ │ │ ├── 075.txt │ │ │ │ ├── 076.txt │ │ │ │ ├── 077.txt │ │ │ │ ├── 078.txt │ │ │ │ ├── 079.txt │ │ │ │ ├── 080.txt │ │ │ │ ├── 081.txt │ │ │ │ ├── 082.txt │ │ │ │ ├── 083.txt │ │ │ │ ├── 084.txt │ │ │ │ ├── 085.txt │ │ │ │ ├── 086.txt │ │ │ │ ├── 087.txt │ │ │ │ ├── 088.txt │ │ │ │ ├── 089.txt │ │ │ │ ├── 090.txt │ │ │ │ ├── 091.txt │ │ │ │ ├── 092.txt │ │ │ │ ├── 093.txt │ │ │ │ ├── 094.txt │ │ │ │ ├── 095.txt │ │ │ │ ├── 096.txt │ │ │ │ ├── 097.txt │ │ │ │ ├── 098.txt │ │ │ │ ├── 099.txt │ │ │ │ ├── 100.txt │ │ │ │ ├── 101.txt │ │ │ │ ├── 102.txt │ │ │ │ ├── 103.txt │ │ │ │ ├── 104.txt │ │ │ │ ├── 105.txt │ │ │ │ ├── 106.txt │ │ │ │ ├── 107.txt │ │ │ │ ├── 108.txt │ │ │ │ ├── 109.txt │ │ │ │ ├── 110.txt │ │ │ │ ├── 111.txt │ │ │ │ ├── 112.txt │ │ │ │ ├── 113.txt │ │ │ │ ├── 114.txt │ │ │ │ ├── 115.txt │ │ │ │ ├── 116.txt │ │ │ │ ├── 117.txt │ │ │ │ ├── 118.txt │ │ │ │ ├── 119.txt │ │ │ │ ├── 120.txt │ │ │ │ ├── 121.txt │ │ │ │ ├── 122.txt │ │ │ │ ├── 123.txt │ │ │ │ ├── 124.txt │ │ │ │ ├── 125.txt │ │ │ │ ├── 126.txt │ │ │ │ ├── 127.txt │ │ │ │ ├── 128.txt │ │ │ │ ├── 129.txt │ │ │ │ ├── 130.txt │ │ │ │ ├── 131.txt │ │ │ │ ├── 132.txt │ │ │ │ ├── 133.txt │ │ │ │ ├── 134.txt │ │ │ │ ├── 135.txt │ │ │ │ ├── 136.txt │ │ │ │ ├── 137.txt │ │ │ │ ├── 138.txt │ │ │ │ ├── 139.txt │ │ │ │ ├── 140.txt │ │ │ │ ├── 141.txt │ │ │ │ ├── 142.txt │ │ │ │ ├── 143.txt │ │ │ │ ├── 144.txt │ │ │ │ ├── 145.txt │ │ │ │ ├── 146.txt │ │ │ │ ├── 147.txt │ │ │ │ ├── 148.txt │ │ │ │ ├── 149.txt │ │ │ │ ├── 150.txt │ │ │ │ ├── 151.txt │ │ │ │ ├── 152.txt │ │ │ │ ├── 153.txt │ │ │ │ ├── 154.txt │ │ │ │ ├── 155.txt │ │ │ │ ├── 156.txt │ │ │ │ ├── 157.txt │ │ │ │ ├── 158.txt │ │ │ │ ├── 159.txt │ │ │ │ ├── 160.txt │ │ │ │ ├── 161.txt │ │ │ │ ├── 162.txt │ │ │ │ ├── 163.txt │ │ │ │ ├── 164.txt │ │ │ │ ├── 165.txt │ │ │ │ ├── 166.txt │ │ │ │ ├── 167.txt │ │ │ │ ├── 168.txt │ │ │ │ ├── 169.txt │ │ │ │ ├── 170.txt │ │ │ │ ├── 171.txt │ │ │ │ ├── 172.txt │ │ │ │ ├── 173.txt │ │ │ │ ├── 174.txt │ │ │ │ ├── 175.txt │ │ │ │ ├── 176.txt │ │ │ │ ├── 177.txt │ │ │ │ ├── 178.txt │ │ │ │ ├── 179.txt │ │ │ │ ├── 180.txt │ │ │ │ ├── 181.txt │ │ │ │ ├── 182.txt │ │ │ │ ├── 183.txt │ │ │ │ ├── 184.txt │ │ │ │ ├── 185.txt │ │ │ │ ├── 186.txt │ │ │ │ ├── 187.txt │ │ │ │ ├── 188.txt │ │ │ │ ├── 189.txt │ │ │ │ ├── 190.txt │ │ │ │ ├── 191.txt │ │ │ │ ├── 192.txt │ │ │ │ ├── 193.txt │ │ │ │ ├── 194.txt │ │ │ │ ├── 195.txt │ │ │ │ ├── 196.txt │ │ │ │ ├── 197.txt │ │ │ │ ├── 198.txt │ │ │ │ ├── 199.txt │ │ │ │ ├── 200.txt │ │ │ │ ├── 201.txt │ │ │ │ ├── 202.txt │ │ │ │ ├── 203.txt │ │ │ │ ├── 204.txt │ │ │ │ ├── 205.txt │ │ │ │ ├── 206.txt │ │ │ │ ├── 207.txt │ │ │ │ ├── 208.txt │ │ │ │ ├── 209.txt │ │ │ │ ├── 210.txt │ │ │ │ ├── 211.txt │ │ │ │ ├── 212.txt │ │ │ │ ├── 213.txt │ │ │ │ ├── 214.txt │ │ │ │ ├── 215.txt │ │ │ │ ├── 216.txt │ │ │ │ ├── 217.txt │ │ │ │ ├── 218.txt │ │ │ │ ├── 219.txt │ │ │ │ ├── 220.txt │ │ │ │ ├── 221.txt │ │ │ │ ├── 222.txt │ │ │ │ ├── 223.txt │ │ │ │ ├── 224.txt │ │ │ │ ├── 225.txt │ │ │ │ ├── 226.txt │ │ │ │ ├── 227.txt │ │ │ │ ├── 228.txt │ │ │ │ ├── 229.txt │ │ │ │ ├── 230.txt │ │ │ │ ├── 231.txt │ │ │ │ ├── 232.txt │ │ │ │ ├── 233.txt │ │ │ │ ├── 234.txt │ │ │ │ ├── 235.txt │ │ │ │ ├── 236.txt │ │ │ │ ├── 237.txt │ │ │ │ ├── 238.txt │ │ │ │ ├── 239.txt │ │ │ │ ├── 240.txt │ │ │ │ ├── 241.txt │ │ │ │ ├── 242.txt │ │ │ │ ├── 243.txt │ │ │ │ ├── 244.txt │ │ │ │ ├── 245.txt │ │ │ │ ├── 246.txt │ │ │ │ ├── 247.txt │ │ │ │ ├── 248.txt │ │ │ │ ├── 249.txt │ │ │ │ ├── 250.txt │ │ │ │ ├── 251.txt │ │ │ │ ├── 252.txt │ │ │ │ ├── 253.txt │ │ │ │ ├── 254.txt │ │ │ │ ├── 255.txt │ │ │ │ ├── 256.txt │ │ │ │ ├── 257.txt │ │ │ │ ├── 258.txt │ │ │ │ ├── 259.txt │ │ │ │ ├── 260.txt │ │ │ │ ├── 261.txt │ │ │ │ ├── 262.txt │ │ │ │ ├── 263.txt │ │ │ │ ├── 264.txt │ │ │ │ ├── 265.txt │ │ │ │ ├── 266.txt │ │ │ │ ├── 267.txt │ │ │ │ ├── 268.txt │ │ │ │ ├── 269.txt │ │ │ │ ├── 270.txt │ │ │ │ ├── 271.txt │ │ │ │ ├── 272.txt │ │ │ │ ├── 273.txt │ │ │ │ ├── 274.txt │ │ │ │ ├── 275.txt │ │ │ │ ├── 276.txt │ │ │ │ ├── 277.txt │ │ │ │ ├── 278.txt │ │ │ │ ├── 279.txt │ │ │ │ ├── 280.txt │ │ │ │ ├── 281.txt │ │ │ │ ├── 282.txt │ │ │ │ ├── 283.txt │ │ │ │ ├── 284.txt │ │ │ │ ├── 285.txt │ │ │ │ ├── 286.txt │ │ │ │ ├── 287.txt │ │ │ │ ├── 288.txt │ │ │ │ ├── 289.txt │ │ │ │ ├── 290.txt │ │ │ │ ├── 291.txt │ │ │ │ ├── 292.txt │ │ │ │ ├── 293.txt │ │ │ │ ├── 294.txt │ │ │ │ ├── 295.txt │ │ │ │ ├── 296.txt │ │ │ │ ├── 297.txt │ │ │ │ ├── 298.txt │ │ │ │ ├── 299.txt │ │ │ │ ├── 300.txt │ │ │ │ ├── 301.txt │ │ │ │ ├── 302.txt │ │ │ │ ├── 303.txt │ │ │ │ ├── 304.txt │ │ │ │ ├── 305.txt │ │ │ │ ├── 306.txt │ │ │ │ ├── 307.txt │ │ │ │ ├── 308.txt │ │ │ │ ├── 309.txt │ │ │ │ ├── 310.txt │ │ │ │ ├── 311.txt │ │ │ │ ├── 312.txt │ │ │ │ ├── 313.txt │ │ │ │ ├── 314.txt │ │ │ │ ├── 315.txt │ │ │ │ ├── 316.txt │ │ │ │ ├── 317.txt │ │ │ │ ├── 318.txt │ │ │ │ ├── 319.txt │ │ │ │ ├── 320.txt │ │ │ │ ├── 321.txt │ │ │ │ ├── 322.txt │ │ │ │ ├── 323.txt │ │ │ │ ├── 324.txt │ │ │ │ ├── 325.txt │ │ │ │ ├── 326.txt │ │ │ │ ├── 327.txt │ │ │ │ ├── 328.txt │ │ │ │ ├── 329.txt │ │ │ │ ├── 330.txt │ │ │ │ ├── 331.txt │ │ │ │ ├── 332.txt │ │ │ │ ├── 333.txt │ │ │ │ ├── 334.txt │ │ │ │ ├── 335.txt │ │ │ │ ├── 336.txt │ │ │ │ ├── 337.txt │ │ │ │ ├── 338.txt │ │ │ │ ├── 339.txt │ │ │ │ ├── 340.txt │ │ │ │ ├── 341.txt │ │ │ │ ├── 342.txt │ │ │ │ ├── 343.txt │ │ │ │ ├── 344.txt │ │ │ │ ├── 345.txt │ │ │ │ ├── 346.txt │ │ │ │ ├── 347.txt │ │ │ │ ├── 348.txt │ │ │ │ ├── 349.txt │ │ │ │ ├── 350.txt │ │ │ │ ├── 351.txt │ │ │ │ ├── 352.txt │ │ │ │ ├── 353.txt │ │ │ │ ├── 354.txt │ │ │ │ ├── 355.txt │ │ │ │ ├── 356.txt │ │ │ │ ├── 357.txt │ │ │ │ ├── 358.txt │ │ │ │ ├── 359.txt │ │ │ │ ├── 360.txt │ │ │ │ ├── 361.txt │ │ │ │ ├── 362.txt │ │ │ │ ├── 363.txt │ │ │ │ ├── 364.txt │ │ │ │ ├── 365.txt │ │ │ │ ├── 366.txt │ │ │ │ ├── 367.txt │ │ │ │ ├── 368.txt │ │ │ │ ├── 369.txt │ │ │ │ ├── 370.txt │ │ │ │ ├── 371.txt │ │ │ │ ├── 372.txt │ │ │ │ ├── 373.txt │ │ │ │ ├── 374.txt │ │ │ │ ├── 375.txt │ │ │ │ ├── 376.txt │ │ │ │ ├── 377.txt │ │ │ │ ├── 378.txt │ │ │ │ ├── 379.txt │ │ │ │ ├── 380.txt │ │ │ │ ├── 381.txt │ │ │ │ ├── 382.txt │ │ │ │ ├── 383.txt │ │ │ │ ├── 384.txt │ │ │ │ ├── 385.txt │ │ │ │ ├── 386.txt │ │ │ │ ├── 387.txt │ │ │ │ ├── 388.txt │ │ │ │ ├── 389.txt │ │ │ │ ├── 390.txt │ │ │ │ ├── 391.txt │ │ │ │ ├── 392.txt │ │ │ │ ├── 393.txt │ │ │ │ ├── 394.txt │ │ │ │ ├── 395.txt │ │ │ │ ├── 396.txt │ │ │ │ ├── 397.txt │ │ │ │ ├── 398.txt │ │ │ │ ├── 399.txt │ │ │ │ ├── 400.txt │ │ │ │ ├── 401.txt │ │ │ │ ├── 402.txt │ │ │ │ ├── 403.txt │ │ │ │ ├── 404.txt │ │ │ │ ├── 405.txt │ │ │ │ ├── 406.txt │ │ │ │ ├── 407.txt │ │ │ │ ├── 408.txt │ │ │ │ ├── 409.txt │ │ │ │ ├── 410.txt │ │ │ │ ├── 411.txt │ │ │ │ ├── 412.txt │ │ │ │ ├── 413.txt │ │ │ │ ├── 414.txt │ │ │ │ ├── 415.txt │ │ │ │ ├── 416.txt │ │ │ │ ├── 417.txt │ │ │ │ ├── 418.txt │ │ │ │ ├── 419.txt │ │ │ │ ├── 420.txt │ │ │ │ ├── 421.txt │ │ │ │ ├── 422.txt │ │ │ │ ├── 423.txt │ │ │ │ ├── 424.txt │ │ │ │ ├── 425.txt │ │ │ │ ├── 426.txt │ │ │ │ ├── 427.txt │ │ │ │ ├── 428.txt │ │ │ │ ├── 429.txt │ │ │ │ ├── 430.txt │ │ │ │ ├── 431.txt │ │ │ │ ├── 432.txt │ │ │ │ ├── 433.txt │ │ │ │ ├── 434.txt │ │ │ │ ├── 435.txt │ │ │ │ ├── 436.txt │ │ │ │ ├── 437.txt │ │ │ │ ├── 438.txt │ │ │ │ ├── 439.txt │ │ │ │ ├── 440.txt │ │ │ │ ├── 441.txt │ │ │ │ ├── 442.txt │ │ │ │ ├── 443.txt │ │ │ │ ├── 444.txt │ │ │ │ ├── 445.txt │ │ │ │ ├── 446.txt │ │ │ │ ├── 447.txt │ │ │ │ ├── 448.txt │ │ │ │ ├── 449.txt │ │ │ │ ├── 450.txt │ │ │ │ ├── 451.txt │ │ │ │ ├── 452.txt │ │ │ │ ├── 453.txt │ │ │ │ ├── 454.txt │ │ │ │ ├── 455.txt │ │ │ │ ├── 456.txt │ │ │ │ ├── 457.txt │ │ │ │ ├── 458.txt │ │ │ │ ├── 459.txt │ │ │ │ ├── 460.txt │ │ │ │ ├── 461.txt │ │ │ │ ├── 462.txt │ │ │ │ ├── 463.txt │ │ │ │ ├── 464.txt │ │ │ │ ├── 465.txt │ │ │ │ ├── 466.txt │ │ │ │ ├── 467.txt │ │ │ │ ├── 468.txt │ │ │ │ ├── 469.txt │ │ │ │ ├── 470.txt │ │ │ │ ├── 471.txt │ │ │ │ ├── 472.txt │ │ │ │ ├── 473.txt │ │ │ │ ├── 474.txt │ │ │ │ ├── 475.txt │ │ │ │ ├── 476.txt │ │ │ │ ├── 477.txt │ │ │ │ ├── 478.txt │ │ │ │ ├── 479.txt │ │ │ │ ├── 480.txt │ │ │ │ ├── 481.txt │ │ │ │ ├── 482.txt │ │ │ │ ├── 483.txt │ │ │ │ ├── 484.txt │ │ │ │ ├── 485.txt │ │ │ │ ├── 486.txt │ │ │ │ ├── 487.txt │ │ │ │ ├── 488.txt │ │ │ │ ├── 489.txt │ │ │ │ ├── 490.txt │ │ │ │ ├── 491.txt │ │ │ │ ├── 492.txt │ │ │ │ ├── 493.txt │ │ │ │ ├── 494.txt │ │ │ │ ├── 495.txt │ │ │ │ ├── 496.txt │ │ │ │ ├── 497.txt │ │ │ │ ├── 498.txt │ │ │ │ ├── 499.txt │ │ │ │ ├── 500.txt │ │ │ │ ├── 501.txt │ │ │ │ ├── 502.txt │ │ │ │ ├── 503.txt │ │ │ │ ├── 504.txt │ │ │ │ ├── 505.txt │ │ │ │ ├── 506.txt │ │ │ │ ├── 507.txt │ │ │ │ ├── 508.txt │ │ │ │ ├── 509.txt │ │ │ │ ├── 510.txt │ │ │ │ └── 511.txt │ │ │ └── tech │ │ │ │ ├── 001.txt │ │ │ │ ├── 002.txt │ │ │ │ ├── 003.txt │ │ │ │ ├── 004.txt │ │ │ │ ├── 005.txt │ │ │ │ ├── 006.txt │ │ │ │ ├── 007.txt │ │ │ │ ├── 008.txt │ │ │ │ ├── 009.txt │ │ │ │ ├── 010.txt │ │ │ │ ├── 011.txt │ │ │ │ ├── 012.txt │ │ │ │ ├── 013.txt │ │ │ │ ├── 014.txt │ │ │ │ ├── 015.txt │ │ │ │ ├── 016.txt │ │ │ │ ├── 017.txt │ │ │ │ ├── 018.txt │ │ │ │ ├── 019.txt │ │ │ │ ├── 020.txt │ │ │ │ ├── 021.txt │ │ │ │ ├── 022.txt │ │ │ │ ├── 023.txt │ │ │ │ ├── 024.txt │ │ │ │ ├── 025.txt │ │ │ │ ├── 026.txt │ │ │ │ ├── 027.txt │ │ │ │ ├── 028.txt │ │ │ │ ├── 029.txt │ │ │ │ ├── 030.txt │ │ │ │ ├── 031.txt │ │ │ │ ├── 032.txt │ │ │ │ ├── 033.txt │ │ │ │ ├── 034.txt │ │ │ │ ├── 035.txt │ │ │ │ ├── 036.txt │ │ │ │ ├── 037.txt │ │ │ │ ├── 038.txt │ │ │ │ ├── 039.txt │ │ │ │ ├── 040.txt │ │ │ │ ├── 041.txt │ │ │ │ ├── 042.txt │ │ │ │ ├── 043.txt │ │ │ │ ├── 044.txt │ │ │ │ ├── 045.txt │ │ │ │ ├── 046.txt │ │ │ │ ├── 047.txt │ │ │ │ ├── 048.txt │ │ │ │ ├── 049.txt │ │ │ │ ├── 050.txt │ │ │ │ ├── 051.txt │ │ │ │ ├── 052.txt │ │ │ │ ├── 053.txt │ │ │ │ ├── 054.txt │ │ │ │ ├── 055.txt │ │ │ │ ├── 056.txt │ │ │ │ ├── 057.txt │ │ │ │ ├── 058.txt │ │ │ │ ├── 059.txt │ │ │ │ ├── 060.txt │ │ │ │ ├── 061.txt │ │ │ │ ├── 062.txt │ │ │ │ ├── 063.txt │ │ │ │ ├── 064.txt │ │ │ │ ├── 065.txt │ │ │ │ ├── 066.txt │ │ │ │ ├── 067.txt │ │ │ │ ├── 068.txt │ │ │ │ ├── 069.txt │ │ │ │ ├── 070.txt │ │ │ │ ├── 071.txt │ │ │ │ ├── 072.txt │ │ │ │ ├── 073.txt │ │ │ │ ├── 074.txt │ │ │ │ ├── 075.txt │ │ │ │ ├── 076.txt │ │ │ │ ├── 077.txt │ │ │ │ ├── 078.txt │ │ │ │ ├── 079.txt │ │ │ │ ├── 080.txt │ │ │ │ ├── 081.txt │ │ │ │ ├── 082.txt │ │ │ │ ├── 083.txt │ │ │ │ ├── 084.txt │ │ │ │ ├── 085.txt │ │ │ │ ├── 086.txt │ │ │ │ ├── 087.txt │ │ │ │ ├── 088.txt │ │ │ │ ├── 089.txt │ │ │ │ ├── 090.txt │ │ │ │ ├── 091.txt │ │ │ │ ├── 092.txt │ │ │ │ ├── 093.txt │ │ │ │ ├── 094.txt │ │ │ │ ├── 095.txt │ │ │ │ ├── 096.txt │ │ │ │ ├── 097.txt │ │ │ │ ├── 098.txt │ │ │ │ ├── 099.txt │ │ │ │ ├── 100.txt │ │ │ │ ├── 101.txt │ │ │ │ ├── 102.txt │ │ │ │ ├── 103.txt │ │ │ │ ├── 104.txt │ │ │ │ ├── 105.txt │ │ │ │ ├── 106.txt │ │ │ │ ├── 107.txt │ │ │ │ ├── 108.txt │ │ │ │ ├── 109.txt │ │ │ │ ├── 110.txt │ │ │ │ ├── 111.txt │ │ │ │ ├── 112.txt │ │ │ │ ├── 113.txt │ │ │ │ ├── 114.txt │ │ │ │ ├── 115.txt │ │ │ │ ├── 116.txt │ │ │ │ ├── 117.txt │ │ │ │ ├── 118.txt │ │ │ │ ├── 119.txt │ │ │ │ ├── 120.txt │ │ │ │ ├── 121.txt │ │ │ │ ├── 122.txt │ │ │ │ ├── 123.txt │ │ │ │ ├── 124.txt │ │ │ │ ├── 125.txt │ │ │ │ ├── 126.txt │ │ │ │ ├── 127.txt │ │ │ │ ├── 128.txt │ │ │ │ ├── 129.txt │ │ │ │ ├── 130.txt │ │ │ │ ├── 131.txt │ │ │ │ ├── 132.txt │ │ │ │ ├── 133.txt │ │ │ │ ├── 134.txt │ │ │ │ ├── 135.txt │ │ │ │ ├── 136.txt │ │ │ │ ├── 137.txt │ │ │ │ ├── 138.txt │ │ │ │ ├── 139.txt │ │ │ │ ├── 140.txt │ │ │ │ ├── 141.txt │ │ │ │ ├── 142.txt │ │ │ │ ├── 143.txt │ │ │ │ ├── 144.txt │ │ │ │ ├── 145.txt │ │ │ │ ├── 146.txt │ │ │ │ ├── 147.txt │ │ │ │ ├── 148.txt │ │ │ │ ├── 149.txt │ │ │ │ ├── 150.txt │ │ │ │ ├── 151.txt │ │ │ │ ├── 152.txt │ │ │ │ ├── 153.txt │ │ │ │ ├── 154.txt │ │ │ │ ├── 155.txt │ │ │ │ ├── 156.txt │ │ │ │ ├── 157.txt │ │ │ │ ├── 158.txt │ │ │ │ ├── 159.txt │ │ │ │ ├── 160.txt │ │ │ │ ├── 161.txt │ │ │ │ ├── 162.txt │ │ │ │ ├── 163.txt │ │ │ │ ├── 164.txt │ │ │ │ ├── 165.txt │ │ │ │ ├── 166.txt │ │ │ │ ├── 167.txt │ │ │ │ ├── 168.txt │ │ │ │ ├── 169.txt │ │ │ │ ├── 170.txt │ │ │ │ ├── 171.txt │ │ │ │ ├── 172.txt │ │ │ │ ├── 173.txt │ │ │ │ ├── 174.txt │ │ │ │ ├── 175.txt │ │ │ │ ├── 176.txt │ │ │ │ ├── 177.txt │ │ │ │ ├── 178.txt │ │ │ │ ├── 179.txt │ │ │ │ ├── 180.txt │ │ │ │ ├── 181.txt │ │ │ │ ├── 182.txt │ │ │ │ ├── 183.txt │ │ │ │ ├── 184.txt │ │ │ │ ├── 185.txt │ │ │ │ ├── 186.txt │ │ │ │ ├── 187.txt │ │ │ │ ├── 188.txt │ │ │ │ ├── 189.txt │ │ │ │ ├── 190.txt │ │ │ │ ├── 191.txt │ │ │ │ ├── 192.txt │ │ │ │ ├── 193.txt │ │ │ │ ├── 194.txt │ │ │ │ ├── 195.txt │ │ │ │ ├── 196.txt │ │ │ │ ├── 197.txt │ │ │ │ ├── 198.txt │ │ │ │ ├── 199.txt │ │ │ │ ├── 200.txt │ │ │ │ ├── 201.txt │ │ │ │ ├── 202.txt │ │ │ │ ├── 203.txt │ │ │ │ ├── 204.txt │ │ │ │ ├── 205.txt │ │ │ │ ├── 206.txt │ │ │ │ ├── 207.txt │ │ │ │ ├── 208.txt │ │ │ │ ├── 209.txt │ │ │ │ ├── 210.txt │ │ │ │ ├── 211.txt │ │ │ │ ├── 212.txt │ │ │ │ ├── 213.txt │ │ │ │ ├── 214.txt │ │ │ │ ├── 215.txt │ │ │ │ ├── 216.txt │ │ │ │ ├── 217.txt │ │ │ │ ├── 218.txt │ │ │ │ ├── 219.txt │ │ │ │ ├── 220.txt │ │ │ │ ├── 221.txt │ │ │ │ ├── 222.txt │ │ │ │ ├── 223.txt │ │ │ │ ├── 224.txt │ │ │ │ ├── 225.txt │ │ │ │ ├── 226.txt │ │ │ │ ├── 227.txt │ │ │ │ ├── 228.txt │ │ │ │ ├── 229.txt │ │ │ │ ├── 230.txt │ │ │ │ ├── 231.txt │ │ │ │ ├── 232.txt │ │ │ │ ├── 233.txt │ │ │ │ ├── 234.txt │ │ │ │ ├── 235.txt │ │ │ │ ├── 236.txt │ │ │ │ ├── 237.txt │ │ │ │ ├── 238.txt │ │ │ │ ├── 239.txt │ │ │ │ ├── 240.txt │ │ │ │ ├── 241.txt │ │ │ │ ├── 242.txt │ │ │ │ ├── 243.txt │ │ │ │ ├── 244.txt │ │ │ │ ├── 245.txt │ │ │ │ ├── 246.txt │ │ │ │ ├── 247.txt │ │ │ │ ├── 248.txt │ │ │ │ ├── 249.txt │ │ │ │ ├── 250.txt │ │ │ │ ├── 251.txt │ │ │ │ ├── 252.txt │ │ │ │ ├── 253.txt │ │ │ │ ├── 254.txt │ │ │ │ ├── 255.txt │ │ │ │ ├── 256.txt │ │ │ │ ├── 257.txt │ │ │ │ ├── 258.txt │ │ │ │ ├── 259.txt │ │ │ │ ├── 260.txt │ │ │ │ ├── 261.txt │ │ │ │ ├── 262.txt │ │ │ │ ├── 263.txt │ │ │ │ ├── 264.txt │ │ │ │ ├── 265.txt │ │ │ │ ├── 266.txt │ │ │ │ ├── 267.txt │ │ │ │ ├── 268.txt │ │ │ │ ├── 269.txt │ │ │ │ ├── 270.txt │ │ │ │ ├── 271.txt │ │ │ │ ├── 272.txt │ │ │ │ ├── 273.txt │ │ │ │ ├── 274.txt │ │ │ │ ├── 275.txt │ │ │ │ ├── 276.txt │ │ │ │ ├── 277.txt │ │ │ │ ├── 278.txt │ │ │ │ ├── 279.txt │ │ │ │ ├── 280.txt │ │ │ │ ├── 281.txt │ │ │ │ ├── 282.txt │ │ │ │ ├── 283.txt │ │ │ │ ├── 284.txt │ │ │ │ ├── 285.txt │ │ │ │ ├── 286.txt │ │ │ │ ├── 287.txt │ │ │ │ ├── 288.txt │ │ │ │ ├── 289.txt │ │ │ │ ├── 290.txt │ │ │ │ ├── 291.txt │ │ │ │ ├── 292.txt │ │ │ │ ├── 293.txt │ │ │ │ ├── 294.txt │ │ │ │ ├── 295.txt │ │ │ │ ├── 296.txt │ │ │ │ ├── 297.txt │ │ │ │ ├── 298.txt │ │ │ │ ├── 299.txt │ │ │ │ ├── 300.txt │ │ │ │ ├── 301.txt │ │ │ │ ├── 302.txt │ │ │ │ ├── 303.txt │ │ │ │ ├── 304.txt │ │ │ │ ├── 305.txt │ │ │ │ ├── 306.txt │ │ │ │ ├── 307.txt │ │ │ │ ├── 308.txt │ │ │ │ ├── 309.txt │ │ │ │ ├── 310.txt │ │ │ │ ├── 311.txt │ │ │ │ ├── 312.txt │ │ │ │ ├── 313.txt │ │ │ │ ├── 314.txt │ │ │ │ ├── 315.txt │ │ │ │ ├── 316.txt │ │ │ │ ├── 317.txt │ │ │ │ ├── 318.txt │ │ │ │ ├── 319.txt │ │ │ │ ├── 320.txt │ │ │ │ ├── 321.txt │ │ │ │ ├── 322.txt │ │ │ │ ├── 323.txt │ │ │ │ ├── 324.txt │ │ │ │ ├── 325.txt │ │ │ │ ├── 326.txt │ │ │ │ ├── 327.txt │ │ │ │ ├── 328.txt │ │ │ │ ├── 329.txt │ │ │ │ ├── 330.txt │ │ │ │ ├── 331.txt │ │ │ │ ├── 332.txt │ │ │ │ ├── 333.txt │ │ │ │ ├── 334.txt │ │ │ │ ├── 335.txt │ │ │ │ ├── 336.txt │ │ │ │ ├── 337.txt │ │ │ │ ├── 338.txt │ │ │ │ ├── 339.txt │ │ │ │ ├── 340.txt │ │ │ │ ├── 341.txt │ │ │ │ ├── 342.txt │ │ │ │ ├── 343.txt │ │ │ │ ├── 344.txt │ │ │ │ ├── 345.txt │ │ │ │ ├── 346.txt │ │ │ │ ├── 347.txt │ │ │ │ ├── 348.txt │ │ │ │ ├── 349.txt │ │ │ │ ├── 350.txt │ │ │ │ ├── 351.txt │ │ │ │ ├── 352.txt │ │ │ │ ├── 353.txt │ │ │ │ ├── 354.txt │ │ │ │ ├── 355.txt │ │ │ │ ├── 356.txt │ │ │ │ ├── 357.txt │ │ │ │ ├── 358.txt │ │ │ │ ├── 359.txt │ │ │ │ ├── 360.txt │ │ │ │ ├── 361.txt │ │ │ │ ├── 362.txt │ │ │ │ ├── 363.txt │ │ │ │ ├── 364.txt │ │ │ │ ├── 365.txt │ │ │ │ ├── 366.txt │ │ │ │ ├── 367.txt │ │ │ │ ├── 368.txt │ │ │ │ ├── 369.txt │ │ │ │ ├── 370.txt │ │ │ │ ├── 371.txt │ │ │ │ ├── 372.txt │ │ │ │ ├── 373.txt │ │ │ │ ├── 374.txt │ │ │ │ ├── 375.txt │ │ │ │ ├── 376.txt │ │ │ │ ├── 377.txt │ │ │ │ ├── 378.txt │ │ │ │ ├── 379.txt │ │ │ │ ├── 380.txt │ │ │ │ ├── 381.txt │ │ │ │ ├── 382.txt │ │ │ │ ├── 383.txt │ │ │ │ ├── 384.txt │ │ │ │ ├── 385.txt │ │ │ │ ├── 386.txt │ │ │ │ ├── 387.txt │ │ │ │ ├── 388.txt │ │ │ │ ├── 389.txt │ │ │ │ ├── 390.txt │ │ │ │ ├── 391.txt │ │ │ │ ├── 392.txt │ │ │ │ ├── 393.txt │ │ │ │ ├── 394.txt │ │ │ │ ├── 395.txt │ │ │ │ ├── 396.txt │ │ │ │ ├── 397.txt │ │ │ │ ├── 398.txt │ │ │ │ ├── 399.txt │ │ │ │ ├── 400.txt │ │ │ │ └── 401.txt │ │ └── ner_test │ │ │ ├── dev.txt │ │ │ ├── test.txt │ │ │ └── train.txt │ ├── run_scripts │ │ ├── __init__.py │ │ └── find_ideal_query.py │ └── simulated_active_learning.py ├── annotator_tool │ ├── __init__.py │ ├── active_learner.py │ ├── active_learner_async.py │ ├── al4nlp_adaptor │ │ ├── __init__.py │ │ └── al4nlp_adaptor.py │ ├── annotation_converter.py │ ├── annotation_converter_bio.py │ ├── annotator_widget.py │ ├── libact_adaptor │ │ ├── __init__.py │ │ ├── libact_adaptor.py │ │ └── utils.py │ ├── path_selector_widget.py │ ├── ui_widget.py │ └── visualizers │ │ ├── __init__.py │ │ ├── image.py │ │ ├── seq_annotation.py │ │ └── textarea.py ├── post_processing │ ├── __init__.py │ ├── calculate_tracin_score.py │ ├── pipeline_plasm.py │ └── plasm.py └── visualize │ ├── __init__.py │ ├── extract_metrics.py │ ├── plot_figure.py │ └── visualize.py ├── distillation ├── README.md ├── __init__.py ├── distil.sh ├── distil_distributed.sh ├── distiller.py ├── grouped_batch_sampler.py ├── lm_seqs_dataset.py ├── prepare_data.sh ├── requirements.txt ├── scripts │ ├── binarized_data.py │ ├── extract.py │ ├── extract_distilbert.py │ └── token_counts.py ├── token_count.sh ├── train.py ├── training_configs │ ├── distilbert-base-cased.json │ ├── distilbert-base-multilingual-cased.json │ ├── distilbert-base-uncased.json │ ├── distilelectra.json │ ├── distilgpt2.json │ └── distilroberta-base.json └── utils.py ├── docker ├── Dockerfile ├── Dockerfile_cuda113 ├── README.md ├── build.sh ├── build_cuda113.sh ├── cache.sh ├── fix-permissions ├── initial-condarc ├── jupyter_notebook_config.json ├── jupyter_server_config.py ├── start-notebook.sh ├── start-singleuser.sh └── start.sh ├── docs ├── Makefile ├── Readme.md ├── _static │ └── figures │ │ └── gui.svg ├── basic_usage.rst ├── conf.py ├── config_structure.rst ├── datasets.rst ├── gui.rst ├── index.rst ├── install.rst ├── make.bat ├── models.rst ├── new_strategies_addition.rst ├── new_subsampling_addition.rst ├── post_processing.rst ├── post_processing_usage.rst ├── query_strategies.rst ├── quick_start.rst ├── subsampling_strategies.rst └── testing.rst ├── domain_adaptation ├── README.md ├── __init__.py ├── configs │ ├── da_bert.yaml │ ├── da_bert_large.yaml │ ├── da_roberta.yaml │ ├── da_roberta_large.yaml │ ├── da_sentbert.yaml │ ├── domain_adaptation.yaml │ └── train.yaml ├── hf_dataset_to_sent.py ├── monitor_loss.ipynb ├── run_lm.py └── train.sh ├── examples ├── al │ ├── config.yaml │ └── train.sh ├── benchmark_custom_model.ipynb ├── benchmark_custom_strategy.ipynb ├── cache_necessary_files │ └── run.sh ├── custom_models │ ├── pytorch_cls.py │ └── pytorch_ner.py └── custom_strategy │ ├── __pycache__ │ ├── least_confidence.cpython-38.pyc │ └── top_from_previous_iteration_subsampling.cpython-38.pyc │ ├── least_confidence.py │ └── top_from_previous_iteration_subsampling.py ├── gui.svg ├── hooks ├── README.md └── pre-push ├── init.sh ├── jupyterlab_demo ├── active_learning_system.py ├── cls_demo.ipynb ├── configs │ ├── al_cls.yaml │ ├── al_ner.yaml │ ├── cls_plasm.yaml │ ├── default.py │ ├── framework │ │ ├── allennlp.yaml │ │ ├── flair.yaml │ │ ├── pytorch.yaml │ │ └── transformers.yaml │ └── ner_plasm.yaml ├── data │ ├── .gitkeep │ ├── CLS │ │ └── ag_news │ │ │ ├── labeled.json │ │ │ ├── tags.json │ │ │ ├── test.json │ │ │ └── unlabeled.json │ └── NER │ │ └── conll2003 │ │ ├── labeled.json │ │ ├── tags.json │ │ ├── test.json │ │ └── unlabeled.json ├── demo_system.py ├── logs │ └── .gitkeep ├── ner_demo.ipynb ├── prepare_cls_dataset.ipynb ├── prepare_ner_dataset.ipynb └── utils_data.py ├── nltk_download.py ├── reports ├── coverage │ └── coverage-badge.svg └── junit │ ├── junit.xml │ └── tests-badge.svg ├── scripts ├── calculate_overlap_among_queries.py ├── find_optimal_hyperparameters.py ├── run_active_learning.py ├── run_active_learning_known_queries.py ├── run_full_data.py └── run_tasks_on_multiple_gpus.py ├── setup.py └── tests ├── __init__.py ├── full_tests.py ├── test_al.py └── test_parameters.py /acleto/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/constructors/__init__.py: -------------------------------------------------------------------------------- 1 | from .construct_active_learner import construct_active_learner 2 | from .construct_wrapper import construct_wrapper 3 | -------------------------------------------------------------------------------- /acleto/al4nlp/model_wrappers/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/model_wrappers/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/model_wrappers/flair/__init__.py: -------------------------------------------------------------------------------- 1 | from .wrapper_flair import FlairModelWrapper 2 | -------------------------------------------------------------------------------- /acleto/al4nlp/model_wrappers/pytorch/__init__.py: -------------------------------------------------------------------------------- 1 | from .wrapper_cls_pytorch import PytorchClsWrapper 2 | from .wrapper_ner_pytorch import PytorchNerWrapper 3 | -------------------------------------------------------------------------------- /acleto/al4nlp/model_wrappers/pytorch/wrapper_cls_pytorch.py: -------------------------------------------------------------------------------- 1 | import logging 2 | 3 | from .wrapper_encoder_pytorch import PytorchEncoderWrapper 4 | from ..transformers.wrapper_cls import WrapperCls 5 | from ...models import PYTORCH_INIT_MODELS_DICT 6 | 7 | log = logging.getLogger() 8 | 9 | 10 | class PytorchClsWrapper(PytorchEncoderWrapper, WrapperCls): 11 | def __init__(self, **kwargs): 12 | super().__init__(**kwargs) 13 | 14 | def get_model_class(self): 15 | try: 16 | return PYTORCH_INIT_MODELS_DICT[self.model_config.checkpoint]["model_class"] 17 | except: 18 | model_class, _ = self.get_model_from_path() 19 | return model_class 20 | -------------------------------------------------------------------------------- /acleto/al4nlp/model_wrappers/pytorch/wrapper_ner_pytorch.py: -------------------------------------------------------------------------------- 1 | import logging 2 | 3 | from ..pytorch.wrapper_encoder_pytorch import PytorchEncoderWrapper 4 | from ..transformers.wrapper_ner import WrapperNer 5 | from ...models import PYTORCH_INIT_MODELS_DICT 6 | 7 | log = logging.getLogger() 8 | 9 | 10 | class PytorchNerWrapper(PytorchEncoderWrapper, WrapperNer): 11 | def __init__(self, **kwargs): 12 | super().__init__(**kwargs) 13 | 14 | def get_model_class(self): 15 | try: 16 | return PYTORCH_INIT_MODELS_DICT[self.model_config.checkpoint]["model_class"] 17 | except: 18 | model_class, _ = self.get_model_from_path() 19 | return model_class 20 | -------------------------------------------------------------------------------- /acleto/al4nlp/model_wrappers/transformers/__init__.py: -------------------------------------------------------------------------------- 1 | from .wrapper_cls import WrapperCls 2 | from .wrapper_ner import WrapperNer 3 | from .wrapper_ats import WrapperAts 4 | from .wrapper_nmt import WrapperNmt 5 | -------------------------------------------------------------------------------- /acleto/al4nlp/models/__init__.py: -------------------------------------------------------------------------------- 1 | from .fnet import init_fnet 2 | from .ner_bilstm import init_tagger, BilstmTagger 3 | from .text_classification_cnn import ( 4 | init_text_classification_cnn, 5 | TextClassificationCNN, 6 | ) 7 | 8 | INIT_MODELS_DICT = {"fnet": init_fnet} 9 | PYTORCH_INIT_MODELS_DICT = { 10 | "cnn": { 11 | "model": init_text_classification_cnn, 12 | "model_class": TextClassificationCNN, 13 | }, 14 | "bilstm": {"model": init_tagger, "model_class": BilstmTagger}, 15 | } 16 | # add list of available models instead of dict to avoid flair import 17 | FLAIR_MODELS = ["bilstm-crf"] 18 | -------------------------------------------------------------------------------- /acleto/al4nlp/models/spectral_normalization/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/models/spectral_normalization/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/pool_subsampling_strategies/__init__.py: -------------------------------------------------------------------------------- 1 | from .naive_subsampling import naive_subsampling 2 | from .random_subsampling import random_subsampling 3 | from .ups_subsampling import ups_subsampling 4 | -------------------------------------------------------------------------------- /acleto/al4nlp/pool_subsampling_strategies/naive_subsampling.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | 3 | 4 | def naive_subsampling(uncertainty_estimates, gamma_or_k_confident_to_save, **kwargs): 5 | if isinstance(gamma_or_k_confident_to_save, float): 6 | gamma_or_k_confident_to_save = int( 7 | gamma_or_k_confident_to_save * len(uncertainty_estimates) 8 | ) 9 | argsort = np.argsort(-uncertainty_estimates) 10 | return argsort[:gamma_or_k_confident_to_save] 11 | -------------------------------------------------------------------------------- /acleto/al4nlp/pool_subsampling_strategies/random_subsampling.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | 3 | 4 | def random_subsampling(uncertainty_estimates, gamma_or_k_confident_to_save, **kwargs): 5 | length = len(uncertainty_estimates) 6 | if isinstance(gamma_or_k_confident_to_save, float): 7 | gamma_or_k_confident_to_save = int(gamma_or_k_confident_to_save * length) 8 | if gamma_or_k_confident_to_save >= length: 9 | return np.arange(length) 10 | return np.random.choice( 11 | np.arange(length), gamma_or_k_confident_to_save, replace=False 12 | ) 13 | -------------------------------------------------------------------------------- /acleto/al4nlp/pool_subsampling_strategies/ups_subsampling.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | 3 | 4 | def get_ups_sampling_probas(argsort, gamma, T): 5 | ranks = argsort.argsort() / len(argsort) 6 | return np.exp(-np.maximum(0, ranks - gamma) / np.maximum(T, 1e-8)) 7 | 8 | 9 | def sample_idxs(sampling_probas): 10 | to_select = [] 11 | for i in range(len(sampling_probas)): 12 | proba_to_choose_i = sampling_probas[i] 13 | if np.random.uniform() < proba_to_choose_i: 14 | to_select.append(i) 15 | return np.array(to_select) 16 | 17 | 18 | def ups_subsampling(uncertainty_estimates, gamma_or_k_confident_to_save, T): 19 | if isinstance(gamma_or_k_confident_to_save, int): 20 | gamma_or_k_confident_to_save /= len(uncertainty_estimates) 21 | argsort = np.argsort(-uncertainty_estimates) 22 | sampling_probas = get_ups_sampling_probas(argsort, gamma_or_k_confident_to_save, T) 23 | return sample_idxs(sampling_probas) 24 | -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/query_strategies/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/entropy.py: -------------------------------------------------------------------------------- 1 | from typing import Union 2 | 3 | import numpy as np 4 | from datasets.arrow_dataset import Dataset 5 | 6 | from .al_strategy_utils import take_idx 7 | from ..utils.transformers_dataset import TransformersDataset 8 | 9 | 10 | def entropy( 11 | model, 12 | X_pool: Union[np.ndarray, Dataset, TransformersDataset], 13 | n_instances: int, 14 | **kwargs, 15 | ): 16 | """ 17 | Selects instances with the largest prediction entropy. 18 | https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.28.9963&rep=rep1&type=pdf 19 | """ 20 | probas = model.predict_proba(X_pool) 21 | uncertainty_estimates = np.sum(-probas * np.log(probas), axis=1) 22 | argsort = np.argsort(-uncertainty_estimates) 23 | query_idx = argsort[:n_instances] 24 | query = X_pool.select(query_idx) 25 | 26 | return query_idx, query, uncertainty_estimates 27 | -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/lc.py: -------------------------------------------------------------------------------- 1 | from typing import Union 2 | 3 | import numpy as np 4 | from datasets.arrow_dataset import Dataset 5 | 6 | from ..utils.transformers_dataset import TransformersDataset 7 | 8 | 9 | def lc( 10 | model, 11 | X_pool: Union[np.ndarray, Dataset, TransformersDataset], 12 | n_instances: int, 13 | **kwargs, 14 | ): 15 | """ 16 | Selects instances with the least prediction confidence (regarding the most likely class) 17 | https://arxiv.org/abs/cmp-lg/9407020. 18 | """ 19 | probas = model.predict_proba(X_pool) 20 | uncertainty_estimates = 1 - np.max(probas, axis=1) 21 | argsort = np.argsort(-uncertainty_estimates) 22 | query_idx = argsort[:n_instances] 23 | query = X_pool.select(query_idx) 24 | 25 | return query_idx, query, uncertainty_estimates 26 | -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/strategy_utils/__init__.py: -------------------------------------------------------------------------------- 1 | from .gmm import compute_density, get_gmm_log_probs, gmm_fit 2 | from .mahalanobis import ( 3 | compute_centroids, 4 | compute_inv_covariance, 5 | mahalanobis_distance_with_known_centroids_sigma_inv, 6 | ) 7 | from .mahalanobis_v2 import ( 8 | compute_inv_covariance_v2, 9 | mahalanobis_distance_with_known_centroids_sigma_inv_v2, 10 | ) 11 | from .batchbald.batchbald import get_batchbald_batch 12 | from .batchbald.consistent_dropout import make_dropouts_consistent 13 | from .ue import UeEstimatorHybrid, TextClassifier 14 | -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/strategy_utils/batchbald/README.md: -------------------------------------------------------------------------------- 1 | ##### Implementation taken from https://github.com/BlackHC/batchbald_redux 2 | 3 | @misc{kirsch2019batchbald, 4 | title={BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learning}, 5 | author={Andreas Kirsch and Joost van Amersfoort and Yarin Gal}, 6 | year={2019}, 7 | eprint={1906.08158}, 8 | archivePrefix={arXiv}, 9 | primaryClass={cs.LG} 10 | } -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/strategy_utils/batchbald/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/query_strategies/strategy_utils/batchbald/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/strategy_utils/ue/__init__.py: -------------------------------------------------------------------------------- 1 | from .text_classifier import TextClassifier 2 | from .ue_estimator_hybrid import UeEstimatorHybrid 3 | -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/strategy_wrappers/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/query_strategies/strategy_wrappers/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/query_strategies/strategy_wrappers/modal_sampling.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | from typing import Union 3 | from datasets.arrow_dataset import Dataset 4 | 5 | from ..al_strategy_utils import take_idx 6 | 7 | from ...utils.transformers_dataset import TransformersDataset 8 | 9 | 10 | def modal_sampling( 11 | model, 12 | X_pool: Union[Dataset, TransformersDataset], 13 | n_instances: int, 14 | modal_strategy, 15 | **kwargs, 16 | ): 17 | from modAL.uncertainty import ( 18 | entropy_sampling, 19 | uncertainty_sampling, 20 | margin_sampling, 21 | ) 22 | 23 | modal_strategy_dict = { 24 | "lc": uncertainty_sampling, 25 | "ent": entropy_sampling, 26 | "margin": margin_sampling, 27 | } 28 | 29 | query_idx = modal_strategy_dict[modal_strategy]( 30 | model, X_pool, n_instances=n_instances, **kwargs 31 | ) 32 | 33 | uncertainty_estimates = np.zeros(len(X_pool)) 34 | query = take_idx(X_pool, query_idx) 35 | 36 | return query_idx, query, uncertainty_estimates 37 | -------------------------------------------------------------------------------- /acleto/al4nlp/utils/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/utils/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/utils/data/__init__.py: -------------------------------------------------------------------------------- 1 | from .load_data import load_data 2 | -------------------------------------------------------------------------------- /acleto/al4nlp/utils/init_model_and_tokenizer.py: -------------------------------------------------------------------------------- 1 | def get_classifier_dropout_kwargs( 2 | pretrained_model_name: str, classifier_dropout: float 3 | ): 4 | if "distilbert" in pretrained_model_name: 5 | key = "seq_classif_dropout" 6 | elif "deberta" in pretrained_model_name: 7 | key = "pooler_dropout" 8 | elif "xlnet" in pretrained_model_name: 9 | key = "summary_last_dropout" 10 | elif "distilrubert" in pretrained_model_name: 11 | key = "dropout" 12 | elif "rubert-base" in pretrained_model_name: 13 | key = "hidden_dropout_prob" 14 | else: 15 | key = "classifier_dropout" 16 | return {key: classifier_dropout} 17 | 18 | 19 | def get_tokenizer_kwargs(pretrained_model_name: str, task: str): 20 | if task == "ner" and "roberta" in pretrained_model_name: 21 | return dict(add_prefix_space=True) 22 | return {} 23 | -------------------------------------------------------------------------------- /acleto/al4nlp/utils/packages/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al4nlp/utils/packages/__init__.py -------------------------------------------------------------------------------- /acleto/al4nlp/utils/time_decorator.py: -------------------------------------------------------------------------------- 1 | from time import time 2 | 3 | 4 | def time_decorator(func): 5 | def run_func_and_calculate_time(*args, **kwargs): 6 | start_time = time() 7 | func(*args, **kwargs) 8 | work_time = time() - start_time 9 | 10 | return run_func_and_calculate_time 11 | -------------------------------------------------------------------------------- /acleto/al_benchmark/__init__.py: -------------------------------------------------------------------------------- 1 | from . import simulated_active_learning 2 | -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/al_strategy/ddu.yaml: -------------------------------------------------------------------------------- 1 | strategy: ddu 2 | kwargs: 3 | use_activation: False 4 | use_spectralnorm: True -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/al_strategy/lc.yaml: -------------------------------------------------------------------------------- 1 | strategy: lc 2 | kwargs: -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/al_strategy/logits_lc.yaml: -------------------------------------------------------------------------------- 1 | strategy: logits_lc 2 | kwargs: -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/al_strategy/mahalanobis.yaml: -------------------------------------------------------------------------------- 1 | strategy: mahalanobis 2 | kwargs: 3 | use_da: False 4 | use_v2: False 5 | use_activation: False 6 | use_spectralnorm: True -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/al_strategy/mnlp.yaml: -------------------------------------------------------------------------------- 1 | strategy: mnlp 2 | kwargs: -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/al_strategy/nuq.yaml: -------------------------------------------------------------------------------- 1 | strategy: nuq 2 | kwargs: 3 | use_activation: False 4 | use_spectralnorm: True -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/al_strategy/random.yaml: -------------------------------------------------------------------------------- 1 | strategy: random 2 | kwargs: -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/exps_al.yaml: -------------------------------------------------------------------------------- 1 | cuda_devices: ??? 2 | task_names: '' 3 | hydra: 4 | run: 5 | dir: ./workdir/logs/${now:%Y-%m-%d}/${now:%H-%M-%S} 6 | 7 | tasks: 8 | - 9 | name: 'electra_mnlp' 10 | config_path: './configs/al_ner.yaml' 11 | command: 'run_active_learning_autoner.py' 12 | environ: '' 13 | args: 'al.strategy=mnlp model.checkpoint=google/electra-base-discriminator model.acquisition_model.name=google/electra-base-discriminator' 14 | seeds: [13, 42, 17] 15 | n_repeats: 1 16 | - 17 | name: 'electra_random' 18 | config_path: './configs/al_ner.yaml' 19 | command: 'run_active_learning_autoner.py' 20 | environ: 'al.strategy=random model.checkpoint=google/electra-base-discriminator model.acquisition_model.name=google/electra-base-discriminator' 21 | args: '' 22 | seeds: [13, 42, 17] 23 | n_repeats: 1 24 | -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/flask_cfg.yaml: -------------------------------------------------------------------------------- 1 | host: '0.0.0.0' 2 | port: '5000' 3 | ls_host: '0.0.0.0' 4 | ls_port: '8080' 5 | ls_token: 'cdd3fad9f2b89055c06ee79900a661eac1f25567' 6 | project_id: '15' -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/ue/beam.yaml: -------------------------------------------------------------------------------- 1 | num_beams: 10 2 | num_return_sequences: 10 3 | mc_iterations: 1 -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/ue/ensemble.yaml: -------------------------------------------------------------------------------- 1 | mc_iterations: 10 -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/ue/mc.yaml: -------------------------------------------------------------------------------- 1 | mc_iterations: 10 2 | seed: 42 3 | to_eval_mode: False -------------------------------------------------------------------------------- /acleto/al_benchmark/configs/ue/sampling.yaml: -------------------------------------------------------------------------------- 1 | mc_iterations: 10 2 | generate_top_p: 0.95 3 | top_k: 20 4 | early_stopping: False 5 | to_eval_mode: True 6 | do_sample: True -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/README.TXT: -------------------------------------------------------------------------------- 1 | Consists of 2225 documents from the BBC news website corresponding to stories in five topical areas from 2004-2005. 2 | Natural Classes: 5 (business, entertainment, politics, sport, tech) 3 | 4 | If you make use of the dataset, please consider citing the publication: 5 | - D. Greene and P. Cunningham. "Practical Solutions to the Problem of Diagonal Dominance in Kernel Document Clustering", Proc. ICML 2006. 6 | 7 | All rights, including copyright, in the content of the original articles are owned by the BBC. 8 | 9 | Contact Derek Greene for further information. 10 | http://mlg.ucd.ie/datasets/bbc.html 11 | 12 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/018.txt: -------------------------------------------------------------------------------- 1 | India's rupee hits five-year high 2 | 3 | India's rupee has hit a five-year high after Standard & Poor's (S&P) raised the country's foreign currency rating. 4 | 5 | The rupee climbed to 43.305 per US dollar on Thursday, up from a close of 43.41. The currency has gained almost 1% in the past three sessions. S&P, which rates borrowers' creditworthiness, lifted India's rating by one notch to 'BB+'. With Indian assets now seen as less of a gamble, more cash is expected to flow into its markets, buoying the rupee. 6 | 7 | "The upgrade is positive and basically people will use it as an excuse to come back to India," said Bhanu Baweja, a strategist at UBS. "Money has moved out from India in the first two or three weeks of January into other markets like Korea and Thailand and this upgrade should lead to a reversal." India's foreign currency rating is now one notch below investment grade, which starts at 'BBB-'. The increase has put it on the same level as Romania, Egypt and El Salvador, and one level below Russia. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/038.txt: -------------------------------------------------------------------------------- 1 | Deutsche Telekom sees mobile gain 2 | 3 | German telecoms firm Deutsche Telekom saw strong fourth quarter profits on the back of upbeat US mobile earnings and better-than-expected asset sales. 4 | 5 | Net profit came in at 1.4bn euros (£960m; $1.85bn), a dramatic change from the loss of 364m euros in 2003. Sales rose 2.8% to 14.96bn euros. Sales of stakes in firms including Russia's OAO Mobile Telesystems raised 1.17bn euros. This was more than expected and helped to bring debt down to 35.8bn euros. 6 | 7 | A year ago, debt was more than 11bn euros higher. T-Mobile USA, the company's American mobile business, made a strong contribution to profits. "It's a seminal achievement that they cut debt so low. That gives them some head room to invest in growth now," said Hannes Wittig, telecoms analyst at Dresdner Kleinwort Wasserstein. The company also said it would resume paying a dividend, after two years in which it focused on cutting debt. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/060.txt: -------------------------------------------------------------------------------- 1 | Insurance bosses plead guilty 2 | 3 | Another three US insurance executives have pleaded guilty to fraud charges stemming from an ongoing investigation into industry malpractice. 4 | 5 | Two executives from American International Group (AIG) and one from Marsh & McLennan were the latest. The investigation by New York attorney general Eliot Spitzer has now obtained nine guilty pleas. The highest ranking executive pleading guilty on Tuesday was former Marsh senior vice president Joshua Bewlay. 6 | 7 | He admitted one felony count of scheming to defraud and faces up to four years in prison. A Marsh spokeswoman said Mr Bewlay was no longer with the company. Mr Spitzer's investigation of the US insurance industry looked at whether companies rigged bids and fixed prices. Last month Marsh agreed to pay $850m (£415m) to settle a lawsuit filed by Mr Spitzer, but under the settlement it "neither admits nor denies the allegations". 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/096.txt: -------------------------------------------------------------------------------- 1 | Burren awarded Egyptian contracts 2 | 3 | British energy firm Burren Energy has been awarded two potentially lucrative oil exploration contracts in Egypt. 4 | 5 | The company successfully bid for the two contracts, granted by government owned oil firms, covering onshore and offshore areas in the Gulf of Suez. Burren Energy already has a presence in Egypt, having been awarded an exploration contract last year. The firm, which floated in 2003, recently announced a deal to buy 26% of Indian firm Hindustan Oil Exploration. The £13.8m deal gives Burren Energy access to the Indian oil and gas industry. This latest contract expands Burren Energy's global exploration and production portfolio - it also holds contracts in Turkmenistan and the Republic of Congo. "These assets significantly increase our exploration portfolio in Egypt and we continue to investigate further opportunities in this region," said chief executive Finian O'Sullivan. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/109.txt: -------------------------------------------------------------------------------- 1 | Lufthansa flies back to profit 2 | 3 | German airline Lufthansa has returned to profit in 2004 after posting huge losses in 2003. 4 | 5 | In a preliminary report, the airline announced net profits of 400m euros ($527.61m; £274.73m), compared with a loss of 984m euros in 2003. Operating profits were at 380m euros, ten times more than in 2003. Lufthansa was hit in 2003 by tough competition and a dip in demand following the Iraq war and the killer SARS virus. It was also hit by troubles at its US catering business. Last year, Lufthansa showed signs of recovery even as some European and US airlines were teetering on the brink of bankruptcy. The board of Lufthansa has recommended paying a 2004 dividend of 0.30 euros per share. In 2003, shareholders did not get a dividend. The company said that it will give all the details of its 2004 results on 23 March. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/124.txt: -------------------------------------------------------------------------------- 1 | Bargain calls widen Softbank loss 2 | 3 | Japanese communications firm Softbank has widened losses after heavy spending on a new cut-rate phone service. 4 | 5 | The service, launched in December and dubbed "Otoku" or "bargain", has had almost 900,000 orders, Softbank said. The firm, a market leader in high-speed internet, had an operating loss for the three months to December of 7.5bn yen ($71.5m; £38.4m). But without the Otoku marketing spend it would have made a profit - and expects to move into the black in 2006. The firm did not give a figure for the extent of profits it expected to make next year. It was born in the 1990s tech boom, investing widely and becoming a fast-rising star, till the end of the tech bubble hit it hard. Its recent return to a high profile came with the purchase of Japan Telecom, the country's third-biggest fixed-line telecoms firm. The acquisition spurred its broadband internet division to pole position in the Japanese market, with more than 5.1 million subscribers at the end of December. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/129.txt: -------------------------------------------------------------------------------- 1 | Iraq to invite phone licence bids 2 | 3 | Iraq is to invite bids for two telephone licences, saying it wants to significantly boost nationwide coverage over the next decade. 4 | 5 | Bids have been invited from local, Arab and foreign companies, Iraq's Ministry of Communications said. The winner will work in partnership with the Iraqi Telecommunications and Post Company (ITPC). The firms will install and operate a fixed phone network, providing voice, fax and internet services. 6 | 7 | The ministry said that it wanted to increase Iraq's "very low telephone service penetration rate from about 4.5% today to about 25% within 10 years." It also hopes to develop a "highly visible and changeable telecommunication sector". Details of the bidding and tender process will be published on the ministry's website on 9 February. It also is planning a road-show for investors in Amman, Jordan. The ministry said it would base its selection on criteria including the speed of implementation, tariff rates, coverage, and the firm's experience and financial strength. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/144.txt: -------------------------------------------------------------------------------- 1 | Oil prices fall back from highs 2 | 3 | Oil prices retreated from four-month highs in early trading on Tuesday after producers' cartel Opec said it was now unlikely to cut production. 4 | 5 | Following the comments by acting Opec secretary general Adnan Shihab-Eldin, US light crude fell 32 cents to $51.43 a barrel. He said that high oil prices meant Opec was unlikely to stick to its plan to cut output in the second quarter. In London, Brent crude fell 32 cents to $49.74 a barrel. 6 | 7 | Opec members are next meeting to discuss production levels on 16 March. On Monday, oil prices rose for a sixth straight session, reaching a four-month high as cold weather in the US threatened stocks of heating oil. US demand for heating oil was predicted to be about 14% above normal this week, while stocks were currently about 7.5% below the levels of a year ago. Cold weather across Europe has also put upward pressure on crude prices. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/158.txt: -------------------------------------------------------------------------------- 1 | Qatar and Shell in $6bn gas deal 2 | 3 | Shell has signed a $6bn (£3.12bn) deal with the Middle Eastern sheikhdom of Qatar to supply liquid natural gas (LNG) to North America and Europe. 4 | 5 | The UK-Dutch group will own 30% of the project, with Qatar's state oil firm owning the rest. The agreement is the latest in a string of deals reached by Qatar, which is trying to make itself a regional leader in natural gas. US oil giant ExxonMobil signed up for a $12.8bn deal earlier on Sunday. France's Total is expected to join the ExxonMobil scheme, dubbed Qatargas-2, on Monday, taking 5 million tonnes of LNG a year. ExxonMobil will be taking some 15 million tonnes each year for 25 years from the end of 2007 under the deal. Shell's agreement, under the name Qatargas-4, foresees the building of new facilities to handle 1.4 billion cubic feet of gas, and 7.8 million tonnes of LNG each year from 2011 onwards. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/233.txt: -------------------------------------------------------------------------------- 1 | Bat spit drug firm goes to market 2 | 3 | A German firm whose main product is derived from the saliva of the vampire bat is looking to raise more than 70m euros ($91m; £49m) on the stock market. 4 | 5 | The firm, Paion, said that it hoped to sell 5 million shares - a third of the firm - for 11-14 euros a share. Its main drug, desmoteplase, is based on a protein in the bat's saliva. The protein stops blood from clotting - which helps the bat to drink from its victims, but could also be used to help stroke sufferers. The company's shares go on sale later this week, and are scheduled to start trading on the Frankfurt Stock Exchange on 10 February. If the final price is at the top of the range, the company could be valued at as much as 200m euros. The money raised will be spent largely on developing the company's other drugs, since desmoteplase has already been licensed to one manufacturer, Forest Laboratories. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/234.txt: -------------------------------------------------------------------------------- 1 | Vodafone appoints new Japan boss 2 | 3 | Vodafone has drafted in its UK chief executive William Morrow to take charge of its troubled Japanese operation. 4 | 5 | Mr Morrow will succeed Shiro Tsuda as president of Vodafone KK, Japan's number three mobile operator, in April. Mr Tsuda, who will become chairman, was appointed president only two months ago but the business has struggled since then, losing customers in January. Vodafone had pinned its hopes on the launch of its 3G phones in November but demand for them has been slow. 6 | 7 | While it has more than 15 million customers in Japan, Vodafone has found it difficult to satisfy Japan's technologically demanding mobile users. It suffered a net loss of more than 58,000 customers in January, its second monthly reverse in the last year. "Vodafone is going to need to put a lot of money into Japan if it wants to rebuild the business," Tetsuro Tsusaka, a telecoms analyst with Deutsche Bank, told Reuters. "I do not know if it will be worth it for them to spend that kind of money just for Japan." 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/276.txt: -------------------------------------------------------------------------------- 1 | India power shares jump on debut 2 | 3 | Shares in India's largest power producer, National Thermal Power Corp (NTPC) have risen 13% on their stock market debut. 4 | 5 | The government's partial sell-off of NTPC is part of a controversial programme to privatise state-run firms. The 865 million share offer, a mix of new shares and sales by the government, raised 54bn rupees($1.2bn). It was India's second $1bn stock debut in three months, coming after the flotation by software firm Tata. The share offer was eleven times oversubscribed. "It is a good investment bet," said Suhas Naik, an investment analyst from ING Mutual Fund. "Power needs in India are set to rise and NTPC will benefit from that." Analysts say the success of the NTPC flotation would encourage the government to reduce stakes in more power companies. NTPC has said it will use the money from the share sale to feed the growing needs of the country's energy-starved economy. The firm is the largest utility company in India, and the sixth largest power producer in the world. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/283.txt: -------------------------------------------------------------------------------- 1 | Stock market eyes Japan recovery 2 | 3 | Japanese shares have ended the year at their highest level since 13 July amidst hopes of an economic recovery during 2005. 4 | 5 | The Nikkei index of leading shares gained 7.6% during the year to close at 11,488.76 points. In 2005 it "will rise toward 13,000", predicted Morgan Stanley equity strategist Naoki Kamiyama. The optimism in the financial markets contrast sharply with pessimism in the Japanese business community. Earlier this month, the quarterly Tankan survey of Japanese manufacturers found that business confidence had weakened for the first time since March 2003. 6 | 7 | Slower economic growth, rising oil prices, a stronger yen and weaker exports were blamed for the fall in confidence. Despite this, traders expect strength in the global economy to benefit Japan, which has been close to sliding into recession in recent months. Structural reform within Japan and an anticipated end to the banking sector's bad debt problems should also help, they say. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/321.txt: -------------------------------------------------------------------------------- 1 | Diageo to buy US wine firm 2 | 3 | Diageo, the world's biggest spirits company, has agreed to buy Californian wine company Chalone for $260m (£134m) in an all-cash deal. 4 | 5 | Although Diageo's best-known brands include Smirnoff vodka and Guinness stout, it already has a US winemaking arm - Diageo Chateau & Estate Wines. Diageo said it expects to get US regulatory approval for the deal during the first quarter of 2005. It said Chalone would be integrated into its existing US wine business. 6 | 7 | "The US wine market represents a growth opportunity for Diageo, with favourable demographic and consumption trends," said Diageo North America president Ivan Menezes. In July, Diageo, which is listed on the London Stock Exchange, reported an annual turnover of £8.89bn, down from £9.28bn a year earlier. It blamed a weaker dollar for its lower turnover. In the year ending 31 December 2003, Chalone reported revenues of $69.4m. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/322.txt: -------------------------------------------------------------------------------- 1 | Tokyo says deflation 'controlled' 2 | 3 | The Japanese government has forecast that the country's economic growth will slow to 1.6% in the next fiscal year starting in April 2005. 4 | 5 | While it predicts this fall from the current 2.1% level, it said it was making progress on ending deflation. The figures were given by economics minister Heizo Takenaka who said the economy would grow by 2% in 2006/07. He said the consumer price index (CPI) would rise 0.1% in the next fiscal year, the first gain since 2000/01. "We are attempting to make real economic conditions better and to overcome deflation. I think we are on track," said Mr Takenaka. Deflation - or falling consumer prices - has plagued Japan for more than five years. To ease the problem the Bank of Japan has regularly flooded the money market with excess cash to keep short term interest rates at 0% in an attempt to spur economic activity. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/343.txt: -------------------------------------------------------------------------------- 1 | Air Jamaica back in state control 2 | 3 | The Jamaican government is regaining control of Air Jamaica in an bid to help the ailing company out of its financial difficulties. 4 | 5 | The firm has failed to make money since the state sold a majority stake to hotel tycoon Gordon Stewart in 1994. In common with many carriers, Air Jamaica, with debts of $560m (£291m), has been hit by high fuel costs and the impact of the 11 September attacks. The company will be restructured with the aim of finding a new buyer. "The administration is committed to a viable national airline that will serve as a major catalyst for our economy," said Finance Minister Omar Davies. The 35-year-old airline transports about 55% of all passengers to the island and its pilots are reportedly among the best paid in the industry, with senior members of staff earning in excess of $234,000 a year. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/352.txt: -------------------------------------------------------------------------------- 1 | Share boost for feud-hit Reliance 2 | 3 | The board of Indian conglomerate Reliance has agreed a share buy-back, to counter the effects of a power struggle in the controlling family. 4 | 5 | The buy-back is a victory for chairman Mukesh Ambani, whose idea it was. His brother Anil, the vice-chairman, said had not been consulted and that the buy-back was "completely inappropriate and unnecessary". The board hopes the move will reverse a 13% fall in Reliance's shares since the feud became public last month. The company has been fractious since founder Dhirubhai Ambani died in 2002, leaving no will. "Today's round has gone to [Mukesh], there is no doubt about it," said Nanik Rupani, president of the Indian Merchants Chamber, a Bombay-based traders' body. The company plans to buy back 52 million shares at 570 rupees (£6.80; $13) apiece, a premium of more than 10% to its current market price. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/372.txt: -------------------------------------------------------------------------------- 1 | Madagascar completes currency switch 2 | 3 | Madagascar has completed the replacement of its Malagasy franc with a new currency, the ariary. 4 | 5 | From Monday, all prices and contracts will have to be quoted in the ariary, which was trading at 1,893 to the US dollar. The Malagasy franc, which lost almost half its value in 2004, is no longer legal tender but will remain exchangeable at banks until 2009. The phasing out of the franc, begun in July 2003, was intended to distance the country from its past under French colonial rule and address the problem of the large amount of counterfeit francs in circulation. "It's above all a question of sovereignty," Reuters quoted a central bank official as saying. "It is symbolic of our independence from the old colonial ways. Since we left the French monetary zone in 1973 we should have our own currency with its own name." The ariary was the name of a pre-colonial currency in the Indian Ocean island state. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/377.txt: -------------------------------------------------------------------------------- 1 | VW considers opening Indian plant 2 | 3 | Volkswagen is considering building a car factory in India, but said it had yet to make a final decision. 4 | 5 | The German giant said it was studying the possibility of opening an assembly plant in the country, but that it remained only a "potential" idea. Its comments came after the industry minister of India's Andhra Pradesh state said a team of VW officials were due to visit to discuss the plans. B. Satyanarayana said he expected VW to co-sign a memorandum of agreement. Several foreign carmakers, including Hyundai, Toyota, Suzuki and Ford, already have Indian production facilities to meet demand for automobiles in Asia's fourth-largest economy. VW's proposed plant would be set up in the port city of Visakhapatnam on India's eastern coast. An Andhra Pradesh official added that VW had already approved a factory site measuring 250 acres. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/394.txt: -------------------------------------------------------------------------------- 1 | US Airways staff agree to pay cut 2 | 3 | A union representing 5,200 flight attendants at bankrupt US Airways have agreed to a new contract that cuts pay by nearly 10%. 4 | 5 | The deal will help the carrier, trying to survive by cutting costs by nearly $1bn (£530m) a year, save about $94m. More than two thirds of its 28,000 staff have now accepted wage cuts. But talks are still continuing with a union representing mechanics, baggage handlers and cleaners, which has so far failed to negotiate a new contract. The seventh largest carrier in the US sought bankruptcy protection for a second time in two years last September. It had been one of the quickest to deal with difficulties faced by the aviation industry after the 9/11 attacks in 2001. But it emerged from Chapter 11 bankruptcy in March 2003 to face competition from low-cost carriers and higher fuel costs. US Airways management has said it may need to start liquidating assets if it does not receive concessions from all staff by the middle of this month. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/396.txt: -------------------------------------------------------------------------------- 1 | Profits jump at China's top bank 2 | 3 | Industrial and Commercial Bank (ICBC), China's biggest lender, has seen an 18% jump in profits during 2004. 4 | 5 | The increase in earnings has allowed the firm to write off bad loans and pave the way for a state bailout and eventual stock-market listing. China is trying to clean up its banking system, which is weighed down by billions of dollars of unpaid loans. It has already pumped $45bn (£24bn) into two of its largest banks, and has identified ICBC as a recipient of aid. ICBC's profits were 74.7bn yuan ($9bn; £4.8bn) in 2004, the bank said in a statement. The percentage of non-performing loans dropped to 19.1%, down about 2 percentage points. ICBC was founded in 1984 and had total assets of 5.3 trillion yuan at the end of 2003. China committed to gradually opening up its banking sector when it joined the World Trade Organisation in 2002. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/408.txt: -------------------------------------------------------------------------------- 1 | South African car demand surges 2 | 3 | Car manufacturers with plants in South Africa, including BMW, General Motors, Toyota and Volkswagen, have seen a surge in demand during 2004. 4 | 5 | New vehicle sales jumped 22% to 449,603 from a year earlier, the National Association of Automobile Manufacturers of South Africa (NAAMSA) said. Strong economic growth and low interest rates have driven demand, and analysts expect the trend to continue. NAAMSA said it expects sales to top 500,000 in 2005. During 2004 "South Africa was one of the best performing markets internationally" for car sales, NAAMSA said. While domestic demand is set to continue to enjoy rapid growth, foreign sales could come under pressure, analysts said. The vehicle industry accounts for about 13% of South Africa's total exports. However, the world auto market has its problems and analysts warn that overcapacity and the strength of the rand could hit exports. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/418.txt: -------------------------------------------------------------------------------- 1 | US firm 'bids for Lacroix label' 2 | 3 | A US firm has said it is in final negotiations with luxury goods group LVMH to buy the loss-making Christian Lacroix haute-couture house. 4 | 5 | Paris-based LVMH has been selling non-core businesses and focusing on its most profitable labels including Moet & Chandon champagne and Louis Vuitton. Privately-held Falic Group bought two cosmetics brands, Hard Candy and Urban Decay, from LVMH in early 2003. The Florida company also own a chain of 90 duty free stores in the US. LVMH refused to comment on the reports. But one of the three brothers behind the Falic Group said the firm had also held talks with the designer Christian Lacroix, and wished to retain him. "We are buying his name," Simon Falic told the Reuters news agency. "We have plans to increase the exposure of the brand and increase the volume of business." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/431.txt: -------------------------------------------------------------------------------- 1 | Mitsubishi in Peugeot link talks 2 | 3 | Trouble-hit Mitsubishi Motors is in talks with French carmaker PSA Peugeot Citroen about a possible alliance. 4 | 5 | On Tuesday Mitsubishi, the only major Japanese car firm in the red, confirmed earlier reports of negotiations. But a spokesman refused to comment on speculation that Mitsubishi could end up building cars for PSA and perhaps its Japanese rival Nissan. Mitsubishi has been hit by a recall scandal and the withdrawal of support from shareholder DaimlerChrysler. The US-German firm, once a majority shareholder, decided last April to stop providing financial backing. Mitsubishi's sales have slid 41% in the past year, catalysed by the revelation that the company had systematically been hiding records of faults and then secretly repairing vehicles. Mitsubishi is due to unveil a recovery plan later in January. Analysts said that alliances with other carmakers would be a necessary part of whatever it came up with, not least because its own slow sales have left its manufacturing capacity under-used. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/433.txt: -------------------------------------------------------------------------------- 1 | Cairn Energy in Indian gas find 2 | 3 | Shares in Cairn Energy rose 3.8% to 1,088 pence on Tuesday after the UK firm announced a fresh gas discovery in northern India. 4 | 5 | The firm, which last year made a number of other new finds in the Rajasthan area, said the latest discovery could lead to large gas volumes. However, chief executive Bill Gammell cautioned that additional evalution was first needed at the site. Cairn has also been granted approval to extend its Rajasthan exploration area. This approval has come from the Indian government. 6 | 7 | A spokesman said the company's decision to carry out further investigations at the new find showed that it believed there was significant gas. But he added: "It's still too early to say what the extent of it is." Cairn's string of finds in Rajasthan last year saw it elevated to the FTSE 100 index of the UK's leading listed companies. The company had bought the rights to explore in the area from oil giant Shell. Mr Gammell is a former Scottish international rugby player. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/470.txt: -------------------------------------------------------------------------------- 1 | Saudi investor picks up the Savoy 2 | 3 | London's famous Savoy hotel has been sold to a group combining Saudi billionaire investor Prince Alwaleed bin Talal and a unit of HBOS bank. 4 | 5 | Financial details of the deal, which includes the nearby Simpson's in the Strand restaurant, were not disclosed. The seller - Irish-based property firm Quinlan Private - bought the Savoy along with the Berkeley, Claridge's and the Connaught for £750m last year. Prince Alwaleed's hotel investments include the luxury George V in Paris. He also has substantial stakes in Fairmont Hotels & Resorts, which will manage the Savoy and Simpson's in the Strand, and Four Seasons. Fairmont said it planned to invest $48m (£26m) in renovating parts of the Savoy including the River Room and suites with views over the River Thames. Work was expected to be completed by summer 2006, Fairmont said. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/473.txt: -------------------------------------------------------------------------------- 1 | Two Nigerian banks set to merge 2 | 3 | Nigerian banks United Bank of Africa and Standard Trust Bank have agreed plans to merge and create the biggest bank in West Africa. 4 | 5 | The deal is also in line with a 2004 directive from the Nigerian central bank that called for more consolidation in the nation's crowded banking sector. The merger was announced in a statement on Standard Trust's website on Tuesday, but no financial details were revealed. United Bank is the third biggest in Nigeria in terms of number of branches. Standard Trust is smaller but more profitable. "The boards of United Bank and Standard Trust, at separate meetings yesterday, approved arrangements to merge both institutions," Standard Trust said. Standard Trust is 100% Nigerian-owned, but United Bank has some foreign investors, including New York-based Global Depository Receipts (32.8%), and Banca Nazionale del Lavoro and Monte del Paschi di Siena, both from Italy, who each have a 2.4% stake. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/480.txt: -------------------------------------------------------------------------------- 1 | Wipro beats forecasts once again 2 | 3 | Wipro, India's third-biggest software firm, has reported a 60% rise in profit, topping market expectations. 4 | 5 | Net income in the last quarter was 4.3bn rupees ($98m; £52m), against 2.7bn a year earlier. Profit had been forecast to be 4.1bn rupees. Wipro offers services such as call centres to foreign clients and has worked for more than half of the companies on the Fortune 500 list. Wipro said demand was strong, allowing it to increase the prices it charged. 6 | 7 | "On the face of it, the results don't look very exciting," said Apurva Shah, an analyst at ASK-Raymond James. "But the guidance is positive and pricing going up is good news." Third-quarter sales rose 34% to 20.9bn rupees. One problem identified by Wipro was the high turnover of its staff. It said that 90% of employees at its business process outsourcing operations had had to be replaced. "We have to get that under control," said vice-chairman Vivek Paul. Wipro is majority owned by India's richest man Azim Premji. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/492.txt: -------------------------------------------------------------------------------- 1 | Parmalat bank barred from suing 2 | 3 | Bank of America has been banned from suing Parmalat, the food group which went bust in 2003 after an accounting scandal. 4 | 5 | The bank - along with investors, auditors and the group's managers - wants damages for being a victim of fraud at the hands of the Italian firm. But a judge has barred Bank of America and two auditors from the case. The bank, and Italaudit - formerly the Italian arm of auditor Grant Thornton - face lawsuits and possible prosecution. A second auditor, Deloitte & Touche, has also been banned from the case. Grant Thornton - now rid of the Italian unit at the centre of the case - is still being permitted to sue, as are Consob, Italy's stock market regulator, hundreds of small investors and Parmalat's new managers. Parmalat collapsed in December 2003 after it emerged that the 4bn euros ($5.2bn; £2.8bn) it supposedly held in a Bank of American offshore account did not in fact exist. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/business/510.txt: -------------------------------------------------------------------------------- 1 | Lacroix label bought by US firm 2 | 3 | Luxury goods group LVMH has sold its loss-making Christian Lacroix clothing label to a US investment group. 4 | 5 | The Paris-based firm has been shedding non-core businesses and focusing on its most profitable brands including Moet & Chandon champagne and Louis Vuitton. LVMH said the French designer's haute couture and ready-to-wear labels had been purchased by the Falic Group for an unspecified sum. The Falic Group bought two cosmetics labels from LVMH in 2003. The sale of the Lacroix label comes as many fashion houses are struggling to make money from their expensive haute couture ranges. The Florida-based Falic group, which also runs a chain of 90 duty free stores in the US, said it planned to expand the brand by opening new stores. Mr Lacroix said he planned to stay at the label he founded in 1987 although exact details are still to be confirmed. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/019.txt: -------------------------------------------------------------------------------- 1 | Slater to star in Broadway play 2 | 3 | Actor Christian Slater is stepping into the role of Tom in the Broadway revival of The Glass Menagerie. 4 | 5 | Slater, 35, is replacing actor Dallas Roberts in the Tennessee Williams drama, which opens next month. No reason was given for Roberts' departure. The role will be played by understudy Joey Collins until Slater joins the show. Slater won rave reviews for his recent performance in One Flew Over the Cuckoo's Nest in London's West End. 6 | 7 | He has also starred in a number of films, including Heathers, Robin Hood: Prince of Thieves and more recently Churchill: The Hollywood Years. Preview performances of The Glass Menagerie will begin at New York's Ethel Barrymore Theatre on Thursday. Philip Rinaldi, a spokesman for the show, said the play's 15 March opening date remains unchanged. The revival, directed by David Leveaux, will also star Jessica Lange as the domineering mother, Amanda Wingfield. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/060.txt: -------------------------------------------------------------------------------- 1 | Snicket tops US box office chart 2 | 3 | The film adaptation of Lemony Snicket novels has topped the North America box office chart, displacing Ocean's 12. 4 | 5 | A Series of Unfortunate Events, starring Jim Carrey, took $30.2m (£15.5m) in its debut weekend. Ocean's 12 fell to number two while new entry Spanglish entered the chart at number three, taking $9m (£4.6m). A Series of Unfortunate Events also stars Scottish comedian Billy Connelly, while Carrey takes on a number of different roles in the surreal film. The only other new entry in the top 10 was re-make, Flight of the Phoenix, at number eight, starring Dennis Quaid. Carrey entertained crowds at the UK premiere of Lemony Snicket in London on Friday. The actor grabbed co-star Meryl Streep and whisked her around Leicester Square. Based on the cult children's books by author Daniel Handler, the film follows three children who are orphaned when their parents die in a fire. They are taken in by their wicked uncle Count Olaf, played by Carrey, who schemes to get his hands on their fortune. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/065.txt: -------------------------------------------------------------------------------- 1 | Animation charms Japan box office 2 | 3 | Oscar-winning director Hayao Miyazaki's latest film has set a new Japanese box office record, with 1.5bn yen ($14.3m) in two days, according to reports. 4 | 5 | Howl's Moving Castle is the follow-up to Miyazaki's Spirited Away, which won best animation at last year's Oscars. It is based on the children's book by English writer Diana Wynne Jones. It has registered the highest opening weekend takings of any Japanese film in the country, according to trade publication Screen Daily. 6 | 7 | The film is about an 18-year-old girl who is trapped in an old woman's body after being put under a spell by a witch. Its two-day takings represented 1.1 million cinema admissions, Screen Daily said. The film's distributor Toho expects 40 million people to see it in total - almost one third of the country's population - it added. The film won the Golden Osella for outstanding technical contribution at this year's Venice Film Festival. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/083.txt: -------------------------------------------------------------------------------- 1 | Cage film's third week at US top 2 | 3 | Nicolas Cage movie National Treasure has topped the US and Canada box office for the third week in a row. 4 | 5 | National Treasure made $17.1m (£8.8m) in ticket sales from Friday to Sunday, according to studio estimates, taking its total to $110.2m (£56.7m). Comedy Christmas with the Kranks, starring Tim Allen and Jamie Lee Curtis, was in second place. The Polar Express entered in third place while Jude Law film Closer made its debut at number six. 6 | 7 | Oliver Stone's big-budget epic Alexander, starring Colin Farrell, followed last week's disappointing sixth-place opening with a slump to seventh place and takings of $4.7m (£2.4m). Critics have savaged the three-hour epic, which reportedly cost $150m (£77m) to make. National Treasure, which sees Cage's character Ben Gates chase a hidden fortune, has been made by Disney Studios. It is Cage's fourth collaboration with Jerry Bruckheimer, who is usually noted for his male-orientated action films. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/087.txt: -------------------------------------------------------------------------------- 1 | Global release for Japan hit film 2 | 3 | Oscar-winning animator Hayao Miyazaki's latest film, Howl's Moving Castle, is to be released in 50 countries around the world, its distributor has said. 4 | 5 | Howl's Moving Castle set a Japanese box office record last week, taking 1.5bn yen (£7.7m) in its first two days. Miyazaki won an Oscar earlier this year for Spirited Away, Japan's first Academy Award for nearly 50 years. Howl's Moving Castle is based on a children's fantasy novel by UK author Diana Wynne Jones. It tells the story of an 18-year-old woman who ages dramatically after falling under a witch's spell. She heads to a moving castle kept by Howl, a wizard, and searches for a way to become normal again. A spokesman for distributor Toho said the film "has received strong interest from domestic audiences and foreign media and viewers alike". "We have a good feeling about this film. We saw very good viewer response," he added. The film's worldwide release will start in South Korea on 24 December, and France on 12 January. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/089.txt: -------------------------------------------------------------------------------- 1 | Oscar nominee Dan O'Herlihy dies 2 | 3 | Irish actor Dan O'Herlihy, who was nominated for best actor at the 1955 Oscars, has died at the age of 85. 4 | 5 | O'Herlihy, whose Oscar nomination was for Luis Bunuel's The Adventures of Robinson Crusoe, died at his home in Malibu, California, a spokesman said. The actor began his film career in the 1940s, playing Macduff to Orson Welles' Macbeth in 1948, and was also a regular in on the Dublin stage. He later appeared in Robocop and its sequel and cult TV show Twin Peaks. He played the CEO of Omni Consumer Products in 1987's Robocop and Robocop 2 three years later, and was saw mill owner Andrew Packard in Twin Peaks, also in 1990. Despite his Oscar nomination, he had few other lead roles and became a familiar supporting actor on TV and in film. The year he was nominated, the Academy Award was won by Marlon Brando for On the Waterfront. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/092.txt: -------------------------------------------------------------------------------- 1 | Actor Foxx sees Globe nominations 2 | 3 | US actor Jamie Foxx has been given two nominations for Golden Globe awards, with Meryl Streep, Morgan Freeman and Cate Blanchett also up for prizes. 4 | 5 | The stars were shortlisted on Monday for supporting roles, with the main nominations still to come. Foxx has starred in Collateral and Ray. Clive Owen, David Carradine and Natalie Portman are also up for awards. The Golden Globes, Hollywood's second most prominent awards, are the first major nominations to be announced. Last year, The Lord Of the Rings: The Return Of the King was named best drama movie while Lost In Translation won best musical or comedy. Sean Penn, Charlize Theron, Tim Robbins and Renee Zellweger all won acting awards - mirroring the eventual Oscars outcome. The Golden Globes ceremony will take place on 16 January, with the Oscars following on 27 February. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/112.txt: -------------------------------------------------------------------------------- 1 | Elvis set to top UK singles chart 2 | 3 | Rock 'n' roll legend Elvis is set to top the UK singles chart on Sunday, 27 years after his death. 4 | 5 | The re-release of hit song Jailhouse Rock was out-selling X Factor winner Steve Brockstein's cover of Against All Odds by 2,000 copies on Tuesday. If the record does make the top spot, it will be Elvis' 19th UK number one. The last time he topped the charts was with the remix of the little-known song A Little Less Conversation, which was number one in June 2002. If Jailhouse Rock does reach number one on Sunday, it will be the 999th in the history of the UK pop charts. The song first topped the charts in 1958. Chart analysts say Elvis could score the 1000th number one as well. His record One Night will be released the following week, followed the week after by A Fool Such As I, as part of his record company SonyBMG's new Elvis campaign. It has called it "the most ambitious singles release campaign in the history of the UK record industry". 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/157.txt: -------------------------------------------------------------------------------- 1 | US charity anthem is re-released 2 | 3 | We Are The World, the American charity anthem inspired by the success of Band Aid, has been re-issued to raise money for Aids research and tsunami victims. 4 | 5 | More than 40 stars sang as group USA For Africa, including Lionel Richie, Diana Ross, Ray Charles, Bob Dylan, Michael Jackson and Bruce Springsteen. It topped the charts in the US and UK, raising millions of dollars for African famine relief. The re-release also marks the 20th anniversary of the original recording. It has been re-issued as part of a two-disc DVD set, which will also feature footage from the recording session of the track in January 1985. The single was originally released in the US on 7 March 1985 and sold 800,000 copies in its first week. It went on to win Grammys for song of the year and record of the year. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/177.txt: -------------------------------------------------------------------------------- 1 | Hillbillies singer Scoggins dies 2 | 3 | Country and Western musician Jerry Scoggins has died in Los Angeles at the age of 93, his family has said. 4 | 5 | Scoggins was best remembered for singing the theme tune to popular US TV show The Beverly Hillbillies. The Texan-born singer approached the producers of the programme with theme tune The Ballad of Jed Clampett for the pilot which was screened in 1962. The show, which told the story of a poor man striking oil and moving to Beverly Hills, ran until 1971. 6 | 7 | Scoggins' daugher Jane Kelly Misel said that her father never tired of the song and would sing it at least once a day. "He'd sing it at birthdays and anniversaries and variety shows. He never stopped performing it," she said. When a film version of The Beverly Hillbillies was made in 1993, Scoggins came out of retirement to perform the theme tune. Scoggins sang the lyrics while bluegrass stars Lester Flatt and Earl Scruggs played guitar and banjo. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/186.txt: -------------------------------------------------------------------------------- 1 | Johnny and Denise lose Passport 2 | 3 | Johnny Vaughan and Denise Van Outen's Saturday night entertainment show Passport to Paradise will not return to screens, the BBC has said. 4 | 5 | The ex-Big Breakfast presenters were recruited to host the BBC One family variety show last July. "There are currently no plans for another series," a spokeswoman said. She added the pair "brought a real warmth to Saturday night, but in the end we felt we had done enough with the format of the show". 6 | 7 | Passport to Paradise involved a combination of games and outside broadcasts with a high level of audience participation. The first instalment attracted more than 4.1 million viewers - but that had dropped to fewer than 2.7 million by the time it ended. The BBC spokeswoman said Graham Norton's Strictly Dance Fever would be a priority for 2005. "That's very much on the cards for next year, and we're concentrating at the moment on Strictly Come Dancing, which is doing phenomenally well," she said. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/200.txt: -------------------------------------------------------------------------------- 1 | Celebrities get to stay in jungle 2 | 3 | All four contestants still remain in I'm A Celebrity ... Get Me Out Of Here as no evictions were made on the television show on Saturday. 4 | 5 | Contestants Paul Burrell, Joe Pasquale, Janet Street-Porter and Fran Cosgrave were told by hosts Ant and Dec. Natalie Appleton's decision to quit the show last Monday had given them all a stay of execution, the group were told. Model Sophie Anderton was the last person to be voted off the ITV1 show, set in the Australian jungle. The four remaining stars will do a joint Bushtucker Trial on Sunday. 6 | 7 | Former All Saints singer Natalie Appleton,31, walked out of the show after learning she would face a fifth so-called Bushtucker Trial. The celebrities are chosen by the viewers to pass trials in order to win food for the rest of the camp. Appleton had endured a torrid time during the programme, including a well-publicised row with Sophie Anderton. And on 26 November singer Brian Harvey quit as a contestant after he had a blazing row with Janet Street-Porter. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/211.txt: -------------------------------------------------------------------------------- 1 | Brookside creator's Channel 4 bid 2 | 3 | The creator of defunct TV soap Brookside has written to the culture minister to offer to buy Channel 4. 4 | 5 | Phil Redmond, now chairman of Mersey TV, told Tessa Jowell he would run it with its current remit intact for the next 10 years. But media watchdog Ofcom has said the the commercially funded public service broadcaster will not be privatised. A spokesman for the Department for Culture, Media and Sport said there were no plans to sell the channel. 6 | 7 | He added that primary legislation would be required for the station to be sold off, which the government was not intending to introduce. Brookside was axed in 2003 after its ratings slumped from a peak of seven million to just 1.5 million. Redmond also brought teen soap Hollyoaks to Channel 4 and created Grange Hill, the school-based drama serial which was first broadcast on BBC One in 1978. He was awarded the CBE for services to drama earlier this year. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/222.txt: -------------------------------------------------------------------------------- 1 | School tribute for TV host Carson 2 | 3 | More than 1,000 people turned out on Sunday to pay tribute to late US TV presenter Johnny Carson in the Nebraska town where he grew up. 4 | 5 | Carson, who hosted the Tonight Show for 30 years, died on 23 January from respiratory disease emphysema. He lived in Norfolk, Nebraska from the age of eight until he joined the Navy, but returned regularly and donated $5m (£2.7m) to local causes. Old school friends were among the crowd at the school's Johnny Carson Theater. 6 | 7 | Carson, who was one of the best-loved TV personalities in the US, asked not to have a public memorial in Los Angeles, where he lived in later life. He began his showbusiness career in Norfolk, performing magic under the name The Great Carsoni from the age of 14. His donations included $600,000 to Norfolk High School in the 1980s to build a new performing arts centre. When Carson died, President Bush led the public tributes, saying the presenter "had a profound influence on American life and entertainment". 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/249.txt: -------------------------------------------------------------------------------- 1 | Soul sensation ready for awards 2 | 3 | South West teenage singing sensation, Joss Stone, has been nominated in three categories in Wednesday's Brit awards. 4 | 5 | The 17-year-old from a small east Devon village near Cullumpton, received nominations for Best Solo Female, Best Urban Act and Best Breakthrough Artist. Her second album Mind, Body & Soul reached number one in the UK charts last October and went straight into the US charts at number 11. Ms Stone is due to perform at the 25th award ceremony at London's Earls Court. The teenager also has Grammy nominations in the US, normally dominated by home-grown acts. Born Jocelyn Stoker, the Devon diva started her career in a BBC talent programme, and was then discovered at a New York audition by a US record executive, Steve Greenberg. The 17-year-old singer is hoping to tour in Japan, Australia and the US in 2005. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/274.txt: -------------------------------------------------------------------------------- 1 | TV presenter Deeley drops CD:UK 2 | 3 | Cat Deeley has resigned as host of ITV1's Saturday morning children's music show CD:UK after six years. 4 | 5 | The 28-year-old host who also presents Stars In Their Eyes said she wanted to concentrate on new television projects. "I've had a great few years, met some really incredible people and been given some wonderful opportunities," the presenter said. Deeley will host her last show on 5 March and guest presenters will stand in until a replacement is found. 6 | 7 | The long-running presenter said: "I would like to take this opportunity to thank the CD:UK team and all the viewers that have watched, supported the show and let me gate crash their Saturday morning. "Whilst obviously I am really sad to say goodbye, I'm immensely excited about the future." Deeley formerly co-hosted SM:tv, the children's show which airs before CD:UK on a Saturday morning, alongside hit presenters Ant and Dec. She won a Children's Bafta award for best presenter in December 2001 and also fronted BBC One's Fame Academy series. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/302.txt: -------------------------------------------------------------------------------- 1 | Actress Roberts takes spider role 2 | 3 | Actress Julia Roberts will play the part of a spider in a new film version of children's classic Charlotte's Web. 4 | 5 | She will voice Charlotte, who teams up with a girl to save their friend Wilbur the pig, in the story by EB White. The film - a mix of live action and animation - will be Roberts' first project since the birth of her twins, Hazel and Phinnaeus, two months ago. Oprah Winfrey will voice a goose, John Cleese will voice a sheep and Steve Buscemi a rat in the 2006 film. 6 | 7 | Ten-year-old Dakota Fanning will play Fern, the girl at the centre of the story, in the film to be directed by 13 Going on 30 film-maker Gary Winick. Filming is due to begin in Melbourne, Australia, later this month. Charlotte's Web has sold 45 million copies since it was published in 1952. An animated version was made in 1973 but this will be the first live action film. The actor who will voice Wilbur the pig has yet to be revealed. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/309.txt: -------------------------------------------------------------------------------- 1 | Double win for Sea Inside 2 | 3 | Spanish movie The Sea Inside has been named best picture and star Javier Bardem best actor at the Bangkok International Film Festival. 4 | 5 | Based on the true story of a paralysed man's battle to end his life, The Sea Inside was best foreign language film at the 2005 Golden Globe Awards. Annette Bening and Ana Geislerova shared the best actress award at the Bangkok festival. This year's event was scaled down following the Asian tsunami disaster. 6 | 7 | Falling Down and The Phantom of the Opera director Joel Schumacher won a lifetime achievement award while Hong Kong director Wong Kar-Wai was honoured for promoting Asian cinema. "It's a great honour, I want to share this honour with my Thai colleagues," Wong said. French writer Christophe Barratie and Korean writer Park Chan-Chook shared the best screenplay award for Les Choristes and Old Boy respectively. Now in its third year, the Bangkok International Film Festival is donating profits from ticket sales to the tsunami relief fund. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/310.txt: -------------------------------------------------------------------------------- 1 | BBC denies Blackadder TV comeback 2 | 3 | The BBC has said there are no plans in the pipeline for a new series of hit comedy Blackadder, which ended in 1989. 4 | 5 | Tony Robinson, who played the servant Baldrick, told ITV1's This Morning the show's star, Rowan Atkinson, was "more keen than he has been in the past". Robinson added he would "love" to do another series, each of which was set in a different era, ranging from the 15th century to World War I. But the BBC said on Thursday there were no plans for a comeback. 6 | 7 | In the final series all the main characters were killed off charging towards German lines after being ordered out of their trench. The poignant finale was later voted the best farewell episode of a TV series. A host of other UK actors, including Hugh Laurie, Stephen Fry and Miranda Richardson, also appeared in the show. Blackadder returned for a one-off special filmed to celebrate the arrival of the millennium in 1999. It was shown at the Millennium Dome in Greenwich before being screened on BSkyB. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/317.txt: -------------------------------------------------------------------------------- 1 | Patti Smith to host arts festival 2 | 3 | Rock star Patti Smith has been made artistic director of this year's Meltdown festival, to be held at London's South Bank Centre in June. 4 | 5 | The punk pioneer follows Morrissey, David Bowie and Nick Cave in directing 15 days of concerts and events. Smith has yet to decide the line-up for Meltdown, which begins on 11 June. "I want to touch on all aspects of our culture, perhaps with readings from Genet, and I have an idea for Jeremy Irons to read Proust," she said. 6 | 7 | The 58-year-old US singer of Because the Night told The Observer newspaper that the festival would be her response to these "material, exploitative and greedy times". "It is important we explore the new, but we should also salute the best art there is, aesthetically and spiritually," she said. Last year's Meltdown festival, directed by the ex-Smiths star Morrissey, included appearances by singer Jane Birkin, Nancy Sinatra and playwright Alan Bennett. He also persuaded punk band The New York Dolls to reform for a reunion show. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/318.txt: -------------------------------------------------------------------------------- 1 | Roxy Music on Isle of Wight bill 2 | 3 | Roxy Music will appear at June's Isle of Wight music festival, along with Morrissey, Supergrass and Idlewild. 4 | 5 | REM have already been confirmed as headliners for the three-day event, which takes place from 10-12 June. Original band members Bryan Ferry, Phil Manzerana, Andy Mackay and Paul Thompson will take to the stage on the Saturday night of the festival. The band are also working in the studio on new material, their first since the 1982 album Avalon. Roxy Music were formed in 1971 by Ferry, Mackay and Brian Eno, with Manzanera joining the following year. Their first hit was 1972's Virginia Plain, with other hits including Street Life, Love Is The Drug and the 1981 John Lennon cover, Jealous Guy. Roxy Music's back catalogue was re-released in 2000, leading to the band undertaking a world tour in 2001. Morrissey will also play on Saturday at the festival, while Supergrass and Idlewild have been confirmed for the Friday. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/343.txt: -------------------------------------------------------------------------------- 1 | Applegate's Charity show closes 2 | 3 | US musical Sweet Charity has cancelled its run on Broadway after poor ticket sales for its early shows. 4 | 5 | Star Christina Applegate had to pull out of pre-Broadway performances earlier this month with a broken foot. Producer Barry Weissler said he was "deeply proud" of the show, but said the decision to close it was "painful but fiscally responsible". Applegate, who starred in TV comedy Married With Children, had been hoping to make her Broadway debut in the show. The 33-year-old injured herself while performing in Chicago, and had been hoping to recover in time for its official New York opening on 21 April. She had received mixed reviews for performances in Minneapolis and Chicago. Previews of the $7.5m (£4m) show were due to begin on 4 April. Sweet Charity tells the story of Charity Hope Valentine, a dancer who always falls in love with the wrong man. It was first performed on Broadway in 1966 with Gwen Verdon in the title role, while Shirley MacLaine starred in the 1969 film version. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/entertainment/344.txt: -------------------------------------------------------------------------------- 1 | Ray Charles studio becomes museum 2 | 3 | A museum dedicated to the career of the late legendary singer Ray Charles is to open in his former recording studio in Los Angeles. 4 | 5 | His longtime publicist Jerry Digney said the museum would house "archive materials from recordings, to awards, to ephemera, to wardrobe". A tour bus used by Charles and his entourage over the years will also be on permanent display. It is hoped the museum will be ready for visitors in late 2007. 6 | 7 | Mr Digney said the recording studio and offices had been used by Charles for many years, and was where he recorded much of his last album, Genius Loves Company. It is hoped the museum will also house an education centre. The building had been declared a historic landmark by the city of Los Angeles just before Charles' death in June 2004 at the age of 73. Following his death, Charles won eight Grammy Awards, including album of the year for Genius Loves Company, a collection of duets. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/015.txt: -------------------------------------------------------------------------------- 1 | Talks held on Gibraltar's future 2 | 3 | Two days of talks on the future of Gibraltar begin at Jack Straw's country residence later on Wednesday. 4 | 5 | Officials at the two-day summit at the foreign secretary's official Kent house, Chevening, will plan a new forum on the Rock's future. In October, Mr Straw and his Spanish counterpart Miguel Moratinos agreed to establish a body that would give Gibraltarians a voice in their future. Most Gibraltarians said in a referendum they wanted to remain British. 6 | 7 | Gibraltar's Chief Minister Peter Caruana will represent the British citizens living on the Rock, while Britain's Europe Director Dominick Chilcott will represent the UK. Madrid is being represented by Spain's director general for Europe, Jose Maria Pons. The initiative follows Spain's socialist government's decision to put its long-standing sovereignty ambitions on hold. Gibraltarians rejected plans for the Rock's sovereignty to be shared between Britain and Spain in a referendum organised by Gibraltar government. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/017.txt: -------------------------------------------------------------------------------- 1 | Tsunami debt deal to be announced 2 | 3 | Chancellor Gordon Brown has said he hopes to announce a deal to suspend debt interest repayments by tsunami-hit nations later on Friday. 4 | 5 | The agreement by the G8 group of wealthy nations would save affected countries £3bn pounds a year, he said. The deal is thought to have been hammered out on Thursday night after Japan, one of the biggest creditor nations, finally signed up to it. Mr Brown first proposed the idea earlier this week. 6 | 7 | G8 ministers are also believed to have agreed to instruct the World Bank and the International Monetary Fund to complete a country by country analysis of the reconstruction problems faced by all states hit by the disaster. Mr Brown has been locked in talks with finance ministers of the G8, which Britain now chairs. Germany also proposed a freeze and Canada has begun its own moratorium. The expected deal comes as Foreign Secretary Jack Straw said the number of Britons dead or missing in the disaster have reached 440. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/123.txt: -------------------------------------------------------------------------------- 1 | MPs issued with Blackberry threat 2 | 3 | MPs will be thrown out of the Commons if they use Blackberries in the chamber Speaker Michael Martin has ruled. 4 | 5 | The £200 handheld computers can be used as a phone, pager or to send e-mails. The devices gained new prominence this week after Alastair Campbell used his to accidentally send an expletive-laden message to a Newsnight journalist. Mr Martin revealed some MPs had been using their Blackberries during debates and he also cautioned members against using hidden earpieces. 6 | 7 | The use of electronic devices in the Commons chamber has long been frowned on. The sound of a mobile phone or a pager can result in a strong rebuke from either the Speaker or his deputies. The Speaker chairs debates in the Commons and is charged with ensuring order in the chamber and enforcing rules and conventions of the House. He or she is always an MP chosen by colleagues who, once nominated, gives up all party political allegiances. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/145.txt: -------------------------------------------------------------------------------- 1 | Candidate resigns over BNP link 2 | 3 | A prospective candidate for the UK Independence Party (UKIP) has resigned after admitting a "brief attachment" to the British National Party(BNP). 4 | 5 | Nicholas Betts-Green, who had been selected to fight the Suffolk Coastal seat, quit after reports in a newspaper that he attended a BNP meeting. The former teacher confirmed he had attended the meeting but said that was the only contact he had with the group. Mr Betts-Green resigned after being questioned by the party's leadership. A UKIP spokesman said Mr Betts-Green's resignation followed disclosures in the East Anglian Daily Times last month about his attendance at a BNP meeting. "He did once attend a BNP meeting. He did not like what he saw and heard and will take no further part of it," the spokesman added. A meeting of Suffolk Coastal UKIP members is due to be held next week to discuss a replacement. Mr Betts-Green, of Woodbridge, Suffolk, has also resigned as UKIP's branch chairman. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/212.txt: -------------------------------------------------------------------------------- 1 | Tory expert denies defeatism 2 | 3 | The Conservatives' campaign director has denied a report claiming he warned Michael Howard the party could not win the next general election. 4 | 5 | The Times on Monday said Australian Lynton Crosby told the party leader to focus on trying to increase the Tories' Commons presence by 25 to 30 seats. But Mr Crosby said in a statement: "I have never had any such conversation... and I do not hold that view." Mr Howard later added there was not "one iota" of truth in the report. The strategist helped Australia's PM, John Howard, win four elections. Mr Howard appointed Mr Crosby as his elections chief last October. Mr Crosby's statement said: "The Conservative Party has been making an impact on the issues of lower tax and controlled immigration over the past week." It added: "The Labour Party will be wanting to do all they can to distract attention away from the issues that really matter to people." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/235.txt: -------------------------------------------------------------------------------- 1 | Lib Dems' new election PR chief 2 | 3 | The Lib Dems have appointed a senior figure from BT to be the party's new communications chief for their next general election effort. 4 | 5 | Sandy Walkington will now work with senior figures such as Matthew Taylor on completing the party manifesto. Party chief executive Lord Rennard said the appointment was a "significant strengthening of the Lib Dem team". Mr Walkington said he wanted the party to be ready for any "mischief" rivals or the media tried to throw at it. 6 | 7 | "My role will be to ensure this new public profile is effectively communicated at all levels," he said. "I also know the party will be put under scrutiny in the media and from the other parties as never before - and we will need to show ourselves ready and prepared to counter the mischief and misrepresentation that all too often comes from the party's opponents. "The party is already demonstrating on every issue that it is the effective opposition." Mr Walkington's new job title is director of general election communications. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/236.txt: -------------------------------------------------------------------------------- 1 | Jack Cunningham to stand down 2 | 3 | Veteran Labour MP and former Cabinet minister Jack Cunningham has said he will stand down at the next election. 4 | 5 | One of the few Blair-era ministers to serve under Jim Callaghan, he was given the agriculture portfolio when Labour regained power in 1997. Mr Cunningham went on to become Tony Blair's "cabinet enforcer". He has represented the constituency now known as Copeland since 1970. Mr Blair said he was a "huge figure" in Labour and a "valued, personal friend". 6 | 7 | During Labour's long period in opposition, Mr Cunningham held a number of shadow roles including foreign affairs, the environment and as trade spokesman. As agriculture minister he caused controversy when he decided to ban beef on the bone in the wake of fears over BSE. He quit the government in 1999 and in recent years has served as the chairman of the all-party committee on Lords reform and has been a loyal supporter of the government from the backbenches. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/237.txt: -------------------------------------------------------------------------------- 1 | Kilroy unveils immigration policy 2 | 3 | Ex-chatshow host Robert Kilroy-Silk has attacked UK policy on immigration saying Britain's open door approach is hitting low wage "indigenous" workers. 4 | 5 | The Veritas leader said the only people to benefit from immigrants from places like Poland were employers, landlords, members of the 'metropolitan elite'. The MEP said his party would only admit foreigners who were required because they had specific skills to offer. And he argued asylum cost £2bn a year for 14,000 successful applicants. 6 | 7 | Mr Kilroy-Silk said that worked out at £143,000 per successful asylum seeker. He said Veritas wanted to grant an amnesty for all those in Britain claiming asylum and who have children and deport everyone else. Britain should take its fair share of asylum seekers under the United Nations Convention on Human Rights, he argued. And Mr Kilroy-Silk said he wanted to spend an extra £500m a year to help provide for refugees abroad. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/249.txt: -------------------------------------------------------------------------------- 1 | No election TV debate, says Blair 2 | 3 | Tony Blair has said he will not take part in a TV debate with his political rivals ahead of the next election. 4 | 5 | "We answer this every election campaign and, for the reasons I have given before, the answer is no," he said at his monthly news conference. In October Tory leader Michael Howard said Mr Blair would be running scared if he refused calls to go head-to-head. In recent years the leader of the opposition has always called for a debate, although it has never happened. Before the 2001 election, plans for a debate between Mr Blair, William Hague and Charles Kennedy collapsed. In 1997 a debate between Mr Blair and John Major was also cancelled when a format could not be agreed. Televised debates have become the high point of the US presidential election campaigns. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/269.txt: -------------------------------------------------------------------------------- 1 | Teens 'know little' of politics 2 | 3 | Teenagers questioned for a survey have shown little interest in politics - and have little knowledge. 4 | 5 | Only a quarter of 14-16 year olds knew that Labour was the government, the Tories were the official Opposition and the Lib Dems were the third party. Almost all could identify Tony Blair, but only one in six knew who Michael Howard was, and just one in 10 recognised Charles Kennedy. The ICM survey interviewed 110 pupils for education watchdog Ofsted. Nearly half those pupils polled said it was not important for them to know more about what the political parties stand for. And 4% of those questioned thought the Conservatives were in power - while 2% of them believed the Lib Dems were. The survey also looked at issues of nationality. It found the Union flag and fish and chips topped the list of symbols and foods associated with being British. Many of the pupils also looked on themselves as English, Scottish or Welsh, rather than British; while the notion of being European hardly occurred to anyone. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/304.txt: -------------------------------------------------------------------------------- 1 | UKIP outspent Labour on EU poll 2 | 3 | The UK Independence Party outspent both Labour and the Liberal Democrats in the European elections, new figures show. 4 | 5 | UKIP, which campaigned on a slogan of "Say no to Europe", spent £2.36m on the campaign - second only to the Conservatives' £3.13m. The campaign took UKIP into third place with an extra 10 MEPs. Labour's campaign cost £1.7m, the Lib Dems' £1.19m and the Greens' £404,000, according to figures revealed by the Electoral Commission on Wednesday. Much of the UKIP funding came from Yorkshire millionaire Sir Paul Sykes, who helped bankroll the party's billboard campaign. Critics have accused the party of effectively buying votes. But a UKIP spokesman said Labour and the Conservatives had spent £10m between them on the last general election. "With the advantages of public money the others have, the only way the smaller parties can get their message across is by buying the advertising space," he added. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/310.txt: -------------------------------------------------------------------------------- 1 | MPs issued with Blackberry threat 2 | 3 | MPs will be thrown out of the Commons if they use Blackberries in the chamber Speaker Michael Martin has ruled. 4 | 5 | The £200 handheld computers can be used as a phone, pager or to send e-mails. The devices gained new prominence this week after Alastair Campbell used his to accidentally send an expletive-laden message to a Newsnight journalist. Mr Martin revealed some MPs had been using their Blackberries during debates and he also cautioned members against using hidden earpieces. 6 | 7 | The use of electronic devices in the Commons chamber has long been frowned on. The sound of a mobile phone or a pager can result in a strong rebuke from either the Speaker or his deputies. The Speaker chairs debates in the Commons and is charged with ensuring order in the chamber and enforcing rules and conventions of the House. He or she is always an MP chosen by colleagues who, once nominated, gives up all party political allegiances. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/329.txt: -------------------------------------------------------------------------------- 1 | Tory expert denies defeat warning 2 | 3 | The Conservatives' campaign director has denied a report claiming he warned Michael Howard the party could not win the next general election. 4 | 5 | The Times on Monday said Australian Lynton Crosby told the party leader to focus on trying to increase the Tories' Commons presence by 25 to 30 seats. But Mr Crosby said in a statement: "I have never had any such conversation... and I do not hold that view." Mr Howard later added there was not "one iota" of truth in the report. The strategist helped Australia's PM, John Howard, win four elections. Mr Howard appointed Mr Crosby as his elections chief last October. Mr Crosby's statement said: "The Conservative Party has been making an impact on the issues of lower tax and controlled immigration over the past week." It added: "The Labour Party will be wanting to do all they can to distract attention away from the issues that really matter to people." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/343.txt: -------------------------------------------------------------------------------- 1 | Labour's Cunningham to stand down 2 | 3 | Veteran Labour MP and former Cabinet minister Jack Cunningham has said he will stand down at the next election. 4 | 5 | One of the few Blair-era ministers to serve under Jim Callaghan, he was given the agriculture portfolio when Labour regained power in 1997. Mr Cunningham went on to become Tony Blair's "cabinet enforcer". He has represented the constituency now known as Copeland since 1970. Mr Blair said he was a "huge figure" in Labour and a "valued, personal friend". 6 | 7 | During Labour's long period in opposition, Mr Cunningham held a number of shadow roles including foreign affairs, the environment and as trade spokesman. As agriculture minister he caused controversy when he decided to ban beef on the bone in the wake of fears over BSE. He quit the government in 1999 and in recent years has served as the chairman of the all-party committee on Lords reform and has been a loyal supporter of the government from the backbenches. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/397.txt: -------------------------------------------------------------------------------- 1 | Blunkett hints at election call 2 | 3 | Ex-Home Secretary David Blunkett has given fresh clues that the general election will be announced on Monday. 4 | 5 | He told BBC Radio Five Live: "I'm out in my constituency getting ready for what we presume will be an announcement very shortly at the weekend." He clarified that he meant he would be in his Sheffield seat this weekend, not that he expected an election call then. Tony Blair is tipped to ask the Queen on Monday to dissolve Parliament ready for a 5 May poll. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/politics/398.txt: -------------------------------------------------------------------------------- 1 | Lib Dems predict 'best ever poll' 2 | 3 | The Lib Dems are set for their best results in both the general election and the local council polls, one of their frontbenchers has predicted. 4 | 5 | Local government spokesman Ed Davey was speaking as the party launched its campaign for the local elections being held in 37 English council areas. The flagship pledge is to replace council tax with a local income tax. The Tories say the Lib Dems would make people pay more tax and Labour says the party's sums do not add up. Looking to the coming elections, which are all expected to be held on 5 May, Mr Davey said: "We are going to be winning more votes and winning more seats. "I think we are going to have the best general election results and local election results we have ever had under [party leader] Charles Kennedy. "I couldn't think of a stronger endorsement of a leader." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/002.txt: -------------------------------------------------------------------------------- 1 | O'Sullivan could run in Worlds 2 | 3 | Sonia O'Sullivan has indicated that she would like to participate in next month's World Cross Country Championships in St Etienne. 4 | 5 | Athletics Ireland have hinted that the 35-year-old Cobh runner may be included in the official line-up for the event in France on 19-20 March. Provincial teams were selected after last Saturday's Nationals in Santry and will be officially announced this week. O'Sullivan is at present preparing for the London marathon on 17 April. The participation of O'Sullivan, currentily training at her base in Australia, would boost the Ireland team who won the bronze three years agio. The first three at Santry last Saturday, Jolene Byrne, Maria McCambridge and Fionnualla Britton, are automatic selections and will most likely form part of the long-course team. O'Sullivan will also take part in the Bupa Great Ireland Run on 9 April in Dublin. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/005.txt: -------------------------------------------------------------------------------- 1 | Dibaba breaks 5,000m world record 2 | 3 | Ethiopia's Tirunesh Dibaba set a new world record in winning the women's 5,000m at the Boston Indoor Games. 4 | 5 | Dibaba won in 14 minutes 32.93 seconds to erase the previous world indoor mark of 14:39.29 set by another Ethiopian, Berhane Adera, in Stuttgart last year. But compatriot Kenenisa Bekele's record hopes were dashed when he miscounted his laps in the men's 3,000m and staged his sprint finish a lap too soon. Ireland's Alistair Cragg won in 7:39.89 as Bekele battled to second in 7:41.42. "I didn't want to sit back and get out-kicked," said Cragg. "So I kept on the pace. The plan was to go with 500m to go no matter what, but when Bekele made the mistake that was it. The race was mine." Sweden's Carolina Kluft, the Olympic heptathlon champion, and Slovenia's Jolanda Ceplak had winning performances, too. Kluft took the long jump at 6.63m, while Ceplak easily won the women's 800m in 2:01.52. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/006.txt: -------------------------------------------------------------------------------- 1 | Isinbayeva claims new world best 2 | 3 | Pole vaulter Yelena Isinbayeva broke her own indoor world record by clearing 4.89 metres in Lievin on Saturday. 4 | 5 | It was the Russian's 12th world record of her career and came just a few days after she cleared 4.88m at the Norwich Union Grand Prix in Birmingham. The Olympic champion went on to attempt 5.05m at the meeting on France but failed to clear that height. In the men's 60m, former Olympic 100m champion Maurice Greene could only finish second to Leonard Scott. It was Greene's second consecutive defeat at the hands of his fellow American, who also won in Birmingham last week. "I ran my race perfectly," said Scott, who won in 6.46secs, his best time indoors. "I am happy even if I know that Maurice is a long way from being at his peak at the start of the season." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/007.txt: -------------------------------------------------------------------------------- 1 | O'Sullivan commits to Dublin race 2 | 3 | Sonia O'Sullivan will seek to regain her title at the Bupa Great Ireland Run on 9 April in Dublin. 4 | 5 | The 35-year-old was beaten into fourth at last year's event, having won it a year earlier. "I understand she's had a solid winter's training down in Australia after recovering from a minor injury," said race director Matthew Turnbull. Mark Carroll, Irish record holder at 3km, 5km and 10km, will make his debut in the mass participation 10km race. Carroll has stepped up his form in recent weeks and in late January scored an impressive 3,000m victory over leading American Alan Webb in Boston. Carroll will be facing stiff competition from Australian Craig Mottram, winner in Dublin for the last two years. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/015.txt: -------------------------------------------------------------------------------- 1 | UK Athletics agrees new kit deal 2 | 3 | UK Athletics has agreed a new deal with adidas to supply Great Britain squads of all ages with their kit for the next four years. 4 | 5 | The German-based firm kitted out Team GB at the 2004 Olympics and has deals with 20 other national Olympic bodies. UK Athletics chief David Moorcroft said: "The Athens experience can now be extended to more major championships. "In the year ahead these include the European indoor and World outdoor championships. We are delighted." Moorcroft added: "It is hugely beneficial to the sport that the adidas commitment will also provide for officials and other personnel at our world-class series of live televised events." This week, UK Athletics also agreed a four-year deal with energy drink company, Red Bull, who will be supplying the product to athletics at major domestic meetings and in high performance centres. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/019.txt: -------------------------------------------------------------------------------- 1 | London hope over Chepkemei 2 | 3 | London Marathon organisers are hoping that banned athlete Susan Chepkemei will still take part in this year's race on 17 April. 4 | 5 | Chepkemei was suspended from all competition until the end of the year by Athletics Kenya after failing to report to a national training camp. "We are watching it closely," said London race director David Bedford. "There is a long way to go before the race and we are hoping the situation will be satisfactorily resolved." The camp in Embu was to prepare for the IAAF World Cross Country Championships later this month. Chepkemei however took part and finished third in last Sunday's world best 10K race in Puerto Rico. The 29-year-old has finished second to Paula Radcliffe in the 2002 and 2003 London races as well as in November's New York City marathon. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/024.txt: -------------------------------------------------------------------------------- 1 | Chepkemei joins Edinburgh line-up 2 | 3 | Susan Chepkemei has decided she is fit enough to run in next month's Great Edinburgh International Cross Country. 4 | 5 | The Kenyan was initially unsure if she would have recovered from her gruelling tussle with Paula Radcliffe in the New York Marathon in time to compete. But she has declared herself up to the task and joins a field headed by World cross country champion Benita Johnson. Race director Matthew Turnbull said: "Susan will add even more strength in depth to the world-class line up." Chepkemei, who won the six kilometre event three years ago when it was staged in Newcastle, endured an epic battle with Radcliffe in the Big Apple until the Briton outsprinted her in the final 400m. Tirunesh Dibaba of Ethiopia will defend the title she won last year in Tyneside - before the race was moved north of the border. Recently-crowned European cross country champion Briton Hayley Yelling also competes in Edinburgh on 15 January, as does in-form Scot Kathy Butler. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/026.txt: -------------------------------------------------------------------------------- 1 | Collins appeals against drugs ban 2 | 3 | Sprinter Michelle Collins has lodged an appeal against her eight-year doping ban with the North American Court of Arbitration for Sport (CAS). 4 | 5 | The 33-year-old received the ban last month as a result of her connection to the federal inquiry into the Balco doping scandal. She is the first athlete to be banned without a positive drugs test or an admission of drugs use. CAS has said that a ruling is normally given within four months of an appeal. Collins was suspended by the US Anti-Doping Agency based on patterns observed in her blood and urine tests as well as evidence in the Balco investigation. As well as being hit with the ban, Collins was stripped of her 2003 world and US indoor 200m titles. The San Francisco-based Balco laboratory is at the centre of the scandal which has rocked the sport. The company has been accused of distributing illegal performance-enhancing drugs to elite athletes. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/027.txt: -------------------------------------------------------------------------------- 1 | Holmes feted with further honour 2 | 3 | Double Olympic champion Kelly Holmes has been voted European Athletics (EAA) woman athlete of 2004 in the governing body's annual poll. 4 | 5 | The Briton, made a dame in the New Year Honours List for taking 800m and 1,500m gold, won vital votes from the public, press and EAA member federations. She is only the second British woman to land the title after- Sally Gunnell won for her world 400m hurdles win in 1993. Swedish triple jumper Christian Olsson was voted male athlete of the year. The accolade is the latest in a long list of awards that Holmes has received since her success in Athens. In addition to becoming a dame, she was also named the BBC Sports Personality of the Year in December. Her gutsy victory in the 800m also earned her the International Association of Athletics Federations' award for the best women's performance in the world for 2004. And she scooped two awards at the British Athletics Writers' Association annual dinner in October. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/029.txt: -------------------------------------------------------------------------------- 1 | Isinbayeva heads for Birmingham 2 | 3 | Olympic pole vault champion Yelena Isinbayeva has confirmed she will take part in the 2005 Norwich Union Grand Prix in Birmingham on 18 February. 4 | 5 | "Everybody knows how much I enjoy competing in Britain. I always seem to break records there," said Isinbayeva. "As Olympic champion there will be more attention on me this year, but hopefully I can respond with another record in Birmingham." Kelly Holmes and Carolina Kluft are among other Athens winners competing. The organisers are hoping that Isinbayeva's main rival, fellow Russian Svetlana Feofanova, will also take part in the event. The pair had a thrilling battle in Athens which ended with Isinbayeva finally jumping a world record of 4.91m to claim the gold medal. Isinbayeva, 22, has set 10 world records in the pole vault, three of which have come on British soil. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/031.txt: -------------------------------------------------------------------------------- 1 | Sprinter Walker quits athletics 2 | 3 | Former European 200m champion Dougie Walker is to retire from athletics after a series of six operations left him struggling for fitness. 4 | 5 | Walker had hoped to compete in the New Year Sprint which is staged at Musselburgh Racecourse near Edinburgh on Tuesday and Wednesday. The 31-year-old Scot was suspended for two years in 1998 after testing positive for nandrolone. "I had intended to race but I'm running like a goon," said Walker. He told the Herald newspaper: "I'm not in great shape, after missing about a month of training. "I missed a big chunk of speed work over about three weeks, and then another week working in America. "If I'd had a half-decent mark it might have motivated me more, but I won't be racing. "I still enjoy training, but feel it's time to move on, and concentrate on a career." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/039.txt: -------------------------------------------------------------------------------- 1 | Jones medals 'must go if guilty' 2 | 3 | World Anti-Doping Agency (WADA) chief Dick Pound says Marion Jones should be stripped of all her medals if found guilty of taking banned substances. 4 | 5 | Victor Conte, of Balco Laboratories, claims the American sprinter regularly used drugs to enhance her performance. "If she is found guilty she should be stripped of all her medals and banned for two years," said Pound. Asked if there was a timescale as to what medals could be taken, Pound said: "That is not an issue at all." However, under International Olympic Committee (IOC) rules, athletes can only be stripped of their medals if caught within three years of the event. Jones, who won five medals at the 2000 Olympics, denies using drugs and says she will take legal action over Conte's allegations. Balco Laboratories is the firm at the centre of a wide-reaching investigation into doping in the US. Pound continued: "If she has indeed taken drugs it is going to be a big disappointment for a lot of people." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/052.txt: -------------------------------------------------------------------------------- 1 | Holmes starts 2005 with GB events 2 | 3 | Kelly Holmes will start 2005 with a series of races in Britain. 4 | 5 | Holmes will make her first track appearance on home soil since winning double Olympic gold in January's Norwich Union International in Glasgow. She will also run in the Grand Prix in Birmingham in February and may defend her indoor AAA 800m title in Sheffield earlier that month. "I am still competitive and still want to win," she said. "I'm an athlete and I can't wait to get back on the track." She added: "These events are also a great opportunity to thank the British public for the enormous levels of support they have given me from the moment I stepped off that plane from Greece." The Glasgow meeting will see Holmes compete over 1500m in a five-way match against Sweden, France, Russia and Italy. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/053.txt: -------------------------------------------------------------------------------- 1 | Gebrselassie in London triple bid 2 | 3 | Double Olympic 10,000m champion Haile Gebrselassie will race in the London Marathon for the next three years. 4 | 5 | The Ethiopian legend won Sunday's Almeria half-marathon in Spain on his return from an operation on his Achilles tendon. He was third in London in 2002 in his first serious attempt at the marathon. "It is a coup for us to secure Haile's presence for the next three years and it guarantees a quality race," said race director David Bedford. Gebrselassie will face Olympic champion Stefano Baldini, world champion Jaouad Gharib, and arch-rival Paul Tergat, the current world record holder. "If I didn't think I could win I would not be here," said Gebrselassie, who has set world records on 18 occasions in his illustrious career and is keen to add the marathon record to his collection. "There are a lot of fantastic runners in the race but I shall be doing my utmost to upset them." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/056.txt: -------------------------------------------------------------------------------- 1 | Kluft impressed by Sotherton form 2 | 3 | Olympic heptathlon champion Carolina Kluft was full of admiration for Britain's Kelly Sotherton as the pair prepared to clash in Birmingham. 4 | 5 | Both will be in action on Friday in the 60m hurdles and long jump ahead of the European Indoor Championships later this month in Madrid. Sotherton finished third behind the Swede in Athens, and Kluft said: "I knew about her, she's a great girl. "She looked very good early in the season and was competing really well." Kluft showed impressive early-season form on Tuesday in Stockholm's GE Galan meeting, winning the sprint hurdles, the long jump and the 400m. Sotherton has also displayed promise, with a new high jump personal best in Sheffield at the combined Norwich Union European trials and AAA Championships, and a second place in the long jump behind Jade Johnson. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/067.txt: -------------------------------------------------------------------------------- 1 | Lewis-Francis eyeing world gold 2 | 3 | Mark Lewis-Francis says his Olympic success has made him determined to bag World Championship 100m gold in 2005. 4 | 5 | The 22-year-old pipped Maurice Greene on the last leg of the 4x100m relay in Athens to take top honours for Team GB. But individually, the Birchfield Harrier has yet to build on his World Junior Championship win four years ago. "The gold medal in Athens has made me realise that I can get to the top level and I want to get there again. It can happen, I don't see why not," he said. Lewis-Francis has still to decided what events will feature in his build-up to the worlds - with one exception. He has confirmed his participation in the Norwich Union Grand Prix in Birmingham on 18 February, where he will take on another member of Britain's victorious men's relay team - Jason Gardener - over 60m. He added: "It's a bit too early to make any predictions for Helsinki, but I have my eyes open and I know I can be the best in the world." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/073.txt: -------------------------------------------------------------------------------- 1 | Johnson uncertain about Euro bid 2 | 3 | Jade Johnson is undecided about whether to contest next month's European Indoor Championships in Madrid despite winning the AAAs long jump title on Saturday. 4 | 5 | The 24-year-old delivered a personal best of 6.50m to win the European trials but had to wait until her final jump after four failures. "I don't want to go if I am not going to get a medal," said Johnson. "I will have to see how I am jumping in the next competition and I'll have to have a conversation with my coach." Johnson, who finished seventh in last year's Olympic Games, has not competed indoors since 2000. And the Commonwealth and European silver medallist believes her lack of experience in the early part of the season has knocked her confidence. "It's the stress," said Johnson. "I am not used to feeling this, this early. I am just used to training. "But if I'm doing this kind of thing, then I will have to see how it goes." Johnson next competes in the high-class Birmingham Grand Prix on 18 February. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/074.txt: -------------------------------------------------------------------------------- 1 | Pavey focuses on indoor success 2 | 3 | Jo Pavey will miss January's View From Great Edinburgh International Cross Country to focus on preparing for the European Indoor Championships in March. 4 | 5 | The 31-year-old was third behind Hayley Yelling and Justyna Bak in last week's European Cross Country Championships but she prefers to race on the track. "It was great winning bronze but I'm wary of injuries and must concentrate on the indoor season," she said. "Because of previous injuries I don't even run up hills in training." Pavey, who came fifth in the 5,000m at the Athens Olympics, helped the British cross country team win the team silver medal in Heringsdorf last week. She is likely to start her 3,000m season with a race in either Boston or Stuttgart at the end of January. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/077.txt: -------------------------------------------------------------------------------- 1 | Yelling takes Cardiff hat-trick 2 | 3 | European cross-country champion Hayley Yelling completed a hat-trick of wins in the Reebok Cardiff Cross Challenge in Bute Park on Sunday afternoon. 4 | 5 | The part-time maths teacher beat Irish international Jolene Byrne by 40 metres in the six-kilometre race. Another Great Britain international, Louise Damen, finished third as part of the contingent representing England. Peter Riley, who secured bronze for the GB men's team at last month's European Championships, won the men's 9km race. Riley, representing England, moved away over the last two kilometres to win by 25 metres from Ireland's Gary Murray. Glynn Tromans - the reigning UK Inter-Countries and England Cross-Country champion - came in third place as he continues his comeback from a five-month injury lay-off. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/078.txt: -------------------------------------------------------------------------------- 1 | Butler strikes gold in Spain 2 | 3 | Britain's Kathy Butler continued her impressive year with victory in Sunday's 25th Cross Internacional de Venta de Banos in Spain. 4 | 5 | The Scot, who led GB to World Cross Country bronze earlier this year, moved away from the field with Ines Monteiro halfway into the 6.6km race. She then shrugged off her Portuguese rival to win in 20 minutes 38 seconds. Meanwhile, Briton Karl Keska battled bravely to finish seventh in the men's 10.6km race in a time of 31:41. Kenenisa Bekele of Ethiopia - the reigning world long and short course champion - was never troubled by any of the opposition, winning leisurely in 30.26. Butler said of her success: "I felt great throughout the race and hope this is a good beginning for a marvellous 2005 season for me." Elsewhere, Abebe Dinkessa of Ethiopia won the Brussels IAAF cross-country race on Sunday, completing the 10,500m course in 33.22. Gelete Burka then crowned a great day for Ethiopia by claiming victory in the women's race. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/083.txt: -------------------------------------------------------------------------------- 1 | Johnson too strong for GB runners 2 | 3 | Britain's Kathy Butler and Hayley Yelling were no match for Benita Johnson in the 51st Cross International Zornotza in Amorebieta, Spain. 4 | 5 | Butler and Yelling finished fourth and fifth as Australian world champion Johnson romped to a five-second victory in the 6km race ahead of Edith Masai. Masai's fellow Kenyan Alice Timbilil finished third. Johnson said: "I ran comfortably for the first 3km and then I tried to leave the others but it wasn't an easy task." Butler clocked a time of 22 minutes 45 seconds - 22secs behind the winner but four ahead of Yelling, who last month succeeded Paula Radcliffe as European champion. Johnson, will be one of the star attractions at the Great EdinburghInternational Cross Country on 15 January. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/084.txt: -------------------------------------------------------------------------------- 1 | Holmes starts 2005 with GB events 2 | 3 | Kelly Holmes will start 2005 with a series of races in Britain. 4 | 5 | Holmes will make her first track appearance on home soil since winning double Olympic gold in January's Norwich Union International in Glasgow. She will also run in the Grand Prix in Birmingham in February and may defend her indoor AAA 800m title in Sheffield earlier that month. "I am still competitive and still want to win," she said. "I'm an athlete and I can't wait to get back on the track." She added: "These events are also a great opportunity to thank the British public for the enormous levels of support they have given me from the moment I stepped off that plane from Greece." The Glasgow meeting will see Holmes compete over 1500m in a five-way match against Sweden, France, Russia and Italy. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/085.txt: -------------------------------------------------------------------------------- 1 | Tulu to appear at Caledonian run 2 | 3 | Two-time Olympic 10,000 metres champion Derartu Tulu has confirmed she will take part in the BUPA Great Caledonian Run in Edinburgh on 8 May. 4 | 5 | The 32-year-old Ethiopian is the first star name to enter the event. Tulu has won the Boston, London and Tokyo Marathons, as well as the world 10,000m title in 2001. "We are delighted to have secured the services of one the most decorated competitors the sport has ever seen," said race director Matthew Turnbull. "Her record speaks for herself and there are few other women distance runners who would dare compare their pedigree with Tulu's," he added. "She might be 33 next month, but that didn't stop her winning the Olympic 10,000m bronze medal last summer. She's an ultra-consistent championships racer." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/086.txt: -------------------------------------------------------------------------------- 1 | Disappointed Scott in solid start 2 | 3 | Allan Scott is confident of winning a medal at next week's European Indoor Championships after a solid debut on the international circuit. 4 | 5 | The 22-year-old Scot finished fourth in the 60m hurdles at the Jose M Cagigal Memorial meeting in Madrid. "It was definitely a learning curve and I certainly haven't ruled out challenging for a medal next week," said the East Kilbride athlete. The race was won by Felipe Vivancos, who equalled the Spanish record. 6 | 7 | Sweden's Robert Kronberg was second, with Haiti's Dudley Dorival in third. Scott was slightly disappointed with his run in the final. He won his heat in 7.64secs but ran 0.04secs slower in his first IAAF Indoor Grand Prix circuit final. "I should have done better than that," he said. "I felt I could have won it. I got a poor start - but I still felt I should have ran faster." Vivancos slashed his personal best to equal the Spanish record with a time of 7.60secs while Kronberg and Dorival clocked 7.62secs and 7.63secs respectively. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/092.txt: -------------------------------------------------------------------------------- 1 | Collins calls for Chambers return 2 | 3 | World 100m champion Kim Collins says suspended sprinter Dwain Chambers should be allowed to compete in the Olympics again. 4 | 5 | Chambers was banned for two years after testing positive for the anabolic steroid THG and his suspension runs out in November this year. But Collins says the British Olympic Association should reverse the decision to ban him from the Olympics for life. "It was too harsh," Collins told Radio Five Live. "They should reconsider." Chambers has been in America learning American football but has not ruled out a return to the track. Collins added: "He is a great guy and I have never had any problems with him. We are friends. "I would like to see Dwain come back and compete again. He is a good person. "Even though he made a mistake he understands what he did and should be given a chance once more." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/098.txt: -------------------------------------------------------------------------------- 1 | Ronaldo considering new contract 2 | 3 | Manchester United winger Cristiano Ronaldo said he is close to agreeing to a new contract at Old Trafford. 4 | 5 | The Portugal star, who joined in August 2003 on a five-year-deal, is a regular in the United first-team. "The United board have already made an offer to renew the contract but I'm trying not to think about it," he told the News of the World. "My agent has spoken with the club and it will be resolved soon. I think we'll reach a good agreement for both sides." Ronaldo refused to commit his long-term future to the club. 6 | 7 | "Nobody knows what will come tomorrow. I like being here, but who knows," he added. "There aren't many bigger and better clubs than this one. It's my ambition to be at a big club. I'm happy but nobody knows the future." 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/103.txt: -------------------------------------------------------------------------------- 1 | Duff ruled out of Barcelona clash 2 | 3 | Chelsea's Damien Duff has been ruled out of Wednesday's Champions League clash with Barcelona at the Nou Camp. 4 | 5 | Duff sustained a knee injury in the FA Cup defeat at Newcastle and manager Jose Mourinho said: "He cannot run. His injury is very painful, so he is out." But Mourinho has revealed defender Willian Gallas and striker Didier Drogba will be in the starting line-up. The Blues boss took the unusual step of naming his side a day before the match, with Jole Cole named in midfield. Mourinho said: "We have one more session but I think Drogba will play, and Gallas will play. "Drogba trained on Monday with no problems and will do the same on Tuesday. Gallas feels he can play and wants to play. We are protecting him still but he will be okay to play." Drogba, Chelsea's £24m striker, has missed the last three weeks through injury. 6 | 7 | Cech, Ferreira, Carvalho, Terry, Gallas, Tiago, Makelele, Lampard, Cole, Drogba, Gudjohnsen. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/116.txt: -------------------------------------------------------------------------------- 1 | Bosvelt optimistic over new deal 2 | 3 | Manchester City's Paul Bosvelt will find out "within a month" whether he is to be offered a new one-year deal. 4 | 5 | The 34-year-old Dutch midfielder is out of contract in the summer and, although his age may count against him, he feels he can play on for another season. "I told the club I would like to stay for one more year. They promised me an answer within the next month so I am waiting to see," he said. "The main concern is my age but I think I have proved I am fit enough. Bosvelt joined City from Feyenoord in 2003 and at first he struggled to adapt to life in England. But his professionalism and dedication impressed manager Kevin Keegan. "He realised the pace of the game was faster than anything he was used to but he drove himself back into the team. He is an unsung hero," said Keegan. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/127.txt: -------------------------------------------------------------------------------- 1 | Kewell waits as injury drags on 2 | 3 | Liverpool winger Harry Kewell is struggling to recover from his Achilles tendinitis problem and may not recover until March, claims his agent. 4 | 5 | Kewell, 26, had hoped to play against Fulham last weekend and to be ready for the Champions League tie against Bayer Leverkusen on 22 February. But Bernie Mandic said: "He's had scans which don't show anything substantial. "But despite his best efforts at the moment he simply can't shake off the discomfort and get back on the park." 6 | 7 | Mandic continued in the Sydney Morning Herald: "Harry's struggling a bit but the club are doing everything they can to get him right. "Harry was desperate to be fit for the Fulham game as part of his plan to play himself in for the Leverkusen match." Kewell has not played since December 19 and misses out on international duty this week, with Australia facing South Africa in Durban on Wednesday. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/130.txt: -------------------------------------------------------------------------------- 1 | Weir poised for Sunderland move 2 | 3 | Larne's teenage star Robbie Weir is poised to join Sunderland after turning down a move to Stoke City. 4 | 5 | The 17-year-old Irish League midfielder was also being chased by Rangers and Fulham, but Mick McCarthy's side appear to have won the race. But Larne boss Jimmy McGeough has yet to confirm that Weir is on his way from Inver Park. "I heard on Sunday that he has joined Sunderland, but not from the lad himself," he said. ''Robbie has an agreement with Larne that he can negotiate with interested clubs. 6 | 7 | ''Personally, I would rather see him making an impact at local level to build up his physique before moving into the full-time game. ''He has been on trial with a number of clubs. But it would be great to see him making it at Sunderland.'' 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/135.txt: -------------------------------------------------------------------------------- 1 | Israeli club look to Africa 2 | 3 | Four African players, including Zimbabwe goalkeeper Energy Murambadoro, are all ready to play for Israeli club Hapoel Bnei Sakhnin in the Uefa Cup. 4 | 5 | Bnei Sakhnin are the first Arab side ever to play in European competition and will play English Premiership side Newcastle United in the first round. Warriors' goalkeeper Murambadoro, who made a name for himself at the African Nations Cup finals in Tunisia, helped Bnei Sakhnin overcome Albania's Partizani Tirana 6-1 in the previous round. Murambadoro moved to Israel recently after a brief stint with South African club Hellenic. The club won the Israeli Cup final last season and are based in Sakhnin, which is near Haifa. The club have a strong ethic and are high profile promoters of peace and co-operation within Israel. The three other Africans at the club are former Cameroon defender Ernest Etchi, DR Congo's Alain Masudi and Nigerian midfielder Edith Agoye, who had a stint with Tunisian side Esperance. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/136.txt: -------------------------------------------------------------------------------- 1 | Iranian misses Israel match 2 | 3 | Iranian striker Vahid Hashemian will not travel to Israel for Bayern Munich's Champions League match against Maccabi Tel Aviv on Wednesday. 4 | 5 | The German club said the player had pulled out because of a back injury. Iran refuses to recognise Israel's right to exist and does not allow its citizens to travel to the country. A Bayern spokesman said on Monday that the decision not to take Hashemian to Israel had been motivated only by his physical condition. "He's got back pain and he couldn't train," the spokesman said. "It would have made no sense for him to go." Iran gave judo world champion Arash Miresmaeili a $125,000 reward when he refused to fight an Israeli at the Athens Olympics. The International Judo Federation considered taking action against Miresmaeili but concluded he had been overweight for the fight and could not have taken part. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/141.txt: -------------------------------------------------------------------------------- 1 | Benitez issues warning to Gerrard 2 | 3 | Liverpool manager Rafael Benitez has ordered captain Steven Gerrard not to play down their Champions League ambitions and be more positive. 4 | 5 | Gerrard told the BBC Liverpool were unlikely to win the trophy this year. Benitez responded: "I spoke to Steven and said to him that in future it's better to think we can win the Champions League. Why not?" He said: "We need winners here and everyone thinking only of winning. I always want to win." 6 | 7 | Benitez added: "When we lose I only think of solutions. If you only think about winning the next game, you don't know what the draw will be. "If we can win the next game, maybe we will draw a side that isn't so strong, or a side with injuries or suspensions." Benitez is hoping to win his first trophy since arriving at Liverpool from Valencia when they play Chelsea in the Carling Cup on Sunday in Cardiff. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/143.txt: -------------------------------------------------------------------------------- 1 | Wenger handed summer war chest 2 | 3 | Arsenal boss Arsene Wenger has been guaranteed transfer funds to boost his squad the summer. 4 | 5 | The club's managing director, Keith Edelman, stressed that the development of their new £350m stadium had no affect on Wenger's spending power. "The money is there. Don't worry we've got it," Edelman told BBC Sport. "Hopefully, we'll spend it this summer and in the coming years. Arsene attends all our board meetings and he knows our finances are very strong." 6 | 7 | Edelman added that it was pointless having a brand new stadium if the team did not match the surroundings. "Its great to have nice, new surroundings, but if the team aren't performing on the pitch, then there isn't great respect in having a fabulous stadium," he said. "It's important that we had sufficient funds for our team in place, before we began on the stadium." 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/148.txt: -------------------------------------------------------------------------------- 1 | Palace threat over Cantona masks 2 | 3 | Manchester United fans wearing Eric Cantona masks will not be allowed in Selhurst Park on Saturday. 4 | 5 | United fans are planning to mark the 10th anniversary of Cantona's infamous kung-fu kick on Crystal Palace fan Matthew Simmons by wearing the masks. But Palace stadium manager Kevin Corner has warned: "If they wear them inside Selhurst Park they will be immediately rejected on safety grounds. "And no fans wearing Cantona masks will be allowed entry to the stadium." United fans had declared Saturday as 'Cantona Day' and had planned to wear masks, that were popular during the Frenchman's time as a player at the Old Trafford club. Cantona received a nine-month ban following the incident and Palace are keen to avoid an incident that would stir up ill feeling between the two sets of supporters. Four months after the Cantona incident a Palace fan, Paul Nixon, was killed prior to an FA Cup semi-final at Villa Park. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/161.txt: -------------------------------------------------------------------------------- 1 | Coach Ranieri sacked by Valencia 2 | 3 | Claudio Ranieri has been sacked as Valencia coach just eight months after taking charge at the Primera Liga club for the second time in his career. 4 | 5 | The decision was taken at a board meeting following the side's surprise elimination from the Uefa Cup. "We understand, and he understands, that the results in the last few weeks have not been the most appropriate," said club president Juan Bautista. Former assistant Antonio Lopez will take over as the new coach. Italian Ranieri took over the Valencia job in June 2004 having been replaced at Chelsea by Jose Mourinho. 6 | 7 | Things began well but the Spanish champions extended their winless streak to six after losing to Racing Santander last weekend. That defeat was then followed by a Uefa Cup exit at the hands of Steaua Bucharest. Ranieri first took charge of Valencia in 1997, guiding them to the King's Cup and helping them to qualify for the Champions League. The 54-year-old then moved to Atletico Madrid in 1999, before joining Chelsea the following year. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/162.txt: -------------------------------------------------------------------------------- 1 | Cole faces lengthy injury lay-off 2 | 3 | Aston Villa's Carlton Cole could be out for six weeks with a knee injury. 4 | 5 | The striker, who is on a season-long loan from Chelsea, picked up the knock in an England Under-21 match against Holland earlier this month. "Carlton will be out of action for four to six weeks after a bad challenge," said Villa boss David O'Leary. "I won't be able to tell you whether he will need an operation until maybe next week. Whether he has an operation has got to be left to Chelsea." Cole, who also struggled with an ankle problem earlier in the season, was unable to rest because O'Leary had a shortage of strikers. The return to fitness of Darius Vassell after four months out with a broken ankle and the emergence of Luke Moore has alleviated some of the Villa's manager's problems in that department. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/167.txt: -------------------------------------------------------------------------------- 1 | Robben plays down European return 2 | 3 | Injured Chelsea winger Arjen Robben has insisted that he only has a 10% chance of making a return against Barcelona in the Champions League. 4 | 5 | The 21-year-old has been sidelined since breaking a foot against Blackburn last month. Chelsea face Barcelona at home on 8 March having lost 2-1 in the first leg. And Robben told the Daily Star: "It is not impossible that I will play against Barcelona but it is just a very, very small chance - about 10%." 6 | 7 | Robben has been an inspirational player for Chelsea this season following a switch from PSV Einhoven last summer. He added: "My recovery is going better than we expected a few weeks ago but I think the Barcelona game will come too soon. "I won't take any risks and come back too soon." 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/171.txt: -------------------------------------------------------------------------------- 1 | Boro suffer Morrison injury blow 2 | 3 | Middlesbrough midfielder James Morrison has been ruled out for up to eight weeks after an operation on Tuesday. 4 | 5 | The 18-year-old, who has played in 13 of the club's last 14 games, had surgery to repair a double hernia. A club spokesman confirmed: "It is a bilateral sportsman's hernia, which was operated on yesterday." Morrison was sent for scans after being substituted at half-time during Boro's 2-1 UEFA Cup win over Graz AK in which he scored the equaliser. His injury is the latest blow for the Teessiders, who have been without Gaizka Mendieta, George Boateng and Mark Viduka for extended periods. Meanwhile, the kick-off time for Boro's Uefa Cup match at Sporting Lisbon on 17 March has been brought forward from 2115 GMT to 1945 GMT. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/173.txt: -------------------------------------------------------------------------------- 1 | QPR keeper Day heads for Preston 2 | 3 | Queens Park Rangers keeper Chris Day is set to join Preston on a month's loan. 4 | 5 | Day has been displaced by the arrival of Simon Royce, who is in his second month on loan from Charlton. QPR have also signed Italian Generoso Rossi. R's manager Ian Holloway said: "Some might say it's a risk as he can't be recalled during that month and Simon Royce can now be recalled by Charlton. "But I have other irons in the fire. I have had a 'yes' from a couple of others should I need them." 6 | 7 | Day's Rangers contract expires in the summer. Meanwhile, Holloway is hoping to complete the signing of Middlesbrough defender Andy Davies - either permanently or again on loan - before Saturday's match at Ipswich. Davies impressed during a recent loan spell at Loftus Road. Holloway is also chasing Bristol City midfielder Tom Doherty. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/174.txt: -------------------------------------------------------------------------------- 1 | Cudicini misses Carling Cup final 2 | 3 | Chelsea goalkeeper Carlo Cudicini will miss Sunday's Carling Cup final after the club dropped their appeal against his red card against Newcastle. 4 | 5 | The Italian was sent off for bringing down Shola Ameobi in the final minute of Sunday's match. Blues boss Jose Mourinho had promised to pick Cudicini for the final instead of first-choice keeper Petr Cech. The 31-year-old will now serve a one-match suspension commencing with immediate effect. 6 | 7 | Cudicini kept a club record 24 clean sheets last season for Chelsea, but Petr Cech has established himself as first choice for Mourinho since moving to Stamford Bridge in summer 2004. The 22-year-old Czech Republic international has set a new Premiership record of 961 consecutive minutes without conceding a goal, a mark which is still running. But Mourinho has used Cudicini regularly in the Carling Cup, and the Italian has only let in one goal in his four appearances during Chelsea's run to the final. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/175.txt: -------------------------------------------------------------------------------- 1 | Beattie return calms attack fears 2 | 3 | Everton striker James Beattie has been declared fit for training on Tuesday, despite suffering an alleged assault last Saturday. 4 | 5 | Beattie, 27, is not believed to be as seriously hurt as some reports have suggested after being "the victim of an unprovoked assault" in Birmingham. He was on a night out with his girlfriend and two friends following Everton's 3-1 win over Aston Villa. "He was shaken but not badly injured," said Everton spokesman Ian Ross. He added: "He did speak to the police but will not be pressing charges. "He has spoken to manager David Moyes about what happened." Beattie was suspended for the win at Aston Villa after his red card against Chelsea and he will be available again for the Merseyside derby at Liverpool on 20 March. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/191.txt: -------------------------------------------------------------------------------- 1 | Tottenham bid £8m for Forest duo 2 | 3 | Nottingham Forest have confirmed they have received an £8m bid from Tottenham for Andy Reid and Michael Dawson. 4 | 5 | Reid rejected a move to Southampton after Forest accepted a cash-plus-players offer while Spurs had made previous bids for the 22-year-old. Spurs had also made an undisclosed offer for 21-year-old defender Dawson. Forest chief executive Mark Arthur said: "We've received an £8m bid from Spurs for the two players and we're considering the offer." Southampton's offer for Republic of Ireland international Reid comprised £3m-plus and two players - Brett Ormerod and Darren Kenton. 6 | 7 | Spurs had seen previous bids of £3m, £4m and £5m for Reid rejected by the Championship club. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/194.txt: -------------------------------------------------------------------------------- 1 | Klinsmann issues Lehmann warning 2 | 3 | Germany coach Jurgen Klinsmann has warned goalkeeper Jens Lehmann he may have to quit Arsenal to keep his World Cup dreams alive. 4 | 5 | Lehmann is understudy to Oliver Kahn in the German squad, but has lost his place to Manuel Alumnia at Highbury. Klinsmann said: "It will be difficult for any of our players if he is not a first-choice at his club. "If Jens is not Arsenal's number one keeper, that is a problem for me. He must be playing regularly." Lehmann is desperate to keep his place in the Germany squad when the country hosts the World Cup in 2006. Klinsmann added: "If he is not playing regularly he cannot be Germany's number one keeper, or even number two keeper. "The situation for Jens is that he is currently the number two keeper at Arsenal. This could be critical if it remains the same during next season." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/199.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/al_benchmark/data/bbc_news/sport/199.txt -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/203.txt: -------------------------------------------------------------------------------- 1 | Mutu to hold talks with Juventus 2 | 3 | Disgraced former Chelsea striker Adrian Mutu is to begin talks with Juventus as he looks for a new club. 4 | 5 | "Mutu will be in our club's main stand to watch the match between Juventus and Fiorentina," said a Juventus official, who declined to be named. "Then he will meet some important people of the club," he added. Mutu, 25, sacked by Chelsea and banned for seven months for taking cocaine, is now represented by Alessandro Moggi, son of Juve manager Luciano Moggi. Mutu, banned by the English FA, can resume playing next May. He is to undergo a drug education and rehabilitation programme. Mutu wants to train with Romanian side Dinamo Bucharest whilst serving his ban, say the Bucharest side. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/214.txt: -------------------------------------------------------------------------------- 1 | Mansfield 0-1 Leyton Orient 2 | 3 | An second-half goal from Andy Scott condemned Mansfield to a ninth successive game without a win. 4 | 5 | Early in the second half Wayne Carlisle's cross was met by Scott and he blasted the ball home from just outside the penalty area. As Orient chased a second, Mansfield had to clear Alan White's header off the line and Kevin Pilkington saved well from Michael Simpson. By the end Mansfield fans were chanting for the head of chairman Keith Haslam. 6 | 7 | Pilkington, Talbot, Buxton, Dimech, Artell, Corden, Murray, Curtis, Neil, Warne, Barker. Subs Not Used: White, McIntosh, Wood, Lloyd, Herron. 8 | 9 | Harrison, Lockwood, Donny Barnard, White,Mackie, Scott, Saah, Simpson, Carlisle, Lee Barnard, Ibehre. Subs Not Used: Morris, Wardley, Newey, Zakuani, McMahon. 10 | 11 | Scott 51. 12 | 13 | 3,803 14 | 15 | S Mathieson (Cheshire). 16 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/220.txt: -------------------------------------------------------------------------------- 1 | Curbishley delight for Johansson 2 | 3 | Charlton manager Alan Curbishley paid tribute to two-goal striker Jonatan Johansson after beating Norwich. 4 | 5 | Curbishley said after the 4-0 win: "There was talk about Norwich being interested in the summer but I have a lot of faith in Jonatan. "When there was talk of other clubs I just told him he was going nowhere. "He's part of our squad and he got us a couple of important goals early on. I'm sure he's happy here, but I was delighted with all the players." Curbishley was subjected to criticism from Charlton fans recently, and he said: "The thing about this club is that we have got where we are by all pulling in the same direction. "That's happened for however long I've been here and we don't want people taking sides." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/224.txt: -------------------------------------------------------------------------------- 1 | Bristol City 2-1 Milton Keynes 2 | 3 | Leroy Lita took his goal tally to 13 for the season as his double earned City an LDV Vans Trophy win. 4 | 5 | The striker finished off Scott Murray cross from close range just seconds before half-time. Lita then made it 2-0 on 52 minutes, but Dons' substitute Serge Makofo then netted a great volley to make it 2-1. The visitors almost took the tie to extra time with a late 30-yard bullet from Richard Johnson which was well held by Steve Phillips. 6 | 7 | Phillips, Amankwaah, Coles, Hill, Fortune, Murray (Anyinsah 59), Doherty (Harley 45), Dinning, Bell, Lita (Cotterill 72), Gillespie. Subs Not Used: Orr, Brown. 8 | 9 | Hill. 10 | 11 | Lita 45, 52. 12 | 13 | Bevan, Oyedele, Ntimban-Zeh, Crooks, Puncheon, Kamara (Makofo 64), Chorley, Herve (McKoy 45), Tapp (Johnson 45), Mackie, Pacquette. Subs Not Used: Martin, Palmer. 14 | 15 | Pacquette, Chorley, Johnson, McKoy. 16 | 17 | Makofo 66. 18 | 19 | 3,367 20 | 21 | J Ross (Essex). 22 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/225.txt: -------------------------------------------------------------------------------- 1 | Hereford 1-1 Doncaster 2 | 3 | Hereford win 3-1 on penalties. 4 | 5 | Rovers took the lead on 59 minutes as Michael McIndoe's corner found Adriano Rigoglioso, whose stunning volley soared into the top right of the net. Danny Williams equalised for the hosts after 72 minutes with a crisp finish from just insist the penalty box. McIndoe saw an extra-time penalty saved, giving Mills the chance to net the winning penalty. 6 | 7 | Mawson, Travis, Mkandawire, James, Robinson, Daniel Williams, Stanley, Hyde (Pitman 105), Purdie (Mills 83), Brown, Stansfield (Green 102). Subs Not Used: Smith, Scott. 8 | 9 | James. 10 | 11 | Daniel Williams 72. 12 | 13 | Ingham, Mulligan, Fenton, Morley, Priet (Ryan 79), McIndoe, Ravenhill (Price 90), Rigoglioso, Tierney (Jackson 79), Beardsley, Coppinger. Subs Not Used: Warrington, Maloney. 14 | 15 | Rigoglioso 59. 16 | 17 | 1,375 18 | 19 | T Parkes (W Midlands). 20 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/229.txt: -------------------------------------------------------------------------------- 1 | McClaren targets Champions League 2 | 3 | Middlesbrough boss Steve McClaren believes his side can clinch a top-four spot in the Premiership and secure qualification for the Champions League. 4 | 5 | After their 3-2 win over Manchester City, McClaren said: "We are playing exciting football, it's a magnificent result to keep us in the top five. "But how well we do depends how often we can get our best team out. "Once we got the third goal it should have been four or five but we nearly paid for it in the end." McClaren also praised winger Stewart Downing and strikers Jimmy Floyd Hasselbaink and Mark Viduka, who both ended barren runs in front of goal. He added: "If Stewart keeps playing like this Sven-Goran Eriksson has got to pick him. "And the strikers scored great goals, the combination play between them shows they want to play with each other and they are trying." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/239.txt: -------------------------------------------------------------------------------- 1 | Beckham relief as Real go through 2 | 3 | David Beckham expressed his relief at Real Madrid's passage to the Champions League knockout phase. 4 | 5 | After Real's 3-0 win at Roma, the England skipper admitted another season of under-achievement would not be tolerated at the Bernabeu stadium. Beckham said: "It's expected of Madrid to get through, but it's a relief for the club and players to have won. "We lost momentum last season but we cannot afford to to go another season without winning anything." Real's finish as runners-up in their Champions League group means they cannot face his old club Manchester United in the next round. But Real could be drawn against other Premiership hopefuls, Arsenal or Chelsea, who won their respective groups. "It's going to be great whoever we play, even if we don't get either of the two English teams." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/241.txt: -------------------------------------------------------------------------------- 1 | Hamm bows out for US 2 | 3 | Women's football legend Mia Hamm has played her final game. 4 | 5 | Hamm, 32, who officially retired after this year's Athens Olympics, took to the field for the last time to help the US claim a 5-0 win over Mexico. Hamm ends her career as the most prolific scorer - male or female - in international football, with 158 goals in 276 games for the US. She was twice Fifa women's player of the year and won the World Cup in 1991 and 1999 and two Olympic golds. On her retirement, Hamm said: "There are mixed emotions. There are things in your life that you have had to put on the back-burner and you can focus more time and energy on that now. "I have to see what is out there and not commit to everything at the beginning." The friendly also saw the end of Julie Foudy and Joy Fawcett's US careers. Hamm will now embark on a new life with husband, Chicago Cubs shortstop Nomar Garciaparra, who was in the stands to watch her final game. He has just agreed a new deal with the Cubs after joining them in July. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/249.txt: -------------------------------------------------------------------------------- 1 | Solskjaer raises hopes of return 2 | 3 | Manchester United striker Ole Gunnar Solskjaer said he hoped to return next season following a career-threatening injury to his right knee. 4 | 5 | The 31-year-old Norway international had surgery in Sweden in August to transplant cartilage into the joint. "I'm in no doubt that I'll play again," Solskjaer told Aftenposten Daily. "The problem will be the speed but we're talking about eight months' time. It's going as it should with the knee. I'm slavishly following a programme." Solskjaer last played for United as a late substitute in May's FA Cup final win over Millwall. His contract with United runs out in 2006, by which time he will have been at the Premier League club for 10 years. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/256.txt: -------------------------------------------------------------------------------- 1 | Everton's Weir cools Euro hopes 2 | 3 | Everton defender David Weir has played down talk of European football, despite his team lying in second place in the Premiership after beating Liverpool. 4 | 5 | Weir told BBC Radio Five Live: "We don't want to rest on our laurels and say we have achieved anything yet. "I think you start taking your eye off the ball if you make statements and look too far into the future. "If you start making predictions you soon fall back into trouble. The only thing that matters is the next game." He said: "We are looking after each other and hard work goes a long way in this league. We have definitely shown that. "Also injuries and suspensions haven't cost us too badly and we have a lot of self-belief around the place." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/261.txt: -------------------------------------------------------------------------------- 1 | McClaren eyes Uefa Cup top spot 2 | 3 | Steve McClaren wants his Middlesbrough team to win their Uefa Cup group by beating Partizan Belgrade. 4 | 5 | Boro have already qualified for the knockout stages alongside Partizan and Villareal, at the expense of Lazio. But boss McClaren is looking for a victory which would mean they avoid a team that has played in the Champions League in Friday's third-round draw. "To need a win to finish top is fantastic, but it is going to be a tough one," McClaren said. "When the draw was made, I thought it was the toughest group of them all - and so it has proved. "Lazio were favourites, Villarreal have been semi-finalists, and Partizan have fantastic experience in Europe. "The pleasing thing is we did the business in the first two games. "Winning those two has put us in a great position and it has been a fantastic experience playing these teams." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/262.txt: -------------------------------------------------------------------------------- 1 | Spurs to sign Iceland U21 star 2 | 3 | Tottenham are primed to snap up Iceland Under-21 international Emil Hallfredsson after he impressed on trial at White Hart Lane. 4 | 5 | The 20-year-old midfielder, who plays for FH Hafnarfjordur, also starred in the Uefa Cup match against Scottish side Dunfermline earlier this season. Spurs have agreed a fee for the player, who has yet to agree personal terms. "He had offers from two other clubs but he decided to come to Tottenham," said Spurs sporting director Frank Arnesen. "He is a left-sided player, a position we have been looking at and he showed so much talent in his time here that we decided to take him. "It's down the road of bringing in talent, good prospects and giving them a place at Tottenham where they can improve." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/263.txt: -------------------------------------------------------------------------------- 1 | McLeish ready for criticism 2 | 3 | Rangers manager Alex McLeish accepts he is going to be criticised after their disastrous Uefa Cup exit at the hands of Auxerre at Ibrox on Wednesday. 4 | 5 | McLeish told BBC Radio Five Live: "We were in pole position to get through to the next stage but we blew it, we absolutely blew it. "There's no use burying your head in the sand, we know we are going to get a lot of criticism. "We have to take it as we have done in the past and we must now bounce back." McLeish admitted his team's defending was amateurish after watching them lose 2-0 to Guy Roux's French side. 6 | 7 | "I'm very disappointed because we didn't give ourselves a chance, losing the first goal from our own corner. It was amateur," he added. "The early goal in the second half gave us a mountain to climb and we never created the same kind of chances as we did in the first half. "It's difficult to take positives from the game. We've let the fans down." 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/266.txt: -------------------------------------------------------------------------------- 1 | Benitez 'to launch Morientes bid' 2 | 3 | Liverpool may launch an £8m January bid for long-time target Fernando Morientes, according to reports. 4 | 5 | The Real Madrid striker has been linked with a move to Anfield since the summer and is currently behind Raul, Ronaldo and Michael Owen at the Bernabeu. Liverpool boss Rafael Benitez is keen to bolster his forward options with Djibril Cisse out until next season. "If there is an attractive propostition it could be I would be keen to leave," admitted the 28-year-old Morientes. He added: "Unfortunately, I'm not in control of the situation. I'm under contract to Real and they will make any decisions." The fee could put Liverpool off a prospective deal but Real are keen to net the cash as they are reported to be preparing a massive summer bid for Inter Milan striker Adriano. The Reds are currently sixth in the Premiership, 15 points behind leaders Chelsea. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/271.txt: -------------------------------------------------------------------------------- 1 | Henry tipped for Fifa award 2 | 3 | Fifa president Sepp Blatter hopes Arsenal's Thierry Henry will be named World Player of the Year on Monday. 4 | 5 | Henry is on the Fifa shortlist with Barcelona's Ronaldinho and newly-crowned European Footballer of the Year, AC Milan's Andriy Shevchenko. Blatter said: "Henry, for me, is the personality on the field. He is the man who can run and organise the game." The winner of the accolade will be named at a glittering ceremony at Zurich's Opera house. The three shortlisted candidates for the women's award are Mia Hamm of the United States, Germany's Birgit Prinz and Brazilian youngster Marta. 6 | 7 | Hamm, who recently retired - is looking to regain the women's award, which she lost last year to striker Prinz. Fifa has changed the panel of voters for this year's awards. Male and female captains of every national team will be able to vote, as well as their coaches and Fipro - the global organisation for professional players. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/279.txt: -------------------------------------------------------------------------------- 1 | Newcastle line up Babayaro 2 | 3 | Newcastle manager Graeme Souness is closing in on signing Chelsea defender Celestine Babayaro when the transfer window reopens. 4 | 5 | Souness is bidding to bolster his defence and, according to reports, contract negotiations are at an advanced stage with the player. Babayaro has been in the Premiership since 1997 when he moved to Chelsea for £2.25m from Anderlecht. But the 26-year-old has been surplus to requirements this season. Souness would not be drawn on specifics over individual players. 6 | 7 | But he said: "All I can tell you is that the chairman has worked really hard in the last couple of months to try to do deals. "We have said from day one we want to strengthen, and that is what we are hoping to do in the coming weeks." 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/282.txt: -------------------------------------------------------------------------------- 1 | Ferguson hails Man Utd's resolve 2 | 3 | Manchester United's Alex Ferguson has praised his players' gutsy performance in the 1-0 win at Aston Villa. 4 | 5 | "That was our hardest away game of the season and it was a fantastic game of football, end-to-end with lots of good passing," said the Old Trafford boss. "We showed lots of character and guts and we weren't going to lose. "I look at that fixture and think we've been there and won, while Arsenal and Chelsea have yet to come and Villa may have some players back when they do." Ferguson also hailed senior stars Ryan Giggs and Roy Keane, who came off the bench for the injured John O'Shea. "Roy came on and brought a bit of composure to the midfield which we needed and which no other player has got. "Giggs was a tremendous threat and he brings tremendous penetration. "All we can do is maintain our form, play as we are and we'll get our rewards." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/296.txt: -------------------------------------------------------------------------------- 1 | Fear will help France - Laporte 2 | 3 | France coach Bernard Laporte believes his team will be scared going into their game with England on Sunday, but claims it will work in their favour. 4 | 5 | The French turned in a stuttering performance as they limped to a 16-9 win against Scotland in the opening match of the Six Nations on Saturday. "We will go to Twickenham with a little fear and it'll give us a boost," said the French coach. He added: "We are never good enough when we are favourites." Meanwhile, Perpignan centre Jean-Philippe Granclaude is delighted to have received his first call-up to the France squad. "It's incredible," the youngster said. "I was not expecting it at all. "Playing with the France team has always been a dream and now it has come true and I am about to face England at Twickenham in the Six Nations." Laporte will announce his starting line-up on Wednesday at the French team's training centre in Marcoussis, near Paris. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/305.txt: -------------------------------------------------------------------------------- 1 | Kirwan demands Italy consistency 2 | 3 | Italy coach John Kirwan has challenged his side to match the performance they produced in pushing Ireland close when they meet Wales on Saturday. 4 | 5 | Despite losing 28-17 in Sunday's Six Nations encounter, the Italians confirmed their continuing improvement. "Our goal is to match every side we face and against Ireland we showed we could do that," said Kirwan. "But the most important thing is that we build on that performance when we play Wales on Saturday." Italy's half-backs had a mixed afternoon, with recalled scrum-half Alessandro Troncon impressing but fly-half Luciano Orquera having an off-day with the boot. Kirwan said: "I was very happy with Troncon. He had an incredible game - he was very good in attack and defence. "Orquera's kicking was off but he showed great courage in defence. "He also followed the game plan. We have to give him confidence because he has the capability to do well." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/317.txt: -------------------------------------------------------------------------------- 1 | Munster Cup tie switched to Spain 2 | 3 | Munster's Heineken Cup quarter-final tie against Biarritz on 3 April has been switched to Real Sociedad's Paseo de Anoeta stadium in San Sebastian. 4 | 5 | Real's ground holds 32,000 whereas the Parc des Sports Aguilera in Biarritz has a capacity of just 12,667. The Irish province will be given at least 8,000 tickets. "The decision to move was a difficult one, but as we considered the fans as one of our primary objectives," said Biarritz chairman Marcel Martin. "We hope we will be rewarded as a huge crowd behaving in the best rugby tradition." The match will be the first Heineken Cup fixture to be played in Spain, and is expected to attract the biggest-ever attendance for a rugby match in the country. Ulster were the last Irish team to play at the Paseo de Anoeta stadium where they faced a Euskarians side during a pre-season tour in 1998. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/319.txt: -------------------------------------------------------------------------------- 1 | Pountney handed ban and fine 2 | 3 | Northampton coach Budge Pountney has been fined £2,000 and banned from match-day coaching for six weeks for calling a referee "a disgrace". 4 | 5 | Pountney was found guilty of bringing the game into disrepute at a Rugby Football Union disciplinary hearing in London on Thursday night. Pountney criticised referee Steve Lander's performance in Northampton's defeat by Saracens on 5 February. The last two weeks of the six-week ban are suspended. Pountney pleaded guilty to the offence before a panel consisting of chairman Robert Horner, Nigel Gillingham and Jeff Probyn. The ban means former Scotland international Pountney cannot enter the playing enclosure, technical areas or go near the touchline, tunnel or players and officials' areas on the day of a game. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/324.txt: -------------------------------------------------------------------------------- 1 | Leeds v Saracens (Fri) 2 | 3 | Headingley 4 | 5 | Friday, 25 February 6 | 7 | 2000 GMT 8 | 9 | The Tykes have brought in Newcastle prop Ed Kalman and Tom McGee from the Borders on loan while fly-half Craig McMullen has joined from Narbonne. Raphael Ibanez is named at hooker for Saracens in one of four changes. Simon Raiwalui and Ben Russell are also selected in the pack while Kevin Sorrell comes in at outside centre. 10 | 11 | - Friday's game at Headingley got the go-ahead on Friday after passing an early pitch inspection. Leeds: Balshaw; Rees, Christophers, Bell, Doherty; McMullen, Dickens; McGee, Rawlinson, Gerber; Murphy, Palmer (capt), Morgan, Parks, Popham. Replacements: Kalman, Regan, Hyde, Rigney, McMillan, Rock, Vickerman. Saracens: Bartholomeusz; Castaignede, Sorrell, Harris, Vaikona; Jackson, Bracken; Yates, Ibanez, Visagie; Raiwalui, Fullarton; Randell, Russell, Vyvyan (capt). Replacements: Cairns, Lloyd, Broster, Chesney, Johnston, Rauluni, Little. 12 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/325.txt: -------------------------------------------------------------------------------- 1 | Worcester v Sale (Fri) 2 | 3 | Sixways 4 | 5 | Friday, 25 February 6 | 7 | 2000 GMT 8 | 9 | They make just one change, with Tim Collier replacing Phil Murphy in the second row. In contrast, Sale are missing 14 players due to a combination of international call-ups and injuries. John Payne and Chris Rhys Jones come into the centres while scrum-half Richard Wigglesworth looks set to play on the wing. 10 | 11 | Delport; Pieters, Rasmussen, Lombard, O'Leary; Brown, Powell; Windo, Van Niekerk, Horsman; Collier, Gillies; Hickey, Sanderson, MacLeod- Henderson. 12 | 13 | Replacements: Fortey, Murphy, Daly, Vaili, Cole, Hayes, Trueman. 14 | 15 | Hanley; Mayor, Payne, Rhys Jones, Wigglesworth; Hercus, Redpath (capt); Turner, Roddam, Stewart; Day, Schofield; Caillet, Carter, Chabal. 16 | 17 | Replacements (from): Bozzi, Coutts, Anglesea, Lund, Martens, Riley, Duffy, C Jones. 18 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/356.txt: -------------------------------------------------------------------------------- 1 | Dawson joins England injury list 2 | 3 | Scrum-half Matt Dawson is an injury doubt for England's Six Nations opener against Wales next weekend. 4 | 5 | The World Cup winner missed Wasps' 12-9 loss to Bath on Saturday after injuring his right calf. Wasps coach Warren Gatland said: "He's got a fitness test in the week but he's got a good chance of playing." Gloucester's Andy Hazell and Leicester star Lewis Moody also received knocks during their respective league matches, but should be fit for Wales next week. If Dawson is not fit to face Wales, Robinson will have to choose from Gloucester's Andy Gomarsall or Leicester youngster Harry Ellis. 6 | 7 | Jamie Noon is another player on the sidelines after he limped off in the first half against Saracens on Friday with a dead leg. The centre, who is in line for a first Six Nations start against Wales, will have to wait 48 hours before knowing the state of his injury. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/388.txt: -------------------------------------------------------------------------------- 1 | Wilkinson fit to face Edinburgh 2 | 3 | England captain Jonny Wilkinson will make his long-awaited return from injury against Edinburgh on Saturday. 4 | 5 | Wilkinson, who has not played since injuring his bicep on 17 October, took part in full-contact training with Newcastle Falcons on Wednesday. And the 25-year-old fly-half will start Saturday's Heineken Cup match at Murrayfield on the bench. But Newcastle director of rugby Rob Andrew said: "He's fine and we hope to get him into the game at some stage." The 25-year-old missed England's autumn internationals after aggravating the haematoma in his upper right arm against Saracens. He was subsequently replaced as England captain by full-back Jason Robinson. Sale's Charlie Hodgson took over the number 10 shirt in the internationals against Canada, South Africa and Australia. Wilkinson's year has been disrupted by injury as his muscle problem followed eight months on the sidelines with a shoulder injury sustained in the World Cup final. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/390.txt: -------------------------------------------------------------------------------- 1 | Dawson set for new Wasps contract 2 | 3 | European champions Wasps are set to offer Matt Dawson a new deal. 4 | 5 | The 31-year-old World Cup winning scrum-half has impressed since joining the London side from Northampton this summer on a one-year contract. Wasps coach Warren Gatland told the Daily Mirror: "We have not yet offered Matt a new contract but we will be doing so. "I'm very happy with his contribution and I think he's good enough to play for another couple of years." Dawson played a vital part in England's World Cup win last year but has fallen out of favour with new coach Andy Robinson after missing a training session in September. However he hopes the new deal will help him regain his England place. 6 | 7 | "Rugby is still my priority and there's still a burning desire within me to play the best rugby I possibly can," he said. "I know within myself, if I was given the chance I could play for England again. "I know I'm fit enough, I'm strong enough, I'm skilful enough." 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/393.txt: -------------------------------------------------------------------------------- 1 | Bath faced with Tindall ultimatum 2 | 3 | Mike Tindall's agent has warned Bath they have until next week to improve their contract offer to the England man or risk losing him to a rival club. 4 | 5 | Dipo Alli says he has received an offer for Tindall which dwarfs Bath's deal and that two other clubs want to talk. "Mike does not want to go into the Six Nations worrying about where he will be playing his club rugby next season," Alli told the Guardian newspaper. "It is up to (Bath owner) Andrew Brownsword. He has to make it happen." Tindall is out of contract at the end of the season but it is understood that Brownsword is unwilling to break the club's salary structure to accommodate the 26-year-old's demands. But Alli insists the player is worth more than Bath have put on the table. "Mike has been at Bath for eight years and wants to remain with the club and his demands are anything but excessive," the agent added. "But Brownsword has to recognise Mike's value and we want to resolve things by the end of next week." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/422.txt: -------------------------------------------------------------------------------- 1 | Safin slumps to shock Dubai loss 2 | 3 | Marat Safin suffered a shock loss to unseeded Nicolas Kiefer in round one of the Dubai Tennis Championships. 4 | 5 | Playing his first match since winning the Australian Open, Safin showed some good touches but was beaten 7-6 (7-2) 6-4 by the in-form Kiefer. The German got on top in the first-set tie-break, striking a sweet forehand to win the first point against serve. And he maintained the momentum early in the second set, breaking the Russian with the help of an inspired volley. Spain's Feliciano Lopez lined up a second round clash with Andre Agassi by beating Thailand's Paradorn Srichaphan. Lopez, who lost in three sets to Roger Federer in last year's final, won 6-2 3-6 6-3. Former champion Fabrice Santoro of France was beaten 6-3 6-0 by sixth seeded Russian Nikolay Davydenko. There were also wins for two other Russians, Igor Andreev and seventh seed Mikhail Youzhny. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/440.txt: -------------------------------------------------------------------------------- 1 | Martinez sees off Vinci challenge 2 | 3 | Veteran Spaniard Conchita Martinez came from a set down to beat Italian Roberta Vinci at the Qatar Open in Doha. 4 | 5 | The 1994 Wimbledon champion won 5-7 6-0 6-2 to earn a second round meeting with French Open champion Anastasia Myskina. Fifth seed Patty Schnyder also had a battle as she needed three sets to beat China's Na Li 7-5 3-6 7-5. Slovakian Daniela Hantuchova beat Bulgarian Magdaleena Maleeva 4-6 6-4 6-3 to set up a second round clash with Russian Elena Bovina. The veteran Martinez found herself in trouble early on against Vinci with the Italian clinching the set thanks to breaks in the third and 11th games. But Vinci's game fell to pieces after that and Martinez swept her aside with some crisp cross-court returns and deft volleys. In the day's other matches, Japan's Ai Sugiyama defeated Australian Samantha Stosur 6-2 6-3 while Australian Nicole Pratt beat Tunisian Selima Sfar 7-5 6-2 and will next face compatriot Alicia Molik. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/441.txt: -------------------------------------------------------------------------------- 1 | Serena becomes world number two 2 | 3 | Serena Williams has moved up five places to second in the world rankings after her Australian Open win. 4 | 5 | Williams won her first Grand Slam title since 2003 with victory over Lindsay Davenport, the world number one. Men's champion Marat Safin remains fourth in the ATP rankings while beaten finalist Lleyton Hewitt replaces Andy Roddick as world number two. Roger Federer retains top spot, but Safin has overtaken Hewitt to become the new leader of the Champions Race. Alicia Molik, who lost a three-set thriller against Davenport in the quarter-finals, is in the women's top 10 for the first time in her career. Her rise means Australia have a player in the top 10 of the men's and women's rankings for the first time in 21 years. And Britain's Elena Baltacha, who qualified and then reached the third round, has risen to 120 in the world - a leap of 65 places and her highest ranking yet. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/442.txt: -------------------------------------------------------------------------------- 1 | Hingis to make unexpected return 2 | 3 | Martina Hingis makes her return to competitve tennis after two years out of the game at the Volvo Women's Open in Pattaya, Thailand, on Tuesday. 4 | 5 | She faces German Marlene Weingartner in the first round. "As a competitor and athlete, I always want to win. I hope my body will hold up," said Hingis. "You miss being out there in a Grand Slam final and not competing. It's a big difference between playing and commenting for TV. I miss it a lot." The former world number one was 22 when she retired after having surgery on both ankles, and her last WTA event was in Filderstadt, Germany, in October 2002, when she lost to Elena Dementieva. Only last year she insisted that a comeback was unlikely, but speaking in Thaliand Hingis admitted: "I said that because I didn't know what would happen with my body." Her appearance will also benefit charities in the region and the Swiss star will donate her prize money. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/443.txt: -------------------------------------------------------------------------------- 1 | Clijsters set for February return 2 | 3 | Tennis star Kim Clijsters will make her return from a career-threatening injury at the Antwerp WTA event in February. 4 | 5 | "Kim had considered returning to action in Paris on 7 February," a statement on her website said. "She's decided against this so that she does not risk the final phase of her recovery. If all goes well, Kim will make her return on February 15." The 21-year-old has not played since last October after aggravating a wrist injury at the Belgian Open. Back then, a doctor treating the Belgian feared that her career may be over, with the player having already endured an operation earlier in the season to cure her wrist problem. "I hope she comes back, but I'm pessimistic," said Bruno Willems. Clijsters was also due to marry fellow tennis star Lleyton Hewitt in February but the pair split "for private reasons" back in October. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/444.txt: -------------------------------------------------------------------------------- 1 | Hewitt falls to Dent 2 | 3 | Lleyton Hewitt suffered a shock defeat to Taylor Dent in the quarter-finals of the Australian Hardcourt Championships in Adelaide on Friday. 4 | 5 | The top seed was a strong favourite for the title but went down 7-6 (7-4) 6-3 to the American. Dent will face Juan Ignacio Chela next after the fourth seed was too strong for Jurgen Melzer. Olivier Rochus beat third seed Nicolas Kiefer 6-7 (4-7) 7-6 (8-6) 7-5 and will take on second seed Joachim Johansson. The Swede reached the last four by beating compatriot Thomas Enqvist 6-3 4-6 6-1. "I felt like I was striking the ball much better," said Johansson. "I felt like I had a lot of break chances, I didn't take care of them all, but I broke him four times and he only broke me once. "I felt that was the key to get up in the set early." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/452.txt: -------------------------------------------------------------------------------- 1 | Rochus shocks Coria in Auckland 2 | 3 | Top seed Guillermo Coria went out of the Heineken Open in Auckland on Thursday with a surprise loss to Olivier Rochus of Belgium. 4 | 5 | Coria lost the semi-final 6-4 6-4 to Rochus, who goes on to face Czech Jan Hernych, a 6-4 7-5 winner over Jose Acasuso of Argentina. Fifth seed Fernando Gonzalez eased past American Robby Ginepri 6-3 6-4. The Chilean will meet sixth seed Juan Ignacio Chela next after the Argentine beat Potito Starace 6-1 7-6 (7-5). Rochus made the semi-finals at the Australian hardcourt championships in Adelaide last week and is naturally delighted with his form. "It's been two unbelievable weeks for me," he said. "Today I knew I had nothing to lose. If I beat him great, if I lost, I would be losing to a top-10 player." Coria conceded that Rochus "played just too good," and added: "When you give your best out there you can't be too sad." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/453.txt: -------------------------------------------------------------------------------- 1 | Johansson takes Adelaide victory 2 | 3 | Second seed Joachim Johansson won his second career title with a 7-5 6-3 win over Taylor Dent at the Australian hardcourt championships in Adelaide. 4 | 5 | The Swede was made to graft, American Dent surviving three break points in the fifth game of the match. But Johansson got the breakthrough with a sublime backhand return winner and won the second set with more ease. His first tournament win was at Memphis in 2004, helping him leap from 113th in the world rankings to number 11. Afterwards, Dent said he rated US Open semi-finalist Johansson as a top contender at the Australian Open, which starts on 17 January. "I believe men's tennis is all about holding serve and if he's playing like that on his own serve I don't see how guys are going to break him," said Dent. Johansson was more restrained in his assessment: "I have to improve my serve if I'm going to go all the way in Melbourne." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/454.txt: -------------------------------------------------------------------------------- 1 | Hingis hints at playing comeback 2 | 3 | Martina Hingis has admitted that she might consider a competitive return to tennis if an appearance in Thailand later this month goes well. 4 | 5 | The former world number one will play at the Volvo Women's Open in Pattaya, which starts on 31 January, as part of her charity work in the region. "The tournament is a test," she said. "I don't know how my body will react. "I support several charities in Thailand. I'm also playing to see where I am." Speaking to Le Matin, the 24-year-old Swiss added: "At Pattaya there will not be as many people and the players are ranked between the 30th and 95th in the world." Hingis was 22 when she retired after having surgery on both ankles, and her last WTA event was in Filderstadt, Germany, in October 2002, when she lost to Elena Dementieva. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/456.txt: -------------------------------------------------------------------------------- 1 | Dementieva prevails in Hong Kong 2 | 3 | Elena Dementieva swept aside defending champion Venus Williams 6-3 6-2 to win Hong Kong's Champions Challenge event. 4 | 5 | The Russian, ranked sixth in the world, broke Williams three times in the first set, while losing her service once. Williams saved three championship points before losing the match at the Victoria Park tennis court. "It's really a great start to the year no matter whether it's an exhibition or not. I was trying to play my best and I really did it," said Dementieva. "This will give me all the confidence before the Grand Slams. I was trying so hard to win this tournament." Williams, 24, was disappointed with her display. "She played some nice points, but it was mostly me committing unforced errors - four or five errors in each game," she said. Before the match, organizers auctioned off rackets belonging to the players, raising £115,000 for victims of the tsunami disaster. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/457.txt: -------------------------------------------------------------------------------- 1 | Injury sidelines Philippoussis 2 | 3 | Mark Philippoussis withdrew from the Sydney International tennis tournament as expected on Sunday after suffering a groin injury during the Hopman Cup. 4 | 5 | His participation in the Australian Open, which begins on 17 January in Melbourne, also remains in doubt. Defending women's champion Justine Henin-Hardenne is also out of the Sydney event because of a knee injury. In the only main draw men's or women's singles match on Sunday, Nathalie Dechy beat American Lisa Raymond 7-5 6-3. Number one men's seed Lleyton Hewitt begins his quest for a fourth Sydney title on Tuesday when he plays Karol Beck. Lindsay Davenport, top seed in the women's draw, has been handed a first-round bye and plays France's Dechy in the second round on Tuesday. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/459.txt: -------------------------------------------------------------------------------- 1 | Federer breezes into semi-finals 2 | 3 | Roger Federer reached the last four of the Qatar Open with an easy 6-1 6-2 win over seventh seed Feliciano Lopez. 4 | 5 | The Swiss world number one reeled off a series winners to outclass the Spaniard and set up a semi-final match against Russian Nikolay Davydenko. Federer, who lost in the quarter-final in his last Qatar appearance in 2003, was happy with his form. "I think I played better than against Greg Rusedski and I am happy I am playing so well," said the top seed. Lopez showed glimpses of resolve early in the second set when he held his first service game and came close to breaking Federer. But the Swiss saved a break point and promptly broke serve in the following game to seize control. Davydenko, meanwhile, upset French third seed Sebastien Grosjean 2-6 6-3 6-2. Fabrice Santoro completed a miserable day for France when he was forced to retire when 6-2 3-0 down to Albert Costa. Spaniard Costa will next face Croatian Ivan Ljubicic after the sixth seed beat Rafael Nadal 6-2 6-7 (3/7) 6-3. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/460.txt: -------------------------------------------------------------------------------- 1 | Davenport puts retirement on hold 2 | 3 | Lindsay Davenport has put any talk of retirement on hold after having a largely injury-free 2004 campaign. 4 | 5 | The 28-year-old world number one had said that she would quit at the end of last year, but after a successful season she has had a change of heart. "Finally I felt I put myself in a position to try and win Grand Slams again," said Davenport. "It would be tough to walk away when I feel like I can contend so there's no point in hanging it up quite yet." Davenport has won three Grand Slams, the 2000 Australian Open, Wimbledon in 1999 and the 1998 US Open. Her career has been hit by a series of injuries but last year she started hitting top form and won seven titles. She was due to take part in this week's Hopman Cup in Perth but decided she wanted to rest her knee. "I just really wanted to make sure my right knee was going to be able to really withstand all the rigours of the whole year coming up," she said. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/463.txt: -------------------------------------------------------------------------------- 1 | Prodigy Monfils blows away Gaudio 2 | 3 | French prodigy Gael Monfils underlined his huge promise by beating French Open champion Gaston Gaudio 6-4 7-6 (7-4) in the first round of the Qatar Open. 4 | 5 | The 18-year-old wild card won three of the four junior Grand Slam events last year, including Wimbledon. Fabrice Santoro, the 2000 champion, beat Sweden's Thomas Johansson 6-4 6-2 but fourth seed Mikhail Youzhny lost 6-3 7-6 (7-3) to Rafael Nadal. Roger Federer plays Greg Rusedski in the second round on Wednesday. Monfils, who was given a wildcard into the tournament, said: "This is my first win over a top 10 player and I am delighted. "I play my best tennis when I am fired up on the court and the reason I won today was because I was able to play my natural, attacking game," he said. "Of course I was a bit tired in the second set. But I was confident I could survive had there been a third set." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/467.txt: -------------------------------------------------------------------------------- 1 | Capriati to miss Melbourne 2 | 3 | Jennifer Capriati has become the third leading lady to withdraw from the Australian Open because of injury. 4 | 5 | The organisers of the first grand slam of 2005, which begins on 17 January, said the American has a problem with her right shoulder. It comes as a blow to the women's draw as last year's champion, Justin Henin-Hardenne, and runner-up, Kim Clijsters, will also be absent. Capriati is a two-time champion in Melbourne with wins in 2001 and 2002. She is believed to have picked up the injury at the Advanta Championships at Philadelphia in November and had to pull out of an exhibition match with Wimbledon champion Maria Sharapova on 17 December. Capriati also decided against competing in the Australian Open warm-up event, the Sydney International. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/471.txt: -------------------------------------------------------------------------------- 1 | Capriati out of Australian Open 2 | 3 | Jennifer Capriati has become the third leading lady to withdraw from the Australian Open because of injury. 4 | 5 | The organisers of the first grand slam of 2005, which begins on 17 January, said the American has a problem with her right shoulder. It comes as a blow to the women's draw as last year's champion, Justin Henin-Hardenne, and runner-up, Kim Clijsters, will also be absent. Capriati is a two-time champion in Melbourne with wins in 2001 and 2002. She is believed to have picked up the injury at the Advanta Championships at Philadelphia in November and had to pull out of an exhibition match with Wimbledon champion Maria Sharapova on 17 December. Capriati also decided against competing in the Australian Open warm-up event, the Sydney International. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/482.txt: -------------------------------------------------------------------------------- 1 | Roddick in talks over new coach 2 | 3 | Andy Roddick is reportedly close to confirming US Davis Cup assistant Dean Goldfine as his new coach. 4 | 5 | Roddick ended his 18-month partnership with Brad Gilbert on Monday, and Goldfine admits talks have taken place. "We had a really good conversation and we're on the same page in terms of what I expect from a player in commitment and what he wants," said Goldfine. "The reading I got from him is that I would have a lot of the qualities he's looking for in a coach." Speaking to told South Florida's Sun-Sentinel newspaper, Goldfine added: "That being said, from his standpoint, which is smart, he wants to cover all his bases. "I think Andy wants a long-term relationship and wants to make sure it's the right fit... the best fit." Goldfine, 39, has worked with Todd Martin and Roddick's close friend Mardy Fish, and was an assistant coach with the US Olympic team. Martin is the other name to have been linked to the vacant post alongside Roddick. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/484.txt: -------------------------------------------------------------------------------- 1 | Sydney return for Henin-Hardenne 2 | 3 | Olympic champion Justine Henin-Hardenne will return to action in January's Sydney International tournament. 4 | 5 | The Belgian has not competed since losing her top world ranking at the US Open in September, where she was beaten in the fourth round by Nadia Petrova. She took time out to shake off a virus but will defend her titles in Sydney and at the Australian Open. Women's world number one Lindsay Davenport and French Open champion Anastasia Myskina will also compete. 6 | 7 | In the men's event, world number three Lleyton Hewitt returns to defend his title, along with runner-up Carlos Moya. Moya, Spain's Davis Cup final hero in their recent win over the US, had to retire with an ankle injury in the first set of the final. 8 | 9 | Tournament director Craig Watson said: "I had a message relayed to me from him after Spain's Davis Cup victory, saying he was looking forward to trying to make up for his disappointment in the (2004) final. The tournament will take place from 9-15 January. 10 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/485.txt: -------------------------------------------------------------------------------- 1 | Koubek suspended after drugs test 2 | 3 | Stefan Koubek says he has been banned for three months by the International Tennis Federation (ITF) after testing positive for a banned substance. 4 | 5 | The world number 60 failed a routine drugs test at this year's French Open but now plans to lodge an appeal. Koubek believes an injection given to him by an Austrian doctor to treat a wrist injury is to blame for producing traces of the substance in his system. "I have acted correctly," the 27-year-old Austrian said in a statement. Koubek, who defeated Britain's Greg Rusedski in the decisive rubber of the Davis Cup in September, is now set to miss the start of the season. He said, "A three-month ban would mean that I not only will miss the Australian Open, but also the Davis Cup in Australia." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/498.txt: -------------------------------------------------------------------------------- 1 | Dent continues Adelaide progress 2 | 3 | American Taylor Dent reached the final of the Australian hardcourt event in Adelaide with a crushing 6-1 6-1 win over Argentine Juan Ignacio Chela. 4 | 5 | Dent will meet Swede Joachim Johansson on Sunday after the second seed survived a tense tie-break to defeat Belgium's Olivier Rochus 6-1 7-6 (7/5). Johansson, the boyfriend of Lleyton Hewitt's sister Jaslyn, received strong crowd support on Saturday. "It feels like home for me, because Jaslyn lives here," said Johansson. Rochus was leading 5-4 in the second set tiebreak but his concentration was ruffled by a disputed line call and the match slipped away. "It was so close - one mistake like this and the match is over, it's tough. For me, it was clearly out," Rochus said. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/501.txt: -------------------------------------------------------------------------------- 1 | Ivanovic seals Canberra victory 2 | 3 | Serbia's Ana Ivanovic captured her first WTA title with a straight-sets victory over Hungarian Melinda Czink in the final of the Canberra Classic. 4 | 5 | The 17-year-old took 83 minutes to take the match 7-5 6-1. Ivanovic beat Czink in the last round of qualifying but the Hungarian made the main draw as a lucky loser after Katarina Srebotnik withdrew injured. Ivanovic said: "I was really nervous in the beginning, but I pulled through and didn't do too much wrong." A junior Wimbledon finalist last year, she added: "It's my first WTA title, and to win it has really given me more energy to practice and improve." Ivanovic will play 32nd seed Iveta Benesova of the Czech Republic in the first round.said of the Australian Open in Melbourne. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/504.txt: -------------------------------------------------------------------------------- 1 | Veteran Martinez wins Thai title 2 | 3 | Conchita Martinez won her first title in almost five years with victory over Anna-Lena Groenefeld at the Volvo Women's Open in Pattaya, Thailand. 4 | 5 | The 32-year-old Spaniard came through 6-3 3-6 6-3 for her first title since Berlin in 2000. "It feels really good," said Martinez, who is playing her last season on the Tour. "To come through like that in an important match feels good. "It's been nearly five years and I didn't think I could do it." Groenefeld was the more powerful player but could not match her opponent's relentless accuracy. "It was my first final, a new experience," said the German. "I think she played a good match, a tough match, but I tried to stay in there. I think the whole week was good for me." 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/505.txt: -------------------------------------------------------------------------------- 1 | Soderling wins tense Milan final 2 | 3 | Fifth seed Robin Soderling took the Milan Indoors title with a dramatic win over Radek Stepanek in Sunday's final. 4 | 5 | The 20-year-old Swede edged the final set tie-break for a 6-3 6-7 (2-7) 7-6 (7-5) victory and his second tour title after winning in Lyon last year. "I'm delighted to have won against such a good opponent in a tournament of this importance," said Soderling. "I was really on form, my service was good and I really liked playing on the synthetic surface." Soderling, the world number 37, shrugged off a nervous start to take the opening set on his first set point. A change in approach from fourth seed Stepanek in the second set paid dividends as he took it on a tie-break, but Soderling's superior power proved too much in the third. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/506.txt: -------------------------------------------------------------------------------- 1 | Rusedski forced out in Marseille 2 | 3 | Greg Rusedski was forced to withdraw from the Open 13 in Marseille on Thursday with a rib injury. 4 | 5 | The British number two had been scheduled to play qualifier Sebastien de Chaunac, who beat world number five Guillermo Coria 6-4 7-5 in round one. But Rusedski was unable to take to the court because of a problem with the left-hand side of his rib-cage. American Taylor Dent caused a shock with a 7-6 6-2 victory over second seed David Nalbandian. But third seed Joachim Johansson made it through after beating Frenchman Gilles Simon 7-6 6-3 while in the first match of the day, sixth seed Feliciano Lopez defeated Ivo Karlovic. There were also wins for Slovakia's Karol Beck and Croatian duo Ivan Ljubicic and Mario Ancic. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/sport/508.txt: -------------------------------------------------------------------------------- 1 | Almagro continues Spanish surge 2 | 3 | Unseeded Nicolas Almagro became the fifth Spaniard to reach the last eight at the Buenos Aires Open, ousting eighth seed Mariano Zabaleta. 4 | 5 | He showed admirable resolve to win a rain-affected match 6-7 6-4 6-4. Compatriot and seventh seed Rafael Nadal also reached the last eight, beating Italian Potito Starace 6-1 6-3. Nadal, playing in the outdoor clay event for the first time, hit some powerful forehands to oust Starace in a match delayed over an hour by rain. "It's always a problem to have to stop for rain but one gets used to it," said Spanish teenager Nadal. "Luckily, I was able to keep my pace going throughout the match." He will now play Gaston Gaudio, who beat unseeded Brazilian Flavio Saretta 6-3 6-2 in the day's late match. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/tech/238.txt: -------------------------------------------------------------------------------- 1 | Gamers snap up new Sony PSP 2 | 3 | Gamers have bought almost all of the first batch of Sony's new PlayStation Portable (PSP) games console, which went on sale in Japan on Sunday. 4 | 5 | Thousands of people queued for hours to get hold of one of the 200,000 PSPs which were shipped to retailers. The handheld console can play games, music and movies and goes on sale in Europe and North America next year. Despite the demand Sony said it would not increase the 500,000-strong stock of PSPs it plans to ship by year's end. 6 | 7 | Sony says it intends to ship three million of the consoles by March 2005. The company is hoping to challenge the dominance of Nintendo in the handheld market. Nintendo released its new DS console earlier this year and has already raised shipment targets for the device by 40%. The PSP is selling in Japan for 19,800 yen ($188; £98) while Nintendo's DS console sells in the US and Japan for $150 (£78). Nintendo's goal is to ship 5 million of its new Nintendo DS handheld consoles by March 2005. 8 | -------------------------------------------------------------------------------- /acleto/al_benchmark/data/bbc_news/tech/284.txt: -------------------------------------------------------------------------------- 1 | Hacker threat to Apple's iTunes 2 | 3 | Users of Apple's music jukebox iTunes need to update the software to avoid a potential security threat. 4 | 5 | Hackers can build malicious playlist files which could crash the program and let them seize control of the computer by inserting Trojan code. A new version of iTunes is now available from the Apple website which solves the problem. Security firm iDefence, which notified users of the problem, recommended that users upgrade to iTunes version 4.7.1. The problem affects all users of iTunes - Windows and Mac OS - running versions 4.7 and earlier. Users can automatically upgrade iTunes by opening the "look for updates" window in the program. The security firm says users should avoid clicking on or accessing playlist files - which have the file extension of .pls or .m3u - which have come from unknown sources. Itunes is the world's most popular online music store with more than 200 million songs downloaded since it launched in 2003. 6 | -------------------------------------------------------------------------------- /acleto/al_benchmark/run_scripts/__init__.py: -------------------------------------------------------------------------------- 1 | 2 | -------------------------------------------------------------------------------- /acleto/annotator_tool/__init__.py: -------------------------------------------------------------------------------- 1 | from .active_learner import ActiveLearner 2 | from .annotator_widget import AnnotatorWidget 3 | from .ui_widget import ActiveLearnerUiWidget 4 | from .visualizers import ImageVisualizer, TextAreaVisualizer 5 | -------------------------------------------------------------------------------- /acleto/annotator_tool/al4nlp_adaptor/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/annotator_tool/al4nlp_adaptor/__init__.py -------------------------------------------------------------------------------- /acleto/annotator_tool/annotation_converter.py: -------------------------------------------------------------------------------- 1 | from abc import abstractmethod 2 | 3 | 4 | class AnnotationConverter: 5 | """ Converter from anntations in the widget to annotations for internal purpose. """ 6 | 7 | @abstractmethod 8 | def __call__(self, indexes, answers): 9 | pass 10 | 11 | 12 | class AnnotationConverterDefault(AnnotationConverter): 13 | def __call__(self, indexes, answers): 14 | return answers 15 | -------------------------------------------------------------------------------- /acleto/annotator_tool/libact_adaptor/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/annotator_tool/libact_adaptor/__init__.py -------------------------------------------------------------------------------- /acleto/annotator_tool/visualizers/__init__.py: -------------------------------------------------------------------------------- 1 | from .image import ImageVisualizer 2 | from .textarea import TextAreaVisualizer 3 | -------------------------------------------------------------------------------- /acleto/post_processing/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/acleto/post_processing/__init__.py -------------------------------------------------------------------------------- /acleto/visualize/__init__.py: -------------------------------------------------------------------------------- 1 | from .extract_metrics import collect_data 2 | from .plot_figure import plot_with_confidence_interval 3 | from .visualize import visualize_experiments, plot_query 4 | -------------------------------------------------------------------------------- /distillation/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/distillation/__init__.py -------------------------------------------------------------------------------- /distillation/distil.sh: -------------------------------------------------------------------------------- 1 | python train.py \ 2 | --student_type distilelectra \ 3 | --student_config training_configs/distilelectra.json \ 4 | --teacher_type electra \ 5 | --teacher_name google/electra-base-discriminator \ 6 | --alpha_ce 5.0 --alpha_mlm 2.0 --alpha_cos 1.0 --alpha_act 1.0 --alpha_clm 0.0 --mlm \ 7 | --freeze_pos_embs \ 8 | --data_file data/binarized_text.electra.pickle \ 9 | --token_counts data/token_counts.electra.pickle \ 10 | --dump_path ./serialization_dir/distilelectra \ 11 | --force # overwrites the `dump_path` if it already exists. -------------------------------------------------------------------------------- /distillation/distil_distributed.sh: -------------------------------------------------------------------------------- 1 | export NODE_RANK=0 2 | export N_NODES=1 3 | 4 | export N_GPU_NODE=4 5 | export WORLD_SIZE=4 6 | 7 | pkill -f 'python -u train.py' 8 | 9 | python -m torch.distributed.launch \ 10 | --nproc_per_node=$N_GPU_NODE \ 11 | --nnodes=$N_NODES \ 12 | --node_rank $NODE_RANK \ 13 | train.py \ 14 | --force \ 15 | --n_gpu $WORLD_SIZE \ 16 | --student_type distilelectra \ 17 | --student_config training_configs/distilelectra.json \ 18 | --teacher_type electra \ 19 | --teacher_name google/electra-base-discriminator \ 20 | --alpha_ce 5.0 --alpha_mlm 2.0 --alpha_cos 1.0 --alpha_act 1.0 --alpha_clm 0.0 --mlm \ 21 | --freeze_pos_embs \ 22 | --data_file data/binarized_text.electra.pickle \ 23 | --token_counts data/token_counts.electra.pickle \ 24 | --dump_path ./serialization_dir/distilelectra \ 25 | --force # overwrites the `dump_path` if it already exists. -------------------------------------------------------------------------------- /distillation/prepare_data.sh: -------------------------------------------------------------------------------- 1 | python scripts/binarized_data.py \ 2 | --dataset_name ag_news \ 3 | --tokenizer_type electra \ 4 | --tokenizer_name google/electra-base-discriminator \ 5 | --dump_file ./data/binarized_text & 6 | -------------------------------------------------------------------------------- /distillation/requirements.txt: -------------------------------------------------------------------------------- 1 | transformers 2 | datasets 3 | 4 | hydra-core 5 | gitpython==3.0.2 6 | tensorboard>=1.14.0 7 | tensorboardX==1.8 8 | psutil 9 | scipy>=1.4.1 10 | -------------------------------------------------------------------------------- /distillation/token_count.sh: -------------------------------------------------------------------------------- 1 | python scripts/token_counts.py \ 2 | --data_file data/binarized_text.electra.pickle \ 3 | --token_counts_dump data/token_counts.electra.pickle \ 4 | --vocab_size 30522 & -------------------------------------------------------------------------------- /distillation/training_configs/distilbert-base-cased.json: -------------------------------------------------------------------------------- 1 | { 2 | "activation": "gelu", 3 | "attention_dropout": 0.1, 4 | "dim": 768, 5 | "dropout": 0.1, 6 | "hidden_dim": 3072, 7 | "initializer_range": 0.02, 8 | "max_position_embeddings": 512, 9 | "n_heads": 12, 10 | "n_layers": 6, 11 | "sinusoidal_pos_embds": true, 12 | "tie_weights_": true, 13 | "vocab_size": 28996 14 | } 15 | -------------------------------------------------------------------------------- /distillation/training_configs/distilbert-base-multilingual-cased.json: -------------------------------------------------------------------------------- 1 | { 2 | "activation": "gelu", 3 | "attention_dropout": 0.1, 4 | "dim": 768, 5 | "dropout": 0.1, 6 | "hidden_dim": 3072, 7 | "initializer_range": 0.02, 8 | "max_position_embeddings": 512, 9 | "n_heads": 12, 10 | "n_layers": 6, 11 | "sinusoidal_pos_embds": true, 12 | "tie_weights_": true, 13 | "vocab_size": 119547 14 | } 15 | -------------------------------------------------------------------------------- /distillation/training_configs/distilbert-base-uncased.json: -------------------------------------------------------------------------------- 1 | { 2 | "activation": "gelu", 3 | "attention_dropout": 0.1, 4 | "dim": 768, 5 | "dropout": 0.1, 6 | "hidden_dim": 3072, 7 | "initializer_range": 0.02, 8 | "max_position_embeddings": 512, 9 | "n_heads": 12, 10 | "n_layers": 6, 11 | "sinusoidal_pos_embds": true, 12 | "tie_weights_": true, 13 | "vocab_size": 30522 14 | } 15 | -------------------------------------------------------------------------------- /distillation/training_configs/distilelectra.json: -------------------------------------------------------------------------------- 1 | { 2 | "architectures": [ 3 | "ElectraForPreTraining" 4 | ], 5 | "attention_probs_dropout_prob": 0.1, 6 | "embedding_size": 768, 7 | "hidden_act": "gelu", 8 | "hidden_dropout_prob": 0.1, 9 | "hidden_size": 768, 10 | "initializer_range": 0.02, 11 | "intermediate_size": 3072, 12 | "layer_norm_eps": 1e-12, 13 | "max_position_embeddings": 512, 14 | "model_type": "electra", 15 | "num_attention_heads": 12, 16 | "num_hidden_layers": 6, 17 | "pad_token_id": 0, 18 | "position_embedding_type": "absolute", 19 | "summary_activation": "gelu", 20 | "summary_last_dropout": 0.1, 21 | "summary_type": "first", 22 | "summary_use_proj": true, 23 | "transformers_version": "4.6.0.dev0", 24 | "type_vocab_size": 2, 25 | "vocab_size": 30522 26 | } 27 | -------------------------------------------------------------------------------- /distillation/training_configs/distilgpt2.json: -------------------------------------------------------------------------------- 1 | { 2 | "initializer_range": 0.02, 3 | "layer_norm_epsilon": 0.00001, 4 | "n_embd": 768, 5 | "n_head": 12, 6 | "n_layer": 6, 7 | "n_positions": 1024, 8 | "vocab_size": 50257 9 | } -------------------------------------------------------------------------------- /distillation/training_configs/distilroberta-base.json: -------------------------------------------------------------------------------- 1 | { 2 | "vocab_size": 50265, 3 | "hidden_size": 768, 4 | "num_hidden_layers": 6, 5 | "num_attention_heads": 12, 6 | "intermediate_size": 3072, 7 | "hidden_act": "gelu", 8 | "hidden_dropout_prob": 0.1, 9 | "attention_probs_dropout_prob": 0.1, 10 | "max_position_embeddings": 514, 11 | "type_vocab_size": 1, 12 | "initializer_range": 0.02, 13 | "layer_norm_eps": 0.00001 14 | } -------------------------------------------------------------------------------- /docker/README.md: -------------------------------------------------------------------------------- 1 | # Docker container for demo 2 | To use this library alongside docker, one has to install [docker](https://docs.docker.com/engine/install/ubuntu/) and [nvidia-docker](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html) to run the container with GPU. 3 | To build a docker image, run: 4 | ``` 5 | ./build.sh 6 | ``` 7 | To run the container locally with CUDA support, run: 8 | ``` 9 | docker run -it -p 8888:8888 --gpus all acleto_demo 10 | ``` 11 | To run the demo notebook, open the jupyter notebook and choose a file for your task in ./acleto/jupyterlab_demo. Note that currently only jupyter notebooks are supported - demo in jupyter lab isn't tested. -------------------------------------------------------------------------------- /docker/build.sh: -------------------------------------------------------------------------------- 1 | rsync -av --progress ../ ./acleto --exclude docker --exclude .git 2 | docker build -t acleto_demo ./ 3 | -------------------------------------------------------------------------------- /docker/build_cuda113.sh: -------------------------------------------------------------------------------- 1 | rsync -av --progress ../ ./acleto --exclude docker --exclude .git 2 | docker build -t acleto_demo_cuda113 -f Dockerfile_cuda113 ./ 3 | -------------------------------------------------------------------------------- /docker/cache.sh: -------------------------------------------------------------------------------- 1 | HYDRA_CONFIG_PATH=~/acleto/jupyterlab_demo/configs \ 2 | HYDRA_CONFIG_NAME=al_ner \ 3 | python ~/acleto/acleto/al4nlp/utils/cache_all_necessary_files.py \ 4 | data.dataset_name=conll2003 acquisition_model.checkpoint=distilbert-base-cased cache_model_and_dataset=True cache_dir=~/acleto/jupyterlab_demo/cache/ner 5 | 6 | 7 | HYDRA_CONFIG_PATH=~/acleto/jupyterlab_demo/configs \ 8 | HYDRA_CONFIG_NAME=al_cls \ 9 | python ~/acleto/acleto/al4nlp/utils/cache_all_necessary_files.py \ 10 | data.dataset_name=ag_news acquisition_model.checkpoint=distilbert-base-uncased cache_model_and_dataset=True cache_dir=~/acleto/jupyterlab_demo/cache/cls 11 | -------------------------------------------------------------------------------- /docker/initial-condarc: -------------------------------------------------------------------------------- 1 | # Conda configuration see https://conda.io/projects/conda/en/latest/configuration.html 2 | 3 | auto_update_conda: false 4 | show_channel_urls: true 5 | channels: 6 | - conda-forge 7 | -------------------------------------------------------------------------------- /docker/jupyter_notebook_config.json: -------------------------------------------------------------------------------- 1 | { 2 | "NotebookApp": { 3 | "password": "sha1:3b4b6378355:642693b20f0a33bcad27b94293d0ed7db3408322" 4 | } 5 | } 6 | -------------------------------------------------------------------------------- /docker/start-notebook.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | # Copyright (c) Jupyter Development Team. 3 | # Distributed under the terms of the Modified BSD License. 4 | 5 | set -e 6 | 7 | # The Jupyter command to launch 8 | # JupyterLab by default 9 | DOCKER_STACKS_JUPYTER_CMD="${DOCKER_STACKS_JUPYTER_CMD:=lab}" 10 | 11 | if [[ -n "${JUPYTERHUB_API_TOKEN}" ]]; then 12 | echo "WARNING: using start-singleuser.sh instead of start-notebook.sh to start a server associated with JupyterHub." 13 | exec /usr/local/bin/start-singleuser.sh "$@" 14 | fi 15 | 16 | wrapper="" 17 | if [[ "${RESTARTABLE}" == "yes" ]]; then 18 | wrapper="run-one-constantly" 19 | fi 20 | 21 | # shellcheck disable=SC1091,SC2086 22 | exec /usr/local/bin/start.sh ${wrapper} jupyter ${DOCKER_STACKS_JUPYTER_CMD} ${NOTEBOOK_ARGS} "$@" 23 | -------------------------------------------------------------------------------- /docker/start-singleuser.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | # Copyright (c) Jupyter Development Team. 3 | # Distributed under the terms of the Modified BSD License. 4 | 5 | set -e 6 | 7 | # set default ip to 0.0.0.0 8 | if [[ "${NOTEBOOK_ARGS} $*" != *"--ip="* ]]; then 9 | NOTEBOOK_ARGS="--ip=0.0.0.0 ${NOTEBOOK_ARGS}" 10 | fi 11 | 12 | # shellcheck disable=SC1091,SC2086 13 | . /usr/local/bin/start.sh jupyterhub-singleuser ${NOTEBOOK_ARGS} "$@" 14 | -------------------------------------------------------------------------------- /docs/Makefile: -------------------------------------------------------------------------------- 1 | # Minimal makefile for Sphinx documentation 2 | # 3 | 4 | # You can set these variables from the command line, and also 5 | # from the environment for the first two. 6 | SPHINXOPTS ?= 7 | SPHINXBUILD ?= sphinx-build 8 | SOURCEDIR = . 9 | BUILDDIR = _build 10 | 11 | # Put it first so that "make" without argument is like "make help". 12 | help: 13 | @$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) 14 | 15 | .PHONY: help Makefile 16 | 17 | # Catch-all target: route all unknown targets to Sphinx using the new 18 | # "make mode" option. $(O) is meant as a shortcut for $(SPHINXOPTS). 19 | %: Makefile 20 | @$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) 21 | -------------------------------------------------------------------------------- /docs/Readme.md: -------------------------------------------------------------------------------- 1 | # Documentation 2 | 3 | The docs can be built by using the Makefile: 4 | 5 | ``` 6 | make html 7 | ``` -------------------------------------------------------------------------------- /docs/basic_usage.rst: -------------------------------------------------------------------------------- 1 | .. _basic_usage: 2 | 3 | ================ 4 | Usage 5 | ================ 6 | 7 | The `configs` folder contains config files with general settings. The `experiments` folder contains config files with experimental design. To run an experiment with a chosen configuration, specify config file name in `HYDRA_CONFIG_NAME` variable and run `train.sh` script (see `./examples/al` for details). 8 | 9 | For example to launch PLASM on AG-News with ELECTRA as a successor model: 10 | 11 | 12 | .. code-block:: console 13 | 14 | cd PATH_TO_THIS_REPO 15 | HYDRA_CONFIG_PATH=../experiments/ag_news HYDRA_EXP_CONFIG_NAME=ag_plasm python active_learning/run_tasks_on_multiple_gpus.py 16 | -------------------------------------------------------------------------------- /docs/config_structure.rst: -------------------------------------------------------------------------------- 1 | .. _config_structure: 2 | 3 | ========================== 4 | Config structure explained 5 | ========================== 6 | 7 | - `cuda_devices`: list of CUDA devices to use: one experiment on one CUDA device. `cuda_devices=[0,1]` means using zero-th and first devices. 8 | - `config_name`: name of config from **configs** folder with general settings: dataset, experiment setting (e.g. LC/ASM/PLASM), model checkpoints, hyperparameters etc. 9 | - `config_path`: path to config with general settings. 10 | - `command`: **.py** file to run. For AL experiments, use **run_active_learning.py**. 11 | - `args`: arguments to modify from a general config in the current experiment. `acquisition_model.name=xlnet-base-cased` means that _xlnet-base-cased_ will be used as an acquisition model. 12 | - `seeds`: random seeds to use. `seeds=[4837, 23419]` means that two separate experiments with the same settings (except for **seed**) will be run: one with **seed == 4837**, one with **seed == 23419**. 13 | -------------------------------------------------------------------------------- /docs/datasets.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/docs/datasets.rst -------------------------------------------------------------------------------- /docs/gui.rst: -------------------------------------------------------------------------------- 1 | .. _gui: 2 | 3 | === 4 | GUI 5 | === 6 | 7 | Our framework provides a serverless GUI annotation tool integrated into the Jupyter IDE: 8 | 9 | 10 | .. image:: _static/figures/gui.svg 11 | -------------------------------------------------------------------------------- /docs/install.rst: -------------------------------------------------------------------------------- 1 | .. _installation: 2 | 3 | ============ 4 | Installation 5 | ============ 6 | 7 | You can easily install AlToolBox using pip: 8 | 9 | 10 | .. code-block:: console 11 | 12 | pip install acleto 13 | 14 | To annotate instances for active learning in Jupyter Notebook or Jupyter Lab one have to install additional widget after framework installation. In case of Jupyter Notebook usage run: 15 | 16 | .. code-block:: console 17 | 18 | jupyter nbextension install --py --symlink --sys-prefix text_selector 19 | jupyter nbextension enable --py --sys-prefix text_selector 20 | 21 | In case of Jupyter Lab usage run: 22 | 23 | .. code-block:: console 24 | 25 | jupyter labextension install js 26 | jupyter labextension install text_selector -------------------------------------------------------------------------------- /docs/make.bat: -------------------------------------------------------------------------------- 1 | @ECHO OFF 2 | 3 | pushd %~dp0 4 | 5 | REM Command file for Sphinx documentation 6 | 7 | if "%SPHINXBUILD%" == "" ( 8 | set SPHINXBUILD=sphinx-build 9 | ) 10 | set SOURCEDIR=. 11 | set BUILDDIR=_build 12 | 13 | %SPHINXBUILD% >NUL 2>NUL 14 | if errorlevel 9009 ( 15 | echo. 16 | echo.The 'sphinx-build' command was not found. Make sure you have Sphinx 17 | echo.installed, then set the SPHINXBUILD environment variable to point 18 | echo.to the full path of the 'sphinx-build' executable. Alternatively you 19 | echo.may add the Sphinx directory to PATH. 20 | echo. 21 | echo.If you don't have Sphinx installed, grab it from 22 | echo.https://www.sphinx-doc.org/ 23 | exit /b 1 24 | ) 25 | 26 | if "%1" == "" goto help 27 | 28 | %SPHINXBUILD% -M %1 %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% 29 | goto end 30 | 31 | :help 32 | %SPHINXBUILD% -M help %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% 33 | 34 | :end 35 | popd 36 | -------------------------------------------------------------------------------- /docs/models.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/docs/models.rst -------------------------------------------------------------------------------- /docs/post_processing.rst: -------------------------------------------------------------------------------- 1 | .. _post_processing: 2 | 3 | =============== 4 | Post processing 5 | =============== 6 | 7 | * PLASM postprocessing pipeline for annotated data reusability. 8 | * Acquisition model distillation. 9 | * Domain adaptation of acquisition models. -------------------------------------------------------------------------------- /docs/testing.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/docs/testing.rst -------------------------------------------------------------------------------- /domain_adaptation/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/domain_adaptation/__init__.py -------------------------------------------------------------------------------- /domain_adaptation/train.sh: -------------------------------------------------------------------------------- 1 | export HYDRA_CONFIG_PATH=./configs 2 | export HYDRA_CONFIG_NAME=domain_adaptation.yaml 3 | export output_dir="bert-conll" 4 | export dataset_name="'conll2003'" 5 | 6 | python ./hf_dataset_to_sent.py \ 7 | output_dir=$output_dir \ 8 | data.dataset_name=$dataset_name \ 9 | data.text_name=tokens \ 10 | data.label_name=ner_tags \ 11 | data.source_task=ner \ 12 | 13 | python ./run_lm.py \ 14 | output_dir=$output_dir \ 15 | data.dataset_name=$dataset_name \ 16 | weight_decay=1e-3 \ 17 | learning_rate=1e-6 \ 18 | gradient_accumulation_steps=4 \ 19 | gpus=[0] \ 20 | n_gpu=1 \ 21 | max_steps=200 -------------------------------------------------------------------------------- /examples/al/config.yaml: -------------------------------------------------------------------------------- 1 | cuda_devices: [0,1,2] 2 | task_names: '' 3 | hydra: 4 | run: 5 | dir: ./workdir/logs/${now:%Y-%m-%d}/${now:%H-%M-%S} 6 | 7 | tasks: 8 | - 9 | name: 'distilbert' 10 | config_path: '../acleto/al_benchmark/configs' 11 | config_name: 'al_cls' 12 | command: '../../scripts/run_active_learning.py' 13 | environ: '' 14 | args: 'data.dataset_name=bbc_news data.path=../../data acquisition_model.name=distilbert-base-uncased' 15 | seeds: [4837, 23419, 705525, 10671619, 1084218] 16 | n_repeats: 1 -------------------------------------------------------------------------------- /examples/al/train.sh: -------------------------------------------------------------------------------- 1 | HYDRA_CONFIG_PATH=../examples/al python '../../scripts/run_tasks_on_multiple_gpus.py' -------------------------------------------------------------------------------- /examples/cache_necessary_files/run.sh: -------------------------------------------------------------------------------- 1 | HYDRA_CONFIG_PATH=../../configs \ 2 | HYDRA_CONFIG_NAME=al_cls \ 3 | python ../../active_learning/utils/cache_all_necessary_files.py \ 4 | data.dataset_name=ag_news acquisition_model.name=distilbert-base-uncased cache_model_and_dataset=True -------------------------------------------------------------------------------- /examples/custom_strategy/__pycache__/least_confidence.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/examples/custom_strategy/__pycache__/least_confidence.cpython-38.pyc -------------------------------------------------------------------------------- /examples/custom_strategy/__pycache__/top_from_previous_iteration_subsampling.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/examples/custom_strategy/__pycache__/top_from_previous_iteration_subsampling.cpython-38.pyc -------------------------------------------------------------------------------- /examples/custom_strategy/least_confidence.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | 3 | 4 | def least_confidence(model, X_pool, n_instances, **kwargs): 5 | probas = model.predict_proba(X_pool) 6 | uncertainty_estimates = 1 - probas.max(axis=1) 7 | query_idx = np.argsort(-uncertainty_estimates)[:n_instances] 8 | query = X_pool.select(query_idx) 9 | return query_idx, query, uncertainty_estimates 10 | -------------------------------------------------------------------------------- /examples/custom_strategy/top_from_previous_iteration_subsampling.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | 3 | 4 | def top_from_previous_iteration_subsampling( 5 | uncertainty_estimates, gamma_or_k_confident_to_save, **kwargs 6 | ): 7 | if isinstance(gamma_or_k_confident_to_save, float): 8 | gamma_or_k_confident_to_save = int( 9 | gamma_or_k_confident_to_save * len(uncertainty_estimates) 10 | ) 11 | argsort = np.argsort(-uncertainty_estimates) 12 | return argsort[:gamma_or_k_confident_to_save] 13 | -------------------------------------------------------------------------------- /hooks/README.md: -------------------------------------------------------------------------------- 1 | # Hooks 2 | 3 | To use pre-push hook for tests, copy it to your .git/hooks directory. After that this hook will be called on each push to the specified in the hook branch. -------------------------------------------------------------------------------- /hooks/pre-push: -------------------------------------------------------------------------------- 1 | #!/bin/sh 2 | 3 | remote="$1" 4 | url="$2" 5 | 6 | branch="$(git rev-parse --abbrev-ref HEAD)" 7 | #echo $branch 8 | # run tests and calculate coverage 9 | if [ $branch != "dev" ] 10 | then 11 | cd tests 12 | coverage run --source ../acleto,../distillation,../domain_adaptation -m pytest ./full_tests.py --junitxml=../reports/junit/junit.xml 13 | cd .. 14 | if [ $? -eq 0 ] 15 | then 16 | # generate custom badges, so we dont't use any CI service 17 | coverage xml -o reports/coverage/coverage.xml 18 | genbadge tests --output-file reports/junit/tests-badge.svg 19 | genbadge coverage --output-file reports/coverage/coverage-badge.svg 20 | # add generated badges to the latest commit 21 | git add ./reports 22 | git commit --amend --no-edit 23 | printf "Successful: Tests\n" 24 | printf "Don't forget to make a git pull\n" 25 | else 26 | printf "Error: Tests\n" 27 | exit 1 28 | fi 29 | else 30 | printf "This branch doesn't require tests\n" 31 | fi 32 | exit 0 33 | -------------------------------------------------------------------------------- /init.sh: -------------------------------------------------------------------------------- 1 | python nltk_download.py 2 | pip install jupyter_contrib_nbextensions 3 | pip install git+https://github.com/IINemo/text_selector.git 4 | pip install --upgrade fairscale==0.4.0 5 | mkdir ./acleto/al4nlp/utils/packages 6 | cd ./acleto/al4nlp/utils/packages 7 | git clone https://github.com/Aktsvigun/summac 8 | git clone https://github.com/neulab/BARTScore.git 9 | cp BARTScore/bart_score.py ./ 10 | rm -rf BARTScore 11 | rm -rf summac/.git -------------------------------------------------------------------------------- /jupyterlab_demo/configs/default.py: -------------------------------------------------------------------------------- 1 | MAX_LEN = 475 2 | PRED_BATCH_SIZE = 8 3 | 4 | N_SAMPLES_PER_AL_ITER = 30 5 | MAX_SAMPLES_NUMBER = 30 6 | AL_STRATEGY = "mnlp" 7 | MNLP_SCORE_POSITIVES = True 8 | 9 | N_EPOCHS = 4 10 | BATCH_SIZE = 2 11 | GRAD_ACCUM_STEPS = 2 12 | LEARNING_RATE = 5e-5 13 | VALIDATION_RATIO = 0.01 14 | PATIENCE = 1 15 | 16 | BERT_MODEL_TYPE = "distilbert-base-cased" 17 | CACHE_DIR = "/notebook/models/cache" 18 | 19 | RANDOM_STATE = 42 20 | 21 | UPSAMPLE_POSITIVE = 1024 22 | IS_SYNC = False 23 | 24 | SUBSAMPLE = 0.1 25 | AUTOFILL_SIMILAR = True 26 | 27 | CONLL_MODE = True 28 | -------------------------------------------------------------------------------- /jupyterlab_demo/configs/framework/allennlp.yaml: -------------------------------------------------------------------------------- 1 | name: allennlp 2 | validation_metric_cls: '+accuracy' 3 | validation_metric_ner: '+f1-measure-overall' -------------------------------------------------------------------------------- /jupyterlab_demo/configs/framework/flair.yaml: -------------------------------------------------------------------------------- 1 | name: flair 2 | validation_metric_cls: accuracy 3 | validation_metric_ner: overall_f1 -------------------------------------------------------------------------------- /jupyterlab_demo/configs/framework/pytorch.yaml: -------------------------------------------------------------------------------- 1 | name: pytorch 2 | validation_metric_cls: accuracy 3 | validation_metric_ner: overall_f1 -------------------------------------------------------------------------------- /jupyterlab_demo/configs/framework/transformers.yaml: -------------------------------------------------------------------------------- 1 | name: transformers 2 | validation_metric_cls: accuracy 3 | validation_metric_ner: overall_f1 -------------------------------------------------------------------------------- /jupyterlab_demo/data/.gitkeep: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/jupyterlab_demo/data/.gitkeep -------------------------------------------------------------------------------- /jupyterlab_demo/data/CLS/ag_news/tags.json: -------------------------------------------------------------------------------- 1 | ["World", "Sports", "Business", "Sci/Tech"] -------------------------------------------------------------------------------- /jupyterlab_demo/data/NER/conll2003/tags.json: -------------------------------------------------------------------------------- 1 | ["O", "B-PER", "I-PER", "B-ORG", "I-ORG", "B-LOC", "I-LOC", "B-MISC", "I-MISC"] -------------------------------------------------------------------------------- /jupyterlab_demo/logs/.gitkeep: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/jupyterlab_demo/logs/.gitkeep -------------------------------------------------------------------------------- /nltk_download.py: -------------------------------------------------------------------------------- 1 | import nltk 2 | 3 | nltk.download("punkt") 4 | nltk.download("stopwords") 5 | -------------------------------------------------------------------------------- /scripts/calculate_overlap_among_queries.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/scripts/calculate_overlap_among_queries.py -------------------------------------------------------------------------------- /tests/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/WSNLP/al_toolbox/7d265ee3babe742232ac23d49ecbe4ac761027fe/tests/__init__.py --------------------------------------------------------------------------------