├── __init__.py ├── box ├── README.md ├── test.scala └── train.scala ├── README.md ├── stmt.py └── LICENSE /__init__.py: -------------------------------------------------------------------------------- 1 | from .stmt import STMT 2 | -------------------------------------------------------------------------------- /box/README.md: -------------------------------------------------------------------------------- 1 | Download STMT from [here](http://nlp.stanford.edu/software/tmt/tmt-0.4/) and put it in this directory. 2 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # topbox 2 | A small Python 3 wrapper around the Stanford Topic Modeling Toolbox (STMT) that makes working with L-LDA a bit easier; no need to leave the Python environment. More information on its workings can be found on [my blog](https://cmry.github.io/notes/topbox). 3 | 4 | # Setting up 5 | 6 | Just [download](http://nlp.stanford.edu/software/tmt/tmt-0.4/tmt-0.4.0.jar) STMT and put it in the `box` directory. After, import `topbox` from wherever you left it. 7 | 8 | On Linux, this would look something like this: 9 | 10 | ``` shell 11 | $ cd ~ 12 | $ git clone https://github.com/cmry/topbox 13 | $ cd ~/topbox/box 14 | $ wget http://nlp.stanford.edu/software/tmt/tmt-0.4/tmt-0.4.0.jar 15 | $ cd ~ 16 | $ vi some_topbox_script.py 17 | ``` 18 | 19 | You can paste the code below in the script file to test if it's working. 20 | 21 | # Example 22 | 23 | ``` python 24 | import topbox 25 | 26 | stmt = topbox.STMT('bit_of_testing', epochs=10, mem=15000) 27 | 28 | 29 | space = ['text text more text', 'things to do with text'] 30 | labels = ['label1 label2', 'label1 label3'] 31 | 32 | stmt.train(space, labels) 33 | 34 | 35 | infer = ['this is a text', 'things with more text'] 36 | gs = ['label1 label2', 'label1 label3'] 37 | 38 | stmt.test(infer, gs) 39 | 40 | 41 | from sklearn.metrics import average_precision_score 42 | 43 | # array requires numpy and scipy 44 | y_true, y_score = stmt.results(gs, array=True) 45 | 46 | print(average_precision_score(y_true, y_score)) 47 | ``` 48 | -------------------------------------------------------------------------------- /box/test.scala: -------------------------------------------------------------------------------- 1 | import scalanlp.io._; 2 | import scalanlp.stage._; 3 | import scalanlp.stage.text._; 4 | import scalanlp.text.tokenize._; 5 | import scalanlp.pipes.Pipes.global._; 6 | 7 | import edu.stanford.nlp.tmt.stage._; 8 | import edu.stanford.nlp.tmt.model.lda._; 9 | import edu.stanford.nlp.tmt.model.llda._; 10 | 11 | val modelPath = file("modelfolder"); 12 | 13 | println("Loading "+modelPath); 14 | val model = LoadCVB0LabeledLDA(modelPath).asCVB0LDA; 15 | val source = CSVFile("datafile.csv") ~> IDColumn(1); 16 | 17 | val text = { 18 | source ~> // read from the source file 19 | Column(3) ~> // select column containing text 20 | TokenizeWith(model.tokenizer.get) // tokenize with tokenizer above 21 | } 22 | 23 | val output = file(modelPath, source.meta[java.io.File].getName.replaceAll(".csv","")); 24 | val dataset = LDADataset(text, model.termIndex); 25 | 26 | println("Writing document distributions to "+output+"-document-topic-distributions-res.csv"); 27 | val perDocTopicDistributions = InferCVB0DocumentTopicDistributions(model, dataset); 28 | CSVFile(output+"-document-topic-distributions-res.csv").write(perDocTopicDistributions); 29 | 30 | // println("Writing topic usage to "+output+"-usage-res.csv"); 31 | // val usage = QueryTopicUsage(model, dataset, perDocTopicDistributions); 32 | // CSVFile(output+"-usage-res.csv").write(usage); 33 | 34 | // println("Estimating per-doc per-word topic distributions"); 35 | // val perDocWordTopicDistributions = EstimatePerWordTopicDistributions( 36 | // model, dataset, perDocTopicDistributions); 37 | // CSVFile(output+"-document-word-topic-distributions.csv").write(perDocWordTopicDistributions); 38 | 39 | // println("Writing top terms to "+output+"-top-terms.csv"); 40 | // val topTerms = QueryTopTerms(model, dataset, perDocWordTopicDistributions, numTopTerms=50); 41 | // CSVFile(output+"-top-terms.csv").write(topTerms); 42 | 43 | -------------------------------------------------------------------------------- /box/train.scala: -------------------------------------------------------------------------------- 1 | // Stanford TMT Example 6 - Training a LabeledLDA model 2 | // http://nlp.stanford.edu/software/tmt/0.4/ 3 | 4 | // tells Scala where to find the TMT classes 5 | import scalanlp.io._; 6 | import scalanlp.stage._; 7 | import scalanlp.stage.text._; 8 | import scalanlp.text.tokenize._; 9 | import scalanlp.pipes.Pipes.global._; 10 | import edu.stanford.nlp.tmt.stage._; 11 | import edu.stanford.nlp.tmt.model.lda._; 12 | import edu.stanford.nlp.tmt.model.llda._; 13 | 14 | val source = CSVFile("datafile.csv") ~> IDColumn(1); 15 | 16 | val tokenizer = { 17 | SimpleEnglishTokenizer() ~> // tokenize on space and punctuation 18 | CaseFolder() ~> // lowercase everything 19 | WordsAndNumbersOnlyFilter() ~> // ignore non-words and non-numbers 20 | MinimumLengthFilter(1) // take terms with >=3 characters 21 | } 22 | 23 | val text = { 24 | source ~> // read from the source file 25 | Column(3) ~> // select column containing text 26 | TokenizeWith(tokenizer) ~> // tokenize with tokenizer above 27 | TermCounter() ~> // collect counts (needed below) 28 | TermMinimumDocumentCountFilter(1) ~> // filter terms in <4 docs 29 | TermDynamicStopListFilter(0) ~> // filter out 30 most common terms 30 | DocumentMinimumLengthFilter(1) // take only docs with >=5 terms 31 | } 32 | 33 | // define fields from the dataset we are going to slice against 34 | val labels = { 35 | source ~> // read from the source file 36 | Column(2) ~> // take column two, the year 37 | TokenizeWith(WhitespaceTokenizer()) ~> // turns label field into an array 38 | TermCounter() ~> // collect label counts 39 | TermMinimumDocumentCountFilter(0) // filter labels in < 10 docs 40 | } 41 | 42 | val dataset = LabeledLDADataset(text, labels); 43 | 44 | // define the model parameters 45 | val modelParams = LabeledLDAModelParams(dataset=dataset); 46 | 47 | // Name of the output model folder to generate 48 | val modelPath = file("modelfolder"); 49 | 50 | // Trains the model, writing to the given output path 51 | TrainCVB0LabeledLDA(modelParams, dataset, output = modelPath, maxIterations = 5); 52 | // or could use TrainGibbsLabeledLDA(modelParams, dataset, output = modelPath, maxIterations = 1500); 53 | -------------------------------------------------------------------------------- /stmt.py: -------------------------------------------------------------------------------- 1 | """Python 2 & 3 wrapper around the Stanford Topic Modeling Toolbox.""" 2 | 3 | from csv import writer, reader 4 | from re import sub 5 | from subprocess import call 6 | from os import path, remove, sep 7 | from shutil import rmtree 8 | from glob import glob 9 | from inspect import isgenerator 10 | from sys import version_info 11 | 12 | # Authors: Chris Emmery 13 | # References: Ramage, Hall, Nallapati, Manning (2009) 14 | # License: BSD 3-Clause 15 | # pylint: disable=C0103 16 | 17 | 18 | class STMT(object): 19 | """Stanford Topic Modelling Toolbox Wrapper. 20 | 21 | This is a wrapper Class around the Stanford Topic Modelling Toolbox. It 22 | assumes that you have your vector space in your code, and don't want to 23 | bother with the `csv -> scala -> java -> csv -> extract results` process. 24 | It therefore compresses all of this in a few class interactions. Basically, 25 | you create model by initiating it with a name, set the amount of epochs 26 | and memory as desired, and then start training and testing on data that 27 | you have in Python code. After, the class can handle extracting the correct 28 | results (even in sklearn format), as well as cleaning up once you're done. 29 | Some examples of this will be given below, more information can be found 30 | on https://cmry.github.io/notes/topbox. 31 | 32 | Parameters 33 | ---------- 34 | name : string 35 | The name that will be appended to all the saved files. If you want to 36 | keep the trained model, this name can be used to load it back in. 37 | 38 | epochs : integer, optional, default 20 39 | The amount of iterations you want L-LDA to train and sample; if you 40 | run into some errors, it's a good idea to set this to 1 to save time 41 | whilst debugging. 42 | 43 | mem : integer, optional, default 7000 44 | The amount of memory (in MB) that the model will use. By default it 45 | assumes that you have 8G of memory, so it will account for 1G of os 46 | running. Should be comfortable; adjust if running into OutOfMemory 47 | errors though. 48 | 49 | keep : boolean, optional, default True 50 | If set to False, will remove the data and scala files after training, 51 | and will remove EVERYTHING after the resutls are obtained. This can 52 | be handy when running a quick topic model and save disk space. If 53 | you're running a big model and want to keep it after your session is 54 | done, it might be better to just leave it to True. 55 | 56 | Attributes 57 | ---------- 58 | dir : string 59 | Absolute path where the storage area of the topbox is located. 60 | 61 | Examples 62 | -------- 63 | train = [['sports football', 'this talks about football, or soccer, 64 | with a goal and a ball'], 65 | ['sports rugby', 'here we have some document where we do a scrum 66 | and kick the ball'], 67 | ['music concerts', 'a venue with loud music and a stage'], 68 | ['music instruments', 'thing that have strings or keys, or 69 | whatever']] 70 | 71 | test = [['music', 'the stage was full of string things'], 72 | ['sports', 'we kick a ball around'], 73 | ['rugby', 'now add some confusing sentence with novel words what is 74 | happening']] 75 | 76 | import topbox 77 | 78 | stmt = topbox.STMT('test_model') 79 | stmt = topbox.STMT('test_model', epochs=400, mem=14000) 80 | 81 | train_labels, train_space = zip(*train) 82 | test_labels, test_space = zip(*test) 83 | 84 | stmt.train(train_space, train_labels) 85 | stmt.test(test_space, test_labels) 86 | 87 | y_true, y_score = stmt.results(test_labels, array=True) 88 | 89 | from sklearn.metrics import average_precision_score 90 | average_precision_score(y_true, y_score) 91 | 92 | Notes 93 | ----- 94 | The code and scale examples are obtained from the Stanford website 95 | (http://nlp.stanford.edu/software/tmt/tmt-0.4/). Their code thusly exists 96 | in this repository under equal license. Please respect this. 97 | """ 98 | 99 | def __init__(self, name, epochs=20, mem=7000, keep=True): 100 | """Set paths and variables.""" 101 | self.dir = path.normpath(path.dirname(path.realpath(__file__)) + \ 102 | '{0}box'.format(sep)) + sep 103 | self.name = name 104 | self.keep = keep 105 | self.epochs = epochs 106 | self.mem = mem 107 | 108 | def boot(self, mod): 109 | """Boot script. 110 | 111 | Alters the directories in the .scala files for running and testing 112 | L-LDA (depending on the `mod`). Uses a generic call on the .jar that 113 | STMT resides in. 114 | 115 | Parameters 116 | ---------- 117 | :mod: string 118 | Either 'test' or 'train' for swithing states. 119 | """ 120 | self.scala(mod) 121 | call(["java", "-Xmx" + str(self.mem) + "m", "-jar", self.dir + 122 | "tmt-0.4.0.jar", self.dir + self.name + "_" + mod + ".scala"]) 123 | self.scala(mod, 1) 124 | 125 | def store(self, space, labels, vsp_type): 126 | """Data to csv storage. 127 | 128 | 129 | Stores a given (sub)vectorspace to the .csv format that STMT works 130 | with. The space should be a dict where the key is a tuple with (int, 131 | str), where int is the index number and str the document its topic 132 | labels seperated by a whitespace. The value is your vector stored in 133 | a list. 134 | 135 | If you want to iteratively construct a space, provide a generator that 136 | will feed batches of the space. 137 | 138 | Parameters 139 | ---------- 140 | space : list 141 | The vector space; a list with text. 142 | 143 | labels : list 144 | List with labels where each index corresponds to the text in space. 145 | 146 | vps_type : string 147 | Either train or test as appendix for the filename. 148 | """ 149 | csv_file = open("%s%s_%s.csv" % (self.dir, self.name, vsp_type), 'a') 150 | csv_writer = writer(csv_file) 151 | for i, zipped in enumerate(zip(labels, space)): 152 | line = [str(i + 1), zipped[0], zipped[1]] 153 | if version_info.major < 3: # fix py2 compat 154 | line = [i.encode('utf8') for i in line] 155 | csv_writer.writerow(line) 156 | csv_file.close() 157 | 158 | def regex(self, f, needle, rock): 159 | """File name replacer. 160 | 161 | Function is used to flip the read object file (original .scale file) 162 | and write replaced cotents to this newly created file. 163 | 164 | Parameters 165 | ---------- 166 | f : string 167 | Contents of the original .scala file. 168 | 169 | needle : string 170 | String sequence to be replaced in the original .scala file. 171 | 172 | rock : string 173 | Basically the .read() contents of the original .scala file. 174 | """ 175 | wf = (self.name + '_').join(f.rsplit('_', 1)) 176 | f = ''.join(f.rsplit('_', 1)) 177 | try: 178 | rf = open(wf, 'r') 179 | except IOError: 180 | rf = open(f, 'r') 181 | stack = sub(needle, rock, rf.read()) 182 | rf.close() 183 | with open(wf, 'w') as wf: 184 | wf.write(stack) 185 | 186 | def scala(self, s, r=False): 187 | """Scala code replacer. 188 | 189 | Handles the .scala text replacements. In the basefiles, the replace 190 | targets are `modelfile` by default. This can also be used to flip 191 | number of the iterations. 192 | 193 | Parameters 194 | ---------- 195 | s : string 196 | Has the value of either train or test in the framework. 197 | 198 | r : boolean, optional, default False 199 | Indicates old to new replace by default. 200 | """ 201 | prep, std = 'maxIterations = ', '5' 202 | orig, new = 'modelfolder', self.dir + self.name + '_' + 'train' 203 | o_csv, n_csv = 'datafile.csv', self.dir + self.name + '_' + s + '.csv' 204 | f = self.dir + '_' + s + '.scala' 205 | self.regex(f, o_csv, n_csv) if not r else self.regex(f, n_csv, o_csv) 206 | self.regex(f, orig, new) if not r else self.regex(f, new, orig) 207 | self.regex(f, prep + std, prep + ' ' + str(self.epochs)) if \ 208 | self.epochs else self.regex(f, prep + std, prep + std) 209 | 210 | def m_incidence(self, predicted_row, label_index, gold_standard): 211 | """Matrix to Incidence. 212 | 213 | Extracts the probabilities from the .csvs, and generates an incidence 214 | vector based on the correct topic labels. If a value is 'NaN', it will 215 | be skipped (model might have crapped up somewhere). The result is a 216 | zipped matrix with tuple values giving (incidence, probability). 217 | 218 | Parameters 219 | ---------- 220 | predicted_row : list 221 | Predicted row in the .csv file. 222 | 223 | label_index : list 224 | Lookup list for topics on index number. 225 | 226 | gold_standard : list 227 | Lookup list for correct topics per document. 228 | 229 | Return 230 | ------ 231 | vector : list of lists 232 | Incidence matrix with: list(list(tuple(incidence, probability))). 233 | """ 234 | if 'NaN' in predicted_row: # don't wanna return NaN 235 | return 236 | else: 237 | vector = [(1 if label_index[i] in gold_standard else 0, 238 | float(predicted_row[i + 1])) for i in 239 | range(len(label_index))] 240 | return vector 241 | 242 | def get_scores(self, label_index, predicted_weights, true_labels): 243 | """Grab results. 244 | 245 | Given the labelled and original file, retrieve for each 246 | vector: the correct label, ranks and probabilities. Get 247 | tuple vector, unzip it and add the incidence part to 248 | y_true and the probability part to y_score (these are 249 | sklearn arrays for evluation). 250 | 251 | Parameters 252 | ---------- 253 | label_index : list of tuples 254 | Enumerated list with topic indexes. 255 | 256 | predicted_weights : string 257 | Csv file directory containing label confidences. 258 | 259 | true_labels : string 260 | Csv file directory containing original material. 261 | 262 | Return 263 | ------ 264 | y_true : list of integers 265 | Binary list (incidence matrix). 266 | 267 | y_score : list of floats 268 | Probabilities per topic. 269 | """ 270 | y_true = [] 271 | y_score = [] 272 | for predicted_row, true_row in zip(predicted_weights, true_labels): 273 | gold_standard = true_row.lower().split() 274 | rank, prob = zip(*self.m_incidence(predicted_row, label_index, 275 | gold_standard)) 276 | if 1 in rank: 277 | y_true.append(rank) 278 | y_score.append(prob) 279 | 280 | return y_true, y_score 281 | 282 | def to_array(self, y_true, y_score): 283 | """To sklean-ready array. 284 | 285 | Converts the incidence matrix and its probabilites to a numpy format. 286 | Also cleans out columns that produce a sum of zeroes; this results in 287 | a division by zero error when determining recall. Dependencies are 288 | both numpu and scipy. 289 | 290 | Parameters 291 | ---------- 292 | y_true : list of integers 293 | Binary list (incidence matrix). 294 | 295 | y_score : list of floats 296 | Probabilities per topic. 297 | 298 | Return 299 | ------ 300 | (y_true, y_score): numpy arrays 301 | Filtered and converted version of y_true and y_score input. 302 | """ 303 | from collections import Counter 304 | import scipy 305 | import numpy as np 306 | 307 | def scan_empty(y_true): 308 | c = Counter() 309 | for x in y_true: 310 | for i, y in enumerate(x): 311 | c[i] += y 312 | return [key for key, value in c.items() if value == 0] 313 | 314 | def lab_reduce(y_true, y_score): 315 | empty_indices = scan_empty(y_true) 316 | i = 0 317 | for k in empty_indices: 318 | y_true = scipy.delete(y_true, k-i, 1) 319 | y_score = scipy.delete(y_score, k-i, 1) 320 | i += 1 321 | return y_true, y_score 322 | 323 | return lab_reduce(np.asarray(y_true), np.asarray(y_score)) 324 | 325 | def results(self, true_labels, array=False): 326 | """Results grabber. 327 | 328 | Finds the predicted document topic distribution and label index for the 329 | model, then retrieves the actual labels from the original file and 330 | serves these to self.get_scores. 331 | 332 | labels : list 333 | The original set of labels per document 334 | 335 | array : boolean, optional, default False 336 | Returns a cleaned numpy array where a column cannot be all zeroes. 337 | Has numpy and scipy as dependencies; better handle this outside of 338 | the class if you do not want to work with those. 339 | 340 | Return 341 | ------ 342 | y_true, y_score : list, list 343 | List of lists incidence matrix (binary) and list of lists document 344 | topic probabilities. 345 | """ 346 | DTDA = 'document-topic-distributions-res' # doctop file 347 | LIDX = '00000{0}label-index'.format(sep) # label index 348 | 349 | orf = open("{0}{1}_{2}{3}{4}.txt".format( 350 | self.dir, self.name, 'train', sep, LIDX), 'r') 351 | label_index = orf.read().lower().split('\n')[:-1] 352 | 353 | lbf = \ 354 | open("{0}{1}_{2}{3}{4}_{5}-{6}.csv".format( 355 | self.dir, self.name, 'train', sep, self.name, 'test', DTDA), 356 | 'r') 357 | predicted_weights = reader(lbf) 358 | 359 | y_true, y_score = self.get_scores(label_index, predicted_weights, 360 | true_labels) 361 | 362 | lbf.close() 363 | orf.close() 364 | 365 | if array: 366 | y_true, y_score = self.to_array(y_true, y_score) 367 | 368 | self.cleanup(step='results') 369 | return y_true, y_score 370 | 371 | def cleanup(self, rmall=False, step=False): 372 | """Cleanup module. 373 | 374 | If the user wants the trained model to be kept, it will only remove the 375 | .csvs and wordcounts. Otherwise, it also dumps the fully trained model 376 | in self.train. 377 | 378 | Parameters 379 | ---------- 380 | rmall : bool, optional, default False 381 | Can be used to remove ALL files from box. 382 | 383 | step : bool, optional, default False 384 | Indicates the step so that it will keep the compressed and model 385 | files. 386 | """ 387 | pattern = self.name + '_*' if not rmall else '*_*' 388 | files = glob(self.dir + pattern) 389 | for f in files: 390 | if not self.keep and step != 'results': 391 | rmtree(f) if '.' not in f else remove(f) 392 | else: 393 | remove(f) if '.' in f and '.gz' not in f else None 394 | 395 | def run(self, space, labels, step): 396 | """Main runner. 397 | 398 | Checks if the given space is given in a generator for batching, writes 399 | it out to a csv with self.store, then self.boot-s the model in either 400 | train or test mode. If it's in test, it will return the results so that 401 | self,results does not have to be used. 402 | 403 | Parameters 404 | ---------- 405 | space : list 406 | The vector space; a list with text. 407 | 408 | labels : list 409 | List with labels where each index corresponds to the text in space. 410 | 411 | step : str 412 | Either test or train. 413 | """ 414 | if not isgenerator(space): 415 | space = [space] 416 | labels = [labels] 417 | for batch_space, batch_labels in zip(space, labels): 418 | self.store(batch_space, batch_labels, step) 419 | space, labels = None, None 420 | self.boot(step) 421 | self.cleanup() 422 | 423 | def train(self, space, labels): 424 | """Sugar train. 425 | 426 | Will train a previously untrained STMT instance on the given 427 | vectorspace. Please check the store function for space requirements. 428 | Can accept a generator for both space and labels. 429 | 430 | Parameters 431 | ---------- 432 | space : list 433 | The vector space; a list with text. 434 | 435 | labels : list 436 | List with labels where each index corresponds to the text in space. 437 | """ 438 | self.run(space, labels, 'train') 439 | 440 | def test(self, space, labels): 441 | """Sugar test. 442 | 443 | Will test a previously trained STMT instance on the given vectorspace. 444 | Please check the store function for space requirements. 445 | Can accept a generator for both space and labels. 446 | 447 | Parameters 448 | ---------- 449 | space : list 450 | The vector space; a list with text. 451 | 452 | labels : list 453 | List with labels where each index corresponds to the text in space. 454 | """ 455 | self.run(space, labels, 'test') 456 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | GNU GENERAL PUBLIC LICENSE 2 | Version 2, June 1991 3 | 4 | Copyright (C) 1989, 1991 Free Software Foundation, Inc., 5 | 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 6 | Everyone is permitted to copy and distribute verbatim copies 7 | of this license document, but changing it is not allowed. 8 | 9 | Preamble 10 | 11 | The licenses for most software are designed to take away your 12 | freedom to share and change it. By contrast, the GNU General Public 13 | License is intended to guarantee your freedom to share and change free 14 | software--to make sure the software is free for all its users. This 15 | General Public License applies to most of the Free Software 16 | Foundation's software and to any other program whose authors commit to 17 | using it. (Some other Free Software Foundation software is covered by 18 | the GNU Lesser General Public License instead.) You can apply it to 19 | your programs, too. 20 | 21 | When we speak of free software, we are referring to freedom, not 22 | price. Our General Public Licenses are designed to make sure that you 23 | have the freedom to distribute copies of free software (and charge for 24 | this service if you wish), that you receive source code or can get it 25 | if you want it, that you can change the software or use pieces of it 26 | in new free programs; and that you know you can do these things. 27 | 28 | To protect your rights, we need to make restrictions that forbid 29 | anyone to deny you these rights or to ask you to surrender the rights. 30 | These restrictions translate to certain responsibilities for you if you 31 | distribute copies of the software, or if you modify it. 32 | 33 | For example, if you distribute copies of such a program, whether 34 | gratis or for a fee, you must give the recipients all the rights that 35 | you have. You must make sure that they, too, receive or can get the 36 | source code. And you must show them these terms so they know their 37 | rights. 38 | 39 | We protect your rights with two steps: (1) copyright the software, and 40 | (2) offer you this license which gives you legal permission to copy, 41 | distribute and/or modify the software. 42 | 43 | Also, for each author's protection and ours, we want to make certain 44 | that everyone understands that there is no warranty for this free 45 | software. If the software is modified by someone else and passed on, we 46 | want its recipients to know that what they have is not the original, so 47 | that any problems introduced by others will not reflect on the original 48 | authors' reputations. 49 | 50 | Finally, any free program is threatened constantly by software 51 | patents. We wish to avoid the danger that redistributors of a free 52 | program will individually obtain patent licenses, in effect making the 53 | program proprietary. To prevent this, we have made it clear that any 54 | patent must be licensed for everyone's free use or not licensed at all. 55 | 56 | The precise terms and conditions for copying, distribution and 57 | modification follow. 58 | 59 | GNU GENERAL PUBLIC LICENSE 60 | TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION 61 | 62 | 0. This License applies to any program or other work which contains 63 | a notice placed by the copyright holder saying it may be distributed 64 | under the terms of this General Public License. The "Program", below, 65 | refers to any such program or work, and a "work based on the Program" 66 | means either the Program or any derivative work under copyright law: 67 | that is to say, a work containing the Program or a portion of it, 68 | either verbatim or with modifications and/or translated into another 69 | language. (Hereinafter, translation is included without limitation in 70 | the term "modification".) Each licensee is addressed as "you". 71 | 72 | Activities other than copying, distribution and modification are not 73 | covered by this License; they are outside its scope. The act of 74 | running the Program is not restricted, and the output from the Program 75 | is covered only if its contents constitute a work based on the 76 | Program (independent of having been made by running the Program). 77 | Whether that is true depends on what the Program does. 78 | 79 | 1. You may copy and distribute verbatim copies of the Program's 80 | source code as you receive it, in any medium, provided that you 81 | conspicuously and appropriately publish on each copy an appropriate 82 | copyright notice and disclaimer of warranty; keep intact all the 83 | notices that refer to this License and to the absence of any warranty; 84 | and give any other recipients of the Program a copy of this License 85 | along with the Program. 86 | 87 | You may charge a fee for the physical act of transferring a copy, and 88 | you may at your option offer warranty protection in exchange for a fee. 89 | 90 | 2. You may modify your copy or copies of the Program or any portion 91 | of it, thus forming a work based on the Program, and copy and 92 | distribute such modifications or work under the terms of Section 1 93 | above, provided that you also meet all of these conditions: 94 | 95 | a) You must cause the modified files to carry prominent notices 96 | stating that you changed the files and the date of any change. 97 | 98 | b) You must cause any work that you distribute or publish, that in 99 | whole or in part contains or is derived from the Program or any 100 | part thereof, to be licensed as a whole at no charge to all third 101 | parties under the terms of this License. 102 | 103 | c) If the modified program normally reads commands interactively 104 | when run, you must cause it, when started running for such 105 | interactive use in the most ordinary way, to print or display an 106 | announcement including an appropriate copyright notice and a 107 | notice that there is no warranty (or else, saying that you provide 108 | a warranty) and that users may redistribute the program under 109 | these conditions, and telling the user how to view a copy of this 110 | License. (Exception: if the Program itself is interactive but 111 | does not normally print such an announcement, your work based on 112 | the Program is not required to print an announcement.) 113 | 114 | These requirements apply to the modified work as a whole. If 115 | identifiable sections of that work are not derived from the Program, 116 | and can be reasonably considered independent and separate works in 117 | themselves, then this License, and its terms, do not apply to those 118 | sections when you distribute them as separate works. But when you 119 | distribute the same sections as part of a whole which is a work based 120 | on the Program, the distribution of the whole must be on the terms of 121 | this License, whose permissions for other licensees extend to the 122 | entire whole, and thus to each and every part regardless of who wrote it. 123 | 124 | Thus, it is not the intent of this section to claim rights or contest 125 | your rights to work written entirely by you; rather, the intent is to 126 | exercise the right to control the distribution of derivative or 127 | collective works based on the Program. 128 | 129 | In addition, mere aggregation of another work not based on the Program 130 | with the Program (or with a work based on the Program) on a volume of 131 | a storage or distribution medium does not bring the other work under 132 | the scope of this License. 133 | 134 | 3. You may copy and distribute the Program (or a work based on it, 135 | under Section 2) in object code or executable form under the terms of 136 | Sections 1 and 2 above provided that you also do one of the following: 137 | 138 | a) Accompany it with the complete corresponding machine-readable 139 | source code, which must be distributed under the terms of Sections 140 | 1 and 2 above on a medium customarily used for software interchange; or, 141 | 142 | b) Accompany it with a written offer, valid for at least three 143 | years, to give any third party, for a charge no more than your 144 | cost of physically performing source distribution, a complete 145 | machine-readable copy of the corresponding source code, to be 146 | distributed under the terms of Sections 1 and 2 above on a medium 147 | customarily used for software interchange; or, 148 | 149 | c) Accompany it with the information you received as to the offer 150 | to distribute corresponding source code. (This alternative is 151 | allowed only for noncommercial distribution and only if you 152 | received the program in object code or executable form with such 153 | an offer, in accord with Subsection b above.) 154 | 155 | The source code for a work means the preferred form of the work for 156 | making modifications to it. For an executable work, complete source 157 | code means all the source code for all modules it contains, plus any 158 | associated interface definition files, plus the scripts used to 159 | control compilation and installation of the executable. However, as a 160 | special exception, the source code distributed need not include 161 | anything that is normally distributed (in either source or binary 162 | form) with the major components (compiler, kernel, and so on) of the 163 | operating system on which the executable runs, unless that component 164 | itself accompanies the executable. 165 | 166 | If distribution of executable or object code is made by offering 167 | access to copy from a designated place, then offering equivalent 168 | access to copy the source code from the same place counts as 169 | distribution of the source code, even though third parties are not 170 | compelled to copy the source along with the object code. 171 | 172 | 4. You may not copy, modify, sublicense, or distribute the Program 173 | except as expressly provided under this License. Any attempt 174 | otherwise to copy, modify, sublicense or distribute the Program is 175 | void, and will automatically terminate your rights under this License. 176 | However, parties who have received copies, or rights, from you under 177 | this License will not have their licenses terminated so long as such 178 | parties remain in full compliance. 179 | 180 | 5. You are not required to accept this License, since you have not 181 | signed it. However, nothing else grants you permission to modify or 182 | distribute the Program or its derivative works. These actions are 183 | prohibited by law if you do not accept this License. Therefore, by 184 | modifying or distributing the Program (or any work based on the 185 | Program), you indicate your acceptance of this License to do so, and 186 | all its terms and conditions for copying, distributing or modifying 187 | the Program or works based on it. 188 | 189 | 6. Each time you redistribute the Program (or any work based on the 190 | Program), the recipient automatically receives a license from the 191 | original licensor to copy, distribute or modify the Program subject to 192 | these terms and conditions. You may not impose any further 193 | restrictions on the recipients' exercise of the rights granted herein. 194 | You are not responsible for enforcing compliance by third parties to 195 | this License. 196 | 197 | 7. If, as a consequence of a court judgment or allegation of patent 198 | infringement or for any other reason (not limited to patent issues), 199 | conditions are imposed on you (whether by court order, agreement or 200 | otherwise) that contradict the conditions of this License, they do not 201 | excuse you from the conditions of this License. If you cannot 202 | distribute so as to satisfy simultaneously your obligations under this 203 | License and any other pertinent obligations, then as a consequence you 204 | may not distribute the Program at all. For example, if a patent 205 | license would not permit royalty-free redistribution of the Program by 206 | all those who receive copies directly or indirectly through you, then 207 | the only way you could satisfy both it and this License would be to 208 | refrain entirely from distribution of the Program. 209 | 210 | If any portion of this section is held invalid or unenforceable under 211 | any particular circumstance, the balance of the section is intended to 212 | apply and the section as a whole is intended to apply in other 213 | circumstances. 214 | 215 | It is not the purpose of this section to induce you to infringe any 216 | patents or other property right claims or to contest validity of any 217 | such claims; this section has the sole purpose of protecting the 218 | integrity of the free software distribution system, which is 219 | implemented by public license practices. Many people have made 220 | generous contributions to the wide range of software distributed 221 | through that system in reliance on consistent application of that 222 | system; it is up to the author/donor to decide if he or she is willing 223 | to distribute software through any other system and a licensee cannot 224 | impose that choice. 225 | 226 | This section is intended to make thoroughly clear what is believed to 227 | be a consequence of the rest of this License. 228 | 229 | 8. If the distribution and/or use of the Program is restricted in 230 | certain countries either by patents or by copyrighted interfaces, the 231 | original copyright holder who places the Program under this License 232 | may add an explicit geographical distribution limitation excluding 233 | those countries, so that distribution is permitted only in or among 234 | countries not thus excluded. In such case, this License incorporates 235 | the limitation as if written in the body of this License. 236 | 237 | 9. The Free Software Foundation may publish revised and/or new versions 238 | of the General Public License from time to time. Such new versions will 239 | be similar in spirit to the present version, but may differ in detail to 240 | address new problems or concerns. 241 | 242 | Each version is given a distinguishing version number. If the Program 243 | specifies a version number of this License which applies to it and "any 244 | later version", you have the option of following the terms and conditions 245 | either of that version or of any later version published by the Free 246 | Software Foundation. If the Program does not specify a version number of 247 | this License, you may choose any version ever published by the Free Software 248 | Foundation. 249 | 250 | 10. If you wish to incorporate parts of the Program into other free 251 | programs whose distribution conditions are different, write to the author 252 | to ask for permission. For software which is copyrighted by the Free 253 | Software Foundation, write to the Free Software Foundation; we sometimes 254 | make exceptions for this. Our decision will be guided by the two goals 255 | of preserving the free status of all derivatives of our free software and 256 | of promoting the sharing and reuse of software generally. 257 | 258 | NO WARRANTY 259 | 260 | 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY 261 | FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN 262 | OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES 263 | PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED 264 | OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF 265 | MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS 266 | TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE 267 | PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, 268 | REPAIR OR CORRECTION. 269 | 270 | 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING 271 | WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR 272 | REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, 273 | INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING 274 | OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED 275 | TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY 276 | YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER 277 | PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE 278 | POSSIBILITY OF SUCH DAMAGES. 279 | 280 | END OF TERMS AND CONDITIONS 281 | 282 | How to Apply These Terms to Your New Programs 283 | 284 | If you develop a new program, and you want it to be of the greatest 285 | possible use to the public, the best way to achieve this is to make it 286 | free software which everyone can redistribute and change under these terms. 287 | 288 | To do so, attach the following notices to the program. It is safest 289 | to attach them to the start of each source file to most effectively 290 | convey the exclusion of warranty; and each file should have at least 291 | the "copyright" line and a pointer to where the full notice is found. 292 | 293 | {description} 294 | Copyright (C) {year} {fullname} 295 | 296 | This program is free software; you can redistribute it and/or modify 297 | it under the terms of the GNU General Public License as published by 298 | the Free Software Foundation; either version 2 of the License, or 299 | (at your option) any later version. 300 | 301 | This program is distributed in the hope that it will be useful, 302 | but WITHOUT ANY WARRANTY; without even the implied warranty of 303 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 304 | GNU General Public License for more details. 305 | 306 | You should have received a copy of the GNU General Public License along 307 | with this program; if not, write to the Free Software Foundation, Inc., 308 | 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. 309 | 310 | Also add information on how to contact you by electronic and paper mail. 311 | 312 | If the program is interactive, make it output a short notice like this 313 | when it starts in an interactive mode: 314 | 315 | Gnomovision version 69, Copyright (C) year name of author 316 | Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. 317 | This is free software, and you are welcome to redistribute it 318 | under certain conditions; type `show c' for details. 319 | 320 | The hypothetical commands `show w' and `show c' should show the appropriate 321 | parts of the General Public License. Of course, the commands you use may 322 | be called something other than `show w' and `show c'; they could even be 323 | mouse-clicks or menu items--whatever suits your program. 324 | 325 | You should also get your employer (if you work as a programmer) or your 326 | school, if any, to sign a "copyright disclaimer" for the program, if 327 | necessary. Here is a sample; alter the names: 328 | 329 | Yoyodyne, Inc., hereby disclaims all copyright interest in the program 330 | `Gnomovision' (which makes passes at compilers) written by James Hacker. 331 | 332 | {signature of Ty Coon}, 1 April 1989 333 | Ty Coon, President of Vice 334 | 335 | This General Public License does not permit incorporating your program into 336 | proprietary programs. If your program is a subroutine library, you may 337 | consider it more useful to permit linking proprietary applications with the 338 | library. If this is what you want to do, use the GNU Lesser General 339 | Public License instead of this License. 340 | 341 | --------------------------------------------------------------------------------