├── .gitignore ├── assets ├── overview.png ├── dev_set_results.png └── test_set_results.png ├── JUnit ├── jsoup-1.17.2.jar ├── lombok-1.18.24.jar ├── fastjson2-2.0.41.jar ├── junit-jupiter-api-5.8.1.jar ├── junit-jupiter-engine-5.8.1.jar ├── junit-platform-engine-1.8.1.jar ├── junit-platform-commons-1.8.1.jar ├── junit-platform-launcher-1.7.2.jar └── junit-platform-console-standalone-1.7.0.jar ├── ncb ├── utils.py ├── evaluate.py ├── exec_java.py ├── exec_python.py ├── execution.py └── test_setup.py ├── README.md └── results └── reference ├── reference_ncb_python_en.jsonl └── reference_ncb_python_zh.jsonl /.gitignore: -------------------------------------------------------------------------------- 1 | .idea 2 | __pycache__ 3 | .DS_Store 4 | -------------------------------------------------------------------------------- /assets/overview.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/assets/overview.png -------------------------------------------------------------------------------- /JUnit/jsoup-1.17.2.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/jsoup-1.17.2.jar -------------------------------------------------------------------------------- /JUnit/lombok-1.18.24.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/lombok-1.18.24.jar -------------------------------------------------------------------------------- /JUnit/fastjson2-2.0.41.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/fastjson2-2.0.41.jar -------------------------------------------------------------------------------- /assets/dev_set_results.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/assets/dev_set_results.png -------------------------------------------------------------------------------- /assets/test_set_results.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/assets/test_set_results.png -------------------------------------------------------------------------------- /JUnit/junit-jupiter-api-5.8.1.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/junit-jupiter-api-5.8.1.jar -------------------------------------------------------------------------------- /JUnit/junit-jupiter-engine-5.8.1.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/junit-jupiter-engine-5.8.1.jar -------------------------------------------------------------------------------- /JUnit/junit-platform-engine-1.8.1.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/junit-platform-engine-1.8.1.jar -------------------------------------------------------------------------------- /JUnit/junit-platform-commons-1.8.1.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/junit-platform-commons-1.8.1.jar -------------------------------------------------------------------------------- /JUnit/junit-platform-launcher-1.7.2.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/junit-platform-launcher-1.7.2.jar -------------------------------------------------------------------------------- /JUnit/junit-platform-console-standalone-1.7.0.jar: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/THUDM/NaturalCodeBench/HEAD/JUnit/junit-platform-console-standalone-1.7.0.jar -------------------------------------------------------------------------------- /ncb/utils.py: -------------------------------------------------------------------------------- 1 | import itertools 2 | import json 3 | import os 4 | import shutil 5 | from pathlib import Path 6 | from typing import List, Union 7 | import numpy as np 8 | 9 | 10 | def load_json(path): 11 | return json.load(open(path, encoding='utf-8')) 12 | 13 | 14 | def save_json(obj, path): 15 | with open(path, 'w', encoding='utf-8') as f: 16 | json.dump(obj, f, ensure_ascii=False, indent=2) 17 | 18 | 19 | def load_jsonl(path): 20 | res = [] 21 | with open(path, encoding='utf-8') as f: 22 | for line in f: 23 | res.append(json.loads(line)) 24 | return res 25 | 26 | 27 | def save_jsonl(obj, path): 28 | with open(path, 'w', encoding='utf-8') as f: 29 | for item in obj: 30 | f.write(json.dumps(item, ensure_ascii=False)+'\n') 31 | 32 | 33 | def del_file(path): 34 | for elm in Path(path).glob('*'): 35 | elm.unlink() if elm.is_file() else shutil.rmtree(elm) 36 | if os.path.exists(path): 37 | os.rmdir(path) 38 | 39 | 40 | def estimate_pass_at_k( 41 | num_samples: Union[int, List[int], np.ndarray], 42 | num_correct: Union[List[int], np.ndarray], 43 | k: int 44 | ) -> np.ndarray: 45 | """ 46 | Estimates pass@k of each problem and returns them in an array. 47 | """ 48 | 49 | def estimator(n: int, c: int, k: int) -> float: 50 | """ 51 | Calculates 1 - comb(n - c, k) / comb(n, k). 52 | """ 53 | if n - c < k: 54 | return 1.0 55 | return 1.0 - np.prod(1.0 - k / np.arange(n - c + 1, n + 1)) 56 | 57 | if isinstance(num_samples, int): 58 | num_samples_it = itertools.repeat(num_samples, len(num_correct)) 59 | else: 60 | assert len(num_samples) == len(num_correct) 61 | num_samples_it = iter(num_samples) 62 | 63 | return np.array([estimator(int(n), int(c), k) for n, c in zip(num_samples_it, num_correct)]) 64 | -------------------------------------------------------------------------------- /ncb/evaluate.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | import os 3 | from pathlib import Path 4 | from test_setup import write_test_files, extract_codes 5 | from execution import execution 6 | from utils import load_json, load_jsonl, save_json, save_jsonl, del_file 7 | 8 | 9 | def evaluate_code(data_dir, language, natural_lang, ckpt_name, ks, num_workers=64): 10 | file = f'results/{ckpt_name}/{ckpt_name}_ncb_{language}_{natural_lang}.jsonl' 11 | print(file) 12 | os.makedirs(data_dir, exist_ok=True) 13 | del_file(data_dir / f'{ckpt_name}') 14 | 15 | data = load_jsonl(file) 16 | dataset_size = len(data) 17 | testcases = load_jsonl(f'data/{language}_{natural_lang}/ncb_{language}_{natural_lang}.jsonl') 18 | 19 | print("Extracting code from response") 20 | data = extract_codes(data, testcases, language) 21 | 22 | print("Writing test files") 23 | input_files_path = Path(f'data/{language}_{natural_lang}/input_files') 24 | test_dir = write_test_files(data_dir, input_files_path, data, language, ckpt_name) 25 | 26 | print("Start evaluation") 27 | result = execution(data_dir / test_dir, ckpt_name, language, natural_lang, dataset_size, ks, num_workers) 28 | print(result, '\n') 29 | save_json(result, data_dir / test_dir / 'result.json') 30 | 31 | 32 | if __name__ == '__main__': 33 | parser = argparse.ArgumentParser() 34 | parser.add_argument('--languages', type=str, nargs='+', default=['python', 'java'], help='programming language') 35 | parser.add_argument('--natural_langs', type=str, nargs='+', default=['zh', 'en'], help='natural language') 36 | parser.add_argument('--ckpt_name', type=str, required=True, default='reference', help='the name of ckpt that you want to evaluate') 37 | parser.add_argument('--num_workers', type=int, default=64, help='number of workers') 38 | parser.add_argument('--ks', type=int, nargs='+', default=[1], help='k of the pass@k') 39 | args = parser.parse_args() 40 | 41 | results = [] 42 | for nat_lang in args.natural_langs: 43 | for lang in args.languages: 44 | data_dir = Path(f'data/temp/{lang}_{nat_lang}_test') 45 | evaluate_code(data_dir, lang, nat_lang, args.ckpt_name, args.ks, args.num_workers) 46 | results.append(load_json(data_dir / f'{args.ckpt_name}/result.json')) 47 | save_jsonl(results, f'results/{args.ckpt_name}/results.jsonl') 48 | for res in results: 49 | print(res) 50 | -------------------------------------------------------------------------------- /ncb/exec_java.py: -------------------------------------------------------------------------------- 1 | import subprocess 2 | import copy 3 | import os 4 | from pathlib import Path 5 | from utils import save_json 6 | 7 | 8 | def excute(tmp_dir: str, JUnit_path: str, timeout: int): 9 | if not tmp_dir: 10 | raise ValueError(f"pls enter the path for executing java") 11 | JUnit = f':{JUnit_path}/*' 12 | tr_file = "TestRunner.java" 13 | tr_name = "TestRunner" 14 | compilation_result = subprocess.run(["javac", "-cp",JUnit,tr_file], cwd=tmp_dir, timeout=timeout,capture_output=True) 15 | compile_returncode = compilation_result.returncode 16 | if compile_returncode == 0: 17 | try: 18 | exec_result = subprocess.run(["java", "-cp",JUnit,tr_name], cwd=tmp_dir, timeout=timeout,capture_output=True) 19 | if exec_result.returncode == 0: 20 | res = exec_result.stdout.decode() 21 | elif exec_result.returncode == 1: 22 | res = f"failed: execute error:\n{exec_result.stderr.decode()}" 23 | except subprocess.TimeoutExpired as e: 24 | res = "time out" 25 | except BaseException as e: 26 | res = f"failed: {e}" 27 | else: 28 | compile_error = compilation_result.stderr.decode('utf-8') 29 | res = f'failed: compilation error:\n{compile_error}' 30 | return res 31 | 32 | 33 | def eval_test(class_name, base_dir, JUnit_path, timeout: int): 34 | item = {} 35 | _id = class_name[class_name.find('_')+1:class_name.rfind('_')] 36 | base_dir = copy.deepcopy(base_dir) 37 | tmp_dir = Path(base_dir, class_name) 38 | compile_error = "" 39 | exe_error = "" 40 | _id = _id = str(class_name)[str(class_name).find('_')+1:str(class_name).rfind('_')] 41 | if not os.listdir(tmp_dir): 42 | item['_id'] = _id 43 | item['class_name'] = class_name 44 | item["compile_error"] = compile_error 45 | item["exe_error"] = exe_error 46 | item["test_result"] = 'nocode' 47 | save_json(item, Path(base_dir, tmp_dir, 'eval_report.json')) 48 | return item 49 | res = excute(tmp_dir, JUnit_path, timeout=timeout) 50 | if 'failed: compilation error' in res: 51 | compile_error = res 52 | 53 | if 'failed: execute error:' in res: 54 | exe_error = res 55 | 56 | item['_id'] = _id 57 | item['class_name'] = class_name 58 | item["compile_error"] = compile_error 59 | item["exe_error"] = exe_error 60 | item["test_result"] = res 61 | 62 | save_json(item, Path(base_dir, tmp_dir, 'eval_report.json')) 63 | return item 64 | -------------------------------------------------------------------------------- /ncb/exec_python.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | import os 3 | import sys 4 | sys.path.append(os.getcwd()) 5 | from pathlib import Path 6 | import pytest 7 | from pytest_jsonreport.plugin import JSONReport 8 | from utils import save_jsonl, save_json 9 | 10 | 11 | def level(rate): 12 | if rate == 0: 13 | return '0' 14 | elif 0 < rate < 30: 15 | return '0_30' 16 | elif 30 <= rate < 60: 17 | return '30_60' 18 | elif 60 <= rate < 100: 19 | return '60_100' 20 | elif rate == 100: 21 | return '100' 22 | 23 | 24 | def run_python_test(base_dir, test_dir): 25 | _id = test_dir[test_dir.rfind('_')+1:] 26 | os.chdir(Path(base_dir, test_dir)) 27 | 28 | overall_result = {'total': 0, 'passed': 0, 'failed': 0} 29 | pass_statistics = {'0': 0, '0_30': 0, '30_60': 0, '60_100': 0, '100': 0} 30 | error = {'SyntaxError': 0, 'RuntimeError': 0} 31 | error_report = [] 32 | error_cases = 0 33 | 34 | plugin = JSONReport() 35 | pytest.main(['-s', '--cov', '--cov-report', 'json', '--disable-warnings'], plugins=[plugin]) 36 | 37 | if len(plugin.report.get('tests')) == 0: 38 | longrepr = plugin.report['collectors'][1]['longrepr'] 39 | error_report.append({"_id": _id, 'nodeid': plugin.report['collectors'][1]['nodeid'], 'longrepr': longrepr}) 40 | if "SyntaxError" in longrepr: 41 | error['SyntaxError'] += 1 42 | else: 43 | error['RuntimeError'] += 1 44 | save_json(error, Path(base_dir, test_dir, 'error.json')) 45 | save_jsonl(error_report, Path(base_dir, test_dir, 'error_report.jsonl')) 46 | save_json({'overall_result': overall_result, 'pass_cases': pass_statistics}, 47 | Path(base_dir, test_dir, 'raw_result.json')) 48 | return 49 | try: 50 | error_report.append({"_id": _id, "testcases_traceback": []}) 51 | for case in plugin.report.get('tests'): 52 | print(case) 53 | if 'traceback' not in case['call']: 54 | continue 55 | if 'AssertionError' not in case['call']['traceback'][0]['message']: 56 | error_cases += 1 57 | error_report[0]['testcases_traceback'].append({'nodeid': case['nodeid'], 58 | 'traceback': case['call']['longrepr']}) 59 | except Exception as e: 60 | pass 61 | error['RuntimeError'] += 1 if error_cases != 0 else 0 62 | 63 | summary = plugin.report.get("summary") 64 | passed = summary.get("passed", 0) 65 | total = summary.get("total", 0) 66 | overall_result['passed'] += passed 67 | overall_result['total'] += total 68 | if total != 0: 69 | pass_rate = (passed / total) * 100 70 | pass_statistics[level(pass_rate)] += 1 71 | 72 | if len(error_report[0]['testcases_traceback']) == 0: 73 | error_report = [] 74 | save_jsonl(error_report, Path(base_dir, test_dir, 'error_report.jsonl')) 75 | save_json(error, Path(base_dir, test_dir, 'error.json')) 76 | save_json({'overall_result': overall_result, 'pass_cases': pass_statistics}, 77 | Path(base_dir, test_dir, 'raw_result.json')) 78 | 79 | 80 | if __name__ == '__main__': 81 | parser = argparse.ArgumentParser() 82 | parser.add_argument('--base_dir', type=str, required=True) 83 | parser.add_argument('--test_dir', type=str, required=True) 84 | args = parser.parse_args() 85 | 86 | run_python_test(args.base_dir, args.test_dir) 87 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # NaturalCodeBench (ACL 2024 Findings) 2 | 3 | This repository contains information, data and code of NaturalCodeBench: A Challenging Application-Driven Dataset for Code Synthesis Evaluation. 4 | 5 | ## 📌Introduction 6 | 7 | We propose NaturalCodeBench (NCB), a comprehensive code benchmark designed to mirror the complexity and variety of scenarios in real coding tasks. NCB comprises 402 high-quality problems in Python and Java, meticulously selected from an online coding service, covering 6 different domains. 8 | 9 | ![overview](assets/overview.png) 10 | 11 | The overall framework of NaturalCodeBench is shown in the above image, including the data collection pipeline and the semi-automated pipeline. 12 | 13 | For a full description of NaturalCodeBench, please refer to the paper: https://arxiv.org/abs/2405.04520 14 | 15 | ## Dataset Summary 16 | 17 | To construct a challenging application-driven dataset for code synthesis evaluation, the seed problems of NCB are cleaned from the queries in coding online services, spanning across 6 domains: Artificial Intelligence, Data Science, Algorithm and Data Structure, Front-End, Software Engineering, and System Administration. 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 | 49 | 50 | 51 | 52 | 53 | 54 | 55 | 56 | 57 | 58 | 59 | 60 | 61 | 62 | 63 | 64 | 65 |
Domains#Problems
DevTestTotal
Software Engineering4488132
Data Science3268100
Algorithm and Data Structure227395
System Administration221733
Artificial Intelligence151328
Front-End11314
66 | 67 | 68 | NaturalCodeBench contains 402 high-quality problems in total. We release the development set of NCB, which contains 140 problems (70 in Python and 70 in Java) for research purpose. The data is placed in ... 69 | 70 | The data format is as follows. 71 | 72 | - ```_id```(integer): A unique identifier for each question. 73 | - ```prompt```(string): The prompt involving problem description and instruction. 74 | - ```problem```(string): The problem description 75 | - ```testcases```(string): The code of testcases 76 | - ```setup_code```(string): The code for test setup 77 | - ```reference_solution```(string): A reference answer to solve the problem. 78 | - ```classification```(string): The domain of the problem 79 | 80 | ## Installation 81 | 82 | We provide a docker to setup the environment. Firstly pull the image. 83 | 84 | ``` 85 | docker pull codegeex/codegeex:0.1.23 86 | ``` 87 | 88 | Then start Docker and mount the code directory. 89 | 90 | ```bash 91 | docker run --rm -it --shm-size 32g -v /path/to/NaturalCodeBench:/ncb codegeex/codegeex:0.1.23 /bin/bash 92 | ``` 93 | 94 | After starting the Docker shell, transfer data into the repository. 95 | 96 | ``` 97 | cd /ncb 98 | cp -r /NaturalCodeBench/data . 99 | ``` 100 | 101 | ## Usage 102 | 103 | Generate samples and save them in the following JSON Lines (jsonl) format, where each sample is formatted into a single line like so: 104 | 105 | ``` 106 | {"_id": "NaturalCodeBench Problem ID", "response": "The response of model without prompt"} 107 | ``` 108 | 109 | Place your JSONL files into the `results` directory according to the following directory structure. 110 | 111 | ``` 112 | results/ 113 | └── {model_name}/ 114 | ├── {model_name}_ncb_java_en.jsonl 115 | ├── {model_name}_ncb_java_zh.jsonl 116 | ├── {model_name}_ncb_python_en.jsonl 117 | └── {model_name}_ncb_python_zh.jsonl 118 | ``` 119 | 120 | We provide `reference` under `results` to illustrate the format and help with debugging. 121 | 122 | To evaluate the samples, run 123 | 124 | ``` 125 | python ncb/evaluate.py --languages python java --natural_lang zh en --ckpt_name {your_model_name} --num_workers 64 --ks 1 10 100 126 | ``` 127 | 128 | 129 | 130 | ## Leaderboard 131 | 132 | We report our evaluation results of 39 LLMs on the test and dev datasets. 133 | 134 | Test set results: 135 | 136 | ![overview](assets/test_set_results.png) 137 | 138 | 139 | 140 | Dev set result: 141 | 142 | ![overview](assets/dev_set_results.png) 143 | 144 | 145 | 146 | ## Citation 147 | 148 | ``` 149 | @inproceedings{zhang-etal-2024-naturalcodebench, 150 | title = "{N}atural{C}ode{B}ench: Examining Coding Performance Mismatch on {H}uman{E}val and Natural User Queries", 151 | author = "Zhang, Shudan and 152 | Zhao, Hanlin and 153 | Liu, Xiao and 154 | Zheng, Qinkai and 155 | Qi, Zehan and 156 | Gu, Xiaotao and 157 | Dong, Yuxiao and 158 | Tang, Jie", 159 | editor = "Ku, Lun-Wei and 160 | Martins, Andre and 161 | Srikumar, Vivek", 162 | booktitle = "Findings of the Association for Computational Linguistics ACL 2024", 163 | month = aug, 164 | year = "2024", 165 | address = "Bangkok, Thailand and virtual meeting", 166 | publisher = "Association for Computational Linguistics", 167 | url = "https://aclanthology.org/2024.findings-acl.471", 168 | doi = "10.18653/v1/2024.findings-acl.471", 169 | pages = "7907--7928" 170 | } 171 | 172 | ``` 173 | 174 | -------------------------------------------------------------------------------- /ncb/execution.py: -------------------------------------------------------------------------------- 1 | import os 2 | import re 3 | import time 4 | from collections import defaultdict 5 | from pathlib import Path 6 | import numpy as np 7 | from tqdm import tqdm 8 | import subprocess 9 | from concurrent.futures import ThreadPoolExecutor, as_completed 10 | from utils import load_jsonl, save_jsonl, load_json, save_json, estimate_pass_at_k 11 | from exec_java import eval_test 12 | 13 | 14 | def run_python_test(base_dir, test_dir): 15 | with open(Path(base_dir, test_dir, 'log.txt'), 'wb') as fp: 16 | subprocess.run(['python', 'ncb/exec_python.py', '--base_dir', base_dir, '--test_dir', test_dir], 17 | stdout=fp, stderr=fp, timeout=840) 18 | 19 | 20 | def execution(base_dir, ckpt, language, natural_lang, dataset_size, ks, num_workers=16, debug=False): 21 | start_time = time.time() 22 | work_dir = os.getcwd() 23 | base_dir = Path(work_dir, base_dir) 24 | test_dirs = os.listdir(base_dir) 25 | if 'error.jsonl' in test_dirs: 26 | test_dirs.remove('error.jsonl') 27 | if 'result.json' in test_dirs: 28 | test_dirs.remove('result.json') 29 | if 'raw_result.json' in test_dirs: 30 | test_dirs.remove('raw_result.json') 31 | if 'error_problems.jsonl' in test_dirs: 32 | test_dirs.remove('error_problems.jsonl') 33 | if language == 'python': 34 | with ThreadPoolExecutor(max_workers=num_workers) as executor: 35 | futures = [executor.submit(run_python_test, base_dir, _dir) for _dir in test_dirs] 36 | 37 | for future in tqdm(as_completed(futures), total=len(futures)): 38 | try: 39 | result = future.result() 40 | except subprocess.TimeoutExpired as e: 41 | print(e) 42 | 43 | os.chdir(work_dir) 44 | overall_result = {'total': 0, 'passed': 0, 'failed': 0} 45 | pass_statistics = {'0': 0, '0_30': 0, '30_60': 0, '60_100': 0, '100': 0} 46 | error = {'SyntaxError': 0, 'RuntimeError': 0} 47 | error_report = [] 48 | error_problems = [] 49 | results = defaultdict(list) 50 | for _dir in test_dirs: 51 | try: 52 | error_report.extend(load_jsonl(base_dir / _dir / 'error_report.jsonl')) 53 | error_cases = load_json(base_dir / _dir / 'error.json') 54 | for key in error.keys(): 55 | error[key] += error_cases[key] 56 | raw_result = load_json(base_dir / _dir / 'raw_result.json') 57 | for key in overall_result.keys(): 58 | overall_result[key] += raw_result['overall_result'][key] 59 | for key in pass_statistics.keys(): 60 | pass_statistics[key] += raw_result['pass_cases'][key] 61 | 62 | _id = str(_dir)[str(_dir).find('_')+1:str(_dir).rfind('_')] 63 | results[_id].append(raw_result['pass_cases']) 64 | 65 | if error_cases['RuntimeError'] > 0: 66 | if debug: 67 | print("RuntimeError: ", _dir) 68 | error_problems.append({'problem': _dir, 'error': 'RuntimeError'}) 69 | elif error_cases['SyntaxError'] > 0: 70 | if debug: 71 | print("SyntaxError: ", _dir) 72 | error_problems.append({'problem': _dir, 'error': 'SyntaxError'}) 73 | elif raw_result['pass_cases']['100'] <= 0: 74 | if debug: 75 | print("Not Pass All: ", _dir) 76 | error_problems.append({'problem': _dir, 'error': 'Not Pass All'}) 77 | except Exception as e: 78 | _id = str(_dir)[str(_dir).find('_') + 1:str(_dir).rfind('_')] 79 | results[_id].append({'0': 1, '0_30': 0, '30_60': 0, '60_100': 0, '100': 0}) 80 | print(f"Error occur in {_dir}, error: {e}") 81 | 82 | total, correct = [], [] 83 | for result in results.values(): 84 | passed = [r['100'] for r in result] 85 | total.append(len(passed)) 86 | correct.append(sum(passed)) 87 | total, correct = np.array(total), np.array(correct) 88 | pass_at_k = {f"pass@{k}": round(estimate_pass_at_k(total, correct, k).mean() * 100, 1) 89 | for k in ks if (total >= k).all()} 90 | 91 | end_time = time.time() - start_time 92 | print(f"Total Running time {end_time}") 93 | save_json({'overall_result': overall_result, 'pass_cases': pass_statistics, 'error_cases': error}, 94 | base_dir / 'raw_result.json') 95 | save_jsonl(error_problems, base_dir / 'error_problems.jsonl') 96 | save_jsonl(error_report, base_dir / 'error.jsonl') 97 | return {'result': {'pass@k': pass_at_k, 98 | 'time': end_time, 99 | }, 100 | 'error': {'SyntaxError': round(error['SyntaxError'] / dataset_size * 100, 1), 101 | 'RuntimeError': round(error['RuntimeError'] / dataset_size * 100, 1)}, 102 | 'ckpt_name': ckpt, 103 | 'benchmark': f'ncb_{language}_{natural_lang}'} 104 | elif language == 'java': 105 | JUnit_path = Path(os.getcwd(), 'JUnit') 106 | reports = [] 107 | with ThreadPoolExecutor(max_workers=num_workers) as executor: 108 | futures = [] 109 | for _dir in test_dirs: 110 | if _dir == '.DS_Store' or _dir.endswith('.jsonl') or _dir == 'java' or _dir == 'java_test': 111 | continue 112 | 113 | solution_dir = base_dir / 'java' 114 | test_dir = base_dir / 'java_test' 115 | os.makedirs(solution_dir, exist_ok=True) 116 | os.makedirs(test_dir, exist_ok=True) 117 | futures.append(executor.submit(eval_test, _dir, base_dir, JUnit_path, 420)) 118 | 119 | for future in tqdm(as_completed(futures), total=len(futures)): 120 | reports.append(future.result()) 121 | 122 | found_pattern = f"Found (.*?) tests" 123 | failed_pattern = f"Failed (.*?) tests" 124 | pr_pattern = f"Passed rate: (.*?)\n" 125 | total_c = 0 126 | suc = 0 127 | pr_0 = 0 128 | pr_0_3 = 0 129 | pr_3_6 = 0 130 | pr_6_10 = 0 131 | pr_1 = 0 132 | cp_error = 0 133 | exe_error = 0 134 | results = defaultdict(list) 135 | error_problems = [] 136 | for i in reports: 137 | _id = i['_id'] 138 | if 'failed' not in i['test_result'] and 'Found' in i['test_result']: 139 | total_c += int(re.search(found_pattern, i['test_result'], re.DOTALL).group(1)) 140 | suc += int(re.search(found_pattern, i['test_result'], re.DOTALL).group(1)) - int( 141 | re.search(failed_pattern, i['test_result'], re.DOTALL).group(1)) 142 | ma = re.search(pr_pattern, i['test_result'], re.DOTALL) 143 | pr = float(ma.group(1)) 144 | if pr == 0: 145 | pr_0 += 1 146 | elif pr > 0 and pr < 0.3: 147 | pr_0_3 += 1 148 | elif pr >= 0.3 and pr < 0.6: 149 | pr_3_6 += 1 150 | elif pr >= 0.6 and pr < 1: 151 | pr_6_10 += 1 152 | elif pr == 1: 153 | pr_1 += 1 154 | results[_id].append(pr == 1) 155 | 156 | if pr != 1: 157 | if debug: 158 | print("Not Pass All:", i['class_name']) 159 | error_problems.append({'problem': i['class_name'], 'error': 'Not Pass All'}) 160 | else: 161 | if i['compile_error'] == "" and i['exe_error'] != "": 162 | if debug: 163 | print("EXEC_error:", i['class_name']) 164 | exe_error += 1 165 | error_problems.append({'problem': i['class_name'], 'error': 'EXEC_error'}) 166 | else: 167 | if debug: 168 | print("CP_error:", i['class_name']) 169 | cp_error += 1 170 | error_problems.append({'problem': i['class_name'], 'error': 'CP_error'}) 171 | 172 | results[_id].append(False) 173 | 174 | total, correct = [], [] 175 | for result in results.values(): 176 | passed = [int(r) for r in result] 177 | total.append(len(passed)) 178 | correct.append(sum(passed)) 179 | total, correct = np.array(total), np.array(correct) 180 | pass_at_k = {f"pass@{k}": round(estimate_pass_at_k(total, correct, k).mean() * 100, 1) 181 | for k in ks if (total >= k).all()} 182 | 183 | raw_result = {'overall_result': { 184 | 'total': total_c, 185 | 'passed': suc, 186 | }, 187 | 'pass_cases': { 188 | '0': pr_0, 189 | '0_30': pr_0_3, 190 | '30_60': pr_3_6, 191 | '60_100': pr_6_10, 192 | '100': pr_1 193 | }} 194 | save_json(raw_result, base_dir / 'raw_result.json') 195 | save_jsonl(error_problems, base_dir / 'error_problems.jsonl') 196 | end_time = time.time() - start_time 197 | print(f"Total Running time {end_time}") 198 | return {'result': {'pass@k': pass_at_k, 199 | 'time': end_time}, 200 | 'error': {'CompileError': round(cp_error / dataset_size * 100, 1), 201 | 'ExecutionError': round(exe_error / dataset_size * 100, 1)}, 202 | 'ckpt_name': ckpt, 203 | 'benchmark': f'ncb_{language}_{natural_lang}'} 204 | -------------------------------------------------------------------------------- /ncb/test_setup.py: -------------------------------------------------------------------------------- 1 | import os 2 | import re 3 | import shutil 4 | from pathlib import Path 5 | from tqdm import tqdm 6 | from utils import save_json 7 | 8 | 9 | def extract_python_code(response): 10 | tag_pattern = fr"\[Python\](.*?)\[/Python\]" 11 | tag_pattern1 = fr"\[Python\](.*?)```" 12 | tag_pattern2 = fr"\[Python\](.*?)" 13 | py_pattern1 = fr"```python[\n\r](.*?)[\n\r]```" 14 | py_pattern2 = fr"```[\n\r](.*?)[\n\r]```" 15 | py_pattern3 = fr"def\s(.*)" 16 | py_pattern4 = fr"class\s(.*)" 17 | py_pattern5 = fr"import\s(.*)" 18 | py_pattern6 = fr"from\s(.*)" 19 | 20 | try: 21 | judge = re.search(tag_pattern, response, re.DOTALL) 22 | if judge and '```' not in judge.group(1): 23 | code = judge.group(1) 24 | elif not judge and '[Python]' in response and '```python' not in response: 25 | if '```' in response[response.find('[Python]'):]: 26 | code = re.search(tag_pattern1, response, re.DOTALL).group(1) 27 | else: 28 | code = re.search(tag_pattern2, response, re.DOTALL).group(1) 29 | elif '```python' in response: 30 | code = re.search(py_pattern1, response, re.DOTALL).group(1) 31 | elif '```' in response: 32 | code = re.search(py_pattern2, response, re.DOTALL).group(1) 33 | elif 'from ' in response and 'import ' in response: 34 | if response.find('from') < response.find('import'): 35 | code = re.search(py_pattern6, response, re.DOTALL).group() 36 | else: 37 | code = re.search(py_pattern5, response, re.DOTALL).group() 38 | elif 'import ' in response: 39 | code = re.search(py_pattern5, response, re.DOTALL).group() 40 | elif 'from ' in response: 41 | code = re.search(py_pattern6, response, re.DOTALL).group() 42 | elif 'class ' in response: 43 | code = re.search(py_pattern4, response, re.DOTALL).group() 44 | elif 'def ' in response: 45 | code = re.search(py_pattern3, response, re.DOTALL).group() 46 | else: 47 | code = 'NaN' 48 | except Exception as e: 49 | code = 'NaN' 50 | code = 'import pytest\n' + code 51 | return code 52 | 53 | 54 | def extract_java_code(response): 55 | tag_pattern = fr"\[Java\](.*?)\[/Java\]" 56 | java_pattern1 = fr"```java[\n\r](.*?)[\n\r]```" 57 | java_pattern2 = fr"public\s(.*)}}" 58 | java_pattern3 = fr"```Java(.*?)```" 59 | java_pattern4 = fr"```[\n\r](.*?)[\n\r]```" 60 | java_pattern5 = fr"import\s(.*)}}" 61 | java_pattern6 = fr"class\s(.*)}}" 62 | java_pattern7 = fr"interface\s(.*)}}" 63 | 64 | try: 65 | if '[Java]' in response: 66 | code = [re.search(tag_pattern, response, re.DOTALL).group(1)] 67 | elif '```java' in response or '```Java' in response or '```' in response: 68 | code = re.findall(java_pattern1, response, re.DOTALL) 69 | for c in code: 70 | response = response.replace("```java"+c+"```", "") 71 | code += re.findall(java_pattern3, response, re.DOTALL) 72 | for c in code: 73 | response = response.replace("```Java"+c+"```", "") 74 | code += re.findall(java_pattern4, response, re.DOTALL) 75 | elif 'import ' in response: 76 | code = [re.search(java_pattern5, response, re.DOTALL).group()] 77 | elif 'public ' in response or 'interface ' in response or 'class ' in response: 78 | public_ind = response.find('public ') 79 | interface_ind = response.find('interface ') 80 | class_ind = response.find('class ') 81 | if public_ind != -1 and \ 82 | (public_ind < interface_ind or interface_ind == -1) and \ 83 | (public_ind < class_ind or class_ind == -1): 84 | code = [re.search(java_pattern2, response, re.DOTALL).group()] 85 | elif interface_ind != -1 and \ 86 | (interface_ind < public_ind or public_ind == -1) and \ 87 | (interface_ind < class_ind or class_ind == -1): 88 | code = [re.search(java_pattern7, response, re.DOTALL).group()] 89 | elif class_ind != -1 and \ 90 | (class_ind < public_ind or public_ind) and \ 91 | (class_ind < interface_ind or interface_ind == -1): 92 | code = [re.search(java_pattern6, response, re.DOTALL).group()] 93 | else: 94 | code = "NaN" 95 | print(response) 96 | else: 97 | code = ['NaN'] 98 | except Exception as e: 99 | code = ['NaN'] 100 | return code 101 | 102 | 103 | def extract_codes(data, testcases, language): 104 | for item in data: 105 | for ex in testcases: 106 | if item['_id'] == ex['_id']: 107 | item['testcases'] = ex['testcases'] 108 | item['setup_code'] = ex['setup_code'] 109 | break 110 | 111 | if 'code' not in item.keys(): 112 | if language == 'java': 113 | item['code'] = extract_java_code(item['response']) 114 | elif language == 'python': 115 | item['code'] = extract_python_code(item['response']) 116 | return data 117 | 118 | 119 | def write_python_test_files(data, data_dir, input_files_path, ckpt): 120 | ids = {} 121 | for item in data: 122 | _id = str(item['_id']) 123 | if not item['response'] or item['code'] == 'NaN': 124 | print(f"No extracted code: python_{item['_id']}") 125 | continue 126 | file_name = f'python_{_id}' 127 | if _id in ids: 128 | ids[_id] += 1 129 | id_name = _id + '_' + str(ids[_id]) 130 | else: 131 | ids[_id] = 0 132 | id_name = _id + '_' + str(ids[_id]) 133 | 134 | 135 | test_dir = data_dir / ckpt / f'python_{id_name}' 136 | os.makedirs(test_dir, exist_ok=True) 137 | 138 | save_json(item, test_dir / f'{file_name}.json') 139 | 140 | with open(test_dir / f'{file_name}.py', 'w', encoding='utf-8') as f: 141 | f.write("import pytest\nimport sys\nwith open('stdin.txt', 'w') as f:\n for i in range(1000):\n f.write(\"1\\n\")\nsys.stdin = open(\"stdin.txt\")\n") 142 | f.write(item['code'] + '\n') 143 | 144 | files_for_prob_dir = input_files_path / str(_id) 145 | if os.path.exists(files_for_prob_dir): 146 | for file in os.listdir(files_for_prob_dir): 147 | file_for_prob = files_for_prob_dir / str(file) 148 | if os.path.isfile(file_for_prob): 149 | shutil.copy(file_for_prob, test_dir / str(file)) 150 | elif os.path.isdir(file_for_prob): 151 | shutil.copytree(file_for_prob, test_dir / str(file)) 152 | 153 | with open(test_dir / f'test_{file_name}.py', 'w', encoding='utf-8') as f: 154 | f.write(f"import pytest\nfrom {file_name} import *\n") 155 | f.write(item['testcases'] + '\n' + item['setup_code']) 156 | 157 | 158 | def write_java_test_files(data, data_dir, input_files_path, ckpt): 159 | Testrunner = """import org.junit.platform.launcher.Launcher; 160 | import org.junit.platform.launcher.LauncherDiscoveryRequest; 161 | import org.junit.platform.launcher.TestPlan; 162 | import org.junit.platform.launcher.TestExecutionListener; 163 | import org.junit.platform.launcher.core.LauncherDiscoveryRequestBuilder; 164 | import org.junit.platform.launcher.core.LauncherFactory; 165 | import org.junit.platform.launcher.listeners.SummaryGeneratingListener; 166 | import org.junit.platform.launcher.listeners.TestExecutionSummary; 167 | 168 | import static org.junit.platform.engine.discovery.DiscoverySelectors.selectClass; 169 | 170 | public class TestRunner { 171 | public static void main(String[] args) { 172 | LauncherDiscoveryRequest request = LauncherDiscoveryRequestBuilder.request() 173 | .selectors(selectClass(YourTestClass.class)) // 替换为你的测试类 174 | .build(); 175 | 176 | Launcher launcher = LauncherFactory.create(); 177 | 178 | // 创建和 注册 SummaryGeneratingListener 179 | SummaryGeneratingListener listener = new SummaryGeneratingListener(); 180 | launcher.registerTestExecutionListeners(listener); 181 | 182 | // 执行测试 183 | launcher.execute(request); 184 | 185 | // 获得测试的结果汇总信息 186 | TestExecutionSummary summary = listener.getSummary(); 187 | long testsFoundCount = summary.getTestsFoundCount(); 188 | long testsFailedCount = summary.getTestsFailedCount(); 189 | long testsSucceedCount = summary.getTestsSucceededCount(); 190 | System.out.println("Found " + testsFoundCount + " tests"); 191 | System.out.println("Failed " + testsFailedCount + " tests"); 192 | System.out.println("Passed rate: " + (double)testsSucceedCount/testsFoundCount); 193 | } 194 | }""" 195 | ids = {} 196 | for index, item in tqdm(enumerate(data)): 197 | _id = str(item['_id']) 198 | if not item['response'] or item['code'] == 'NaN': 199 | print(f"no code: java_{item['_id']}") 200 | continue 201 | if _id in ids: 202 | ids[_id] += 1 203 | id_name = _id + '_' + str(ids[_id]) 204 | else: 205 | ids[_id] = 0 206 | id_name = _id + '_' + str(ids[_id]) 207 | test_dir = data_dir / ckpt / f"java_{id_name}" 208 | os.makedirs(test_dir, exist_ok=True) 209 | 210 | save_json(item, test_dir / f'java_{_id}.json') 211 | 212 | file_c = 0 213 | for c in item['code']: 214 | c = '\n' + c 215 | import_pattern = r'(\n|^)(import .*?)\n' 216 | interface_pattern = r"((@.*?)?(\n[^\n]*)?interface .*?[;}]\s*\n+})" 217 | class_pattern = r"((@.*?)?(\n[^\n]*)?class .*?[;}]\s*\n+})" 218 | enum_pattern = r"((@.*?)?(\n[^\n]*)?enum .*?[;}]?\s*\n+})" 219 | 220 | imports = re.findall(import_pattern, c, re.MULTILINE) 221 | imports = [i[1] for i in imports] 222 | interfaces = re.findall(interface_pattern, c, re.DOTALL) 223 | classes = re.findall(class_pattern, c, re.DOTALL) 224 | enums = re.findall(enum_pattern, c, re.DOTALL) 225 | 226 | for it in interfaces: 227 | it = it[0] 228 | interface_name_pattern = r'interface (.*?)\s' 229 | name = re.search(interface_name_pattern, it, re.DOTALL).group(1) 230 | with open(test_dir / f'{name}.java', 'w', encoding='utf-8') as f: 231 | f.write("import org.junit.jupiter.api.Test;\nimport static org.junit.jupiter.api.Assertions.*;\n") 232 | f.write("\n".join(imports) + "\n") 233 | f.write(it+'\n') 234 | file_c += 1 235 | 236 | for cls in classes: 237 | cls = cls[0] 238 | class_name_pattern = r'class (.*?)\s' 239 | name = re.search(class_name_pattern, cls, re.DOTALL).group(1) 240 | with open(test_dir / f'{name}.java', 'w', encoding='utf-8') as f: 241 | f.write("import org.junit.jupiter.api.Test;\nimport static org.junit.jupiter.api.Assertions.*;\n") 242 | f.write("\n".join(imports) + "\n") 243 | f.write(cls+'\n') 244 | file_c += 1 245 | 246 | for enum in enums: 247 | enum = enum[0] 248 | enum_name_pattern = r'enum (.*?)\s' 249 | name = re.search(enum_name_pattern, enum, re.DOTALL).group(1) 250 | with open(test_dir / f'{name}.java', 'w', encoding='utf-8') as f: 251 | f.write("import org.junit.jupiter.api.Test;\nimport static org.junit.jupiter.api.Assertions.*;\n") 252 | f.write("\n".join(imports) + "\n") 253 | f.write(enum + '\n') 254 | file_c += 1 255 | 256 | files_for_prob_dir = input_files_path / str(_id) 257 | if os.path.exists(files_for_prob_dir): 258 | for file in os.listdir(files_for_prob_dir): 259 | file_for_prob = files_for_prob_dir / str(file) 260 | if os.path.isfile(file_for_prob): 261 | shutil.copy(file_for_prob, test_dir / str(file)) 262 | elif os.path.isdir(file_for_prob): 263 | shutil.copytree(file_for_prob, test_dir / str(file)) 264 | 265 | testcases = item['testcases'] 266 | test_file_name = test_dir / f'Test_java_{_id}.java' 267 | test_file_name_pattern = fr"class (.*?)\s" 268 | match4 = re.search(test_file_name_pattern, testcases, re.DOTALL) 269 | with open(test_file_name, 'w', encoding='utf-8') as f: 270 | tmp = testcases.replace(match4.group(1), f'Test_java_{_id}') 271 | f.write(tmp) 272 | testrunner_file = test_dir / 'TestRunner.java' 273 | tr_code = Testrunner.replace('YourTestClass', f'Test_java_{_id}') 274 | with open(testrunner_file, 'w') as f: 275 | f.write(tr_code) 276 | 277 | 278 | def write_test_files(data_dir, input_files_path, data, language, ckpt): 279 | os.makedirs(data_dir / ckpt, exist_ok=True) 280 | if language == 'python': 281 | write_python_test_files(data, data_dir, input_files_path, ckpt) 282 | elif language == 'java': 283 | write_java_test_files(data, data_dir, input_files_path, ckpt) 284 | return ckpt 285 | -------------------------------------------------------------------------------- /results/reference/reference_ncb_python_en.jsonl: -------------------------------------------------------------------------------- 1 | {"_id": 131, "response": "```python\nimport string\nfrom collections import Counter\n\n\ndef word_count(file_path):\n try:\n # 读取文件内容\n with open(file_path, 'r', encoding='utf-8') as file:\n content = file.read()\n\n # 移除标点符号并转换为小写\n translator = str.maketrans(\"\", \"\", string.punctuation)\n content = content.translate(translator).lower()\n\n # 使用 Counter 统计单词出现次数\n words = content.split()\n word_counter = Counter(words)\n\n # 按照出现次数降序排列\n sorted_word_count = sorted(word_counter.items(), key=lambda x: x[1], reverse=True)\n\n for word, count in sorted_word_count:\n print(f\"'{word}': {count}\")\n\n except FileNotFoundError:\n print(f\"Error: File '{file_path}' not found.\")\n```"} 2 | {"_id": 132, "response": "```python\nimport re\n\n\ndef remove_html_tags(s: str) -> str:\n if not isinstance(s, str):\n raise ValueError(\"Input must be a string\")\n\n if not s:\n return \"\"\n\n # 定义新的正则表达式模式\n tag_pattern = r\"<[^>]*>\"\n comment_pattern = r\"\"\n doctype_pattern = r\"]*>\"\n\n # 移除 HTML 注释和 DOCTYPE\n s = re.sub(comment_pattern, \"\", s)\n s = re.sub(doctype_pattern, \"\", s)\n\n # 移除标签\n s = re.sub(tag_pattern, \"\", s)\n\n # 返回处理后的字符串\n return s.strip()\n\n\n\n```"} 3 | {"_id": 133, "response": "```python\nfrom datetime import datetime\n\nclass Schedule:\n def __init__(self):\n self.events = {}\n\n def add_event(self, event_name, event_date):\n if event_date in self.events:\n self.events[event_date].append(event_name)\n else:\n self.events[event_date] = [event_name]\n\n def view_events(self):\n for event_date, event_names in self.events.items():\n print(f\"{event_date.strftime('%Y-%m-%d %H:%M')}:\")\n for event_name in event_names:\n print(f\"- {event_name}\")\n print()\n\n def remove_event(self, event_name, event_date):\n if event_date in self.events:\n if event_name in self.events[event_date]:\n self.events[event_date].remove(event_name)\n if not self.events[event_date]:\n del self.events[event_date]\n print(f\"已删除事件：{event_name}\")\n return\n print(f\"未找到事件'{event_name}'\")\n\n\n```"} 4 | {"_id": 134, "response": "```python\nclass ListNode:\n def __init__(self, value=0, next=None):\n self.value = value\n self.next = next\n\n\ndef add_two_numbers(l1, l2):\n carry = 0\n dummy_head = ListNode()\n current = dummy_head\n\n while l1 or l2 or carry:\n # 获取当前节点的值\n val1 = l1.value if l1 else 0\n val2 = l2.value if l2 else 0\n\n # 计算和以及进位\n total = val1 + val2 + carry\n carry, digit = divmod(total, 10)\n\n # 创建新节点并更新当前节点的next\n current.next = ListNode(digit)\n current = current.next\n\n # 移动到下一个节点\n if l1:\n l1 = l1.next\n if l2:\n l2 = l2.next\n\n return dummy_head.next\n```"} 5 | {"_id": 135, "response": "```python\nfrom typing import List\nimport xml.etree.ElementTree as ET\n\ndef parse_xml(file_path: str, tag_name: str) -> List[str]:\n\n try:\n root = ET.parse(file_path).getroot()\n except ET.ParseError:\n return []\n data_list = []\n for element in root.findall(tag_name):\n data_list.append(element.text)\n return data_list\n\n```"} 6 | {"_id": 136, "response": "```python\nimport re\n\n\ndef is_valid_variable_name(name):\n \"\"\"\n 判断字符串是否是有效的 JavaScript 变量名\n \"\"\"\n if len(name) > 255:\n return False\n if not re.match(r'^[a-zA-Z_$][a-zA-Z0-9_$]*$', name):\n return False\n return True\n\n```"} 7 | {"_id": 137, "response": "```python\nclass Order:\n def __init__(self, order_id):\n self.order_id = order_id\n self.items = []\n self.total_price = 0.0\n self.status = \"未支付\"\n\n def add_item(self, item_name, item_price):\n if not isinstance(item_price, (int, float)) or item_price <= 0:\n return \"商品价格必须是正数\"\n\n item = {\"name\": item_name, \"price\": item_price}\n self.items.append(item)\n\n def remove_item(self, item_name):\n for item in self.items:\n if item[\"name\"] == item_name:\n self.items.remove(item)\n break\n else:\n return \"该商品不存在于订单中\"\n\n def calculate_total_price(self):\n self.total_price = sum(item[\"price\"] for item in self.items)\n\n def pay_order(self):\n if not self.items:\n return \"订单中没有商品\"\n self.status = \"已支付\"\n\n def cancel_order(self):\n self.status = \"已取消\"\n\n def view_order(self):\n print(\"订单编号：\", self.order_id)\n print(\"商品列表：\")\n for item in self.items:\n print(\"- {}: ￥{}\".format(item[\"name\"], item[\"price\"]))\n print(\"总价格：￥{}\".format(self.total_price))\n print(\"状态：\", self.status)\n```"} 8 | {"_id": 138, "response": "```python\ndef is_valid_sequence(s):\n # 判断字符串长度是否在 2 到 10 之间\n if len(s) < 2 or len(s) > 10:\n return False\n\n # 判断字符串中的字符是否都是小写字母\n for char in s:\n if not char.islower():\n return False\n\n # 判断字符串中是否出现连续的重复字符\n for i in range(len(s) - 1):\n if s[i] == s[i + 1]:\n return False\n\n # 判断左右两部分是否都是回文字符串或递增序列\n left, right = s[:len(s) // 2], s[len(s) // 2:]\n if left == right[::-1] or sorted(left) == sorted(right):\n return True\n else:\n return False\n```"} 9 | {"_id": 139, "response": "```python\nimport math\n\nclass RouteSimulator:\n def __init__(self, x, y):\n self.x = x\n self.y = y\n self.move_history = []\n\n def move(self, distance, angle_degrees):\n # 将角度转换为弧度\n angle_radians = math.radians(angle_degrees)\n\n # 计算新的位置\n new_x = self.x + distance * math.cos(angle_radians)\n new_y = self.y + distance * math.sin(angle_radians)\n\n # 更新位置\n self.x, self.y = new_x, new_y\n\n # 记录移动历史\n self.move_history.append({'distance': distance, 'angle_degrees': angle_degrees, 'position': (new_x, new_y)})\n\n def calculate_distance(self, x1, y1, x2, y2):\n # 计算两点之间的距离\n return math.sqrt((x2 - x1)**2 + (y2 - y1)**2)\n\n def print_current_position(self):\n print(f\"当前位置：({self.x:.2f}, {self.y:.2f})\")\n\n def print_move_history(self):\n print(\"移动历史：\")\n for move in self.move_history:\n distance = move['distance']\n angle_degrees = move['angle_degrees']\n position = move['position']\n print(f\"距离: {distance:.2f}, 角度: {angle_degrees:.2f}, 位置: ({position[0]:.2f}, {position[1]:.2f})\")\n\n```"} 10 | {"_id": 140, "response": "```python\ndef utf8_to_ascii(utf8_string):\n # 验证输入字符串是否为空\n if not utf8_string:\n print(\"输入字符串不能为空！\")\n return None\n\n result = []\n\n try:\n # 遍历输入字符串的每个字符\n for i in range(len(utf8_string)):\n if utf8_string[i] == '%' and i + 2 < len(utf8_string) and utf8_string[i + 1:i + 3].isalnum():\n # 如果遇到以%开头的16进制字符编码，则进行解析\n hex_str = utf8_string[i + 1:i + 3]\n ascii_code = int(hex_str, 16)\n result.append(ascii_code)\n i += 2\n else:\n # 普通的UTF-8字符直接转换为ASCII码\n ascii_code = ord(utf8_string[i])\n result.append(ascii_code)\n except IndexError:\n print(\"输入字符串格式错误！\")\n return None\n except ValueError:\n print(\"输入字符串包含非法的字符编码！\")\n return None\n except Exception as e:\n print(\"发生未知错误：\", str(e))\n return None\n\n return result\n\n\n\n```"} 11 | {"_id": 141, "response": "```python\nimport re\nimport datetime\n\n\ndef find_dates(text):\n if not text:\n return \"输入的文本为空，请提供有效的文本内容\"\n\n date_patterns = [\n r\"\\b(\\d{4})-(0[1-9]|1[0-2])-(0[1-9]|[12][0-9]|3[01])\\b\", # YYYY-MM-DD\n r\"\\b(0[1-9]|1[0-2])/(0[1-9]|[12][0-9]|3[01])/(19|20)?\\d{2}\\b\", # MM/DD/YYYY or M/D/YYYY\n r\"\\b(0[1-9]|[12][0-9]|3[01])-(0[1-9]|1[0-2])-(19|20)?\\d{2}\\b\", # DD-MM-YYYY or D-M-YYYY\n r\"\\b(0[1-9]|[12][0-9]|3[01])\\s(January|February|March|April|May|June|July|August|September|October|November|December),?\\s(19|20)?\\d{2}\\b\"\n # DD Month YYYY or D Month YYYY\n ]\n dates = []\n for pattern in date_patterns:\n dates.extend(re.findall(pattern, text))\n\n formatted_dates = []\n for date in dates:\n if date[0]: # 如果匹配到了年份\n year = date[0]\n else:\n year = \"20\" + date[3] if date[3] else \"\" # 如果没有匹配到年份，则根据模式中的分组确定年份\n\n month = date[1].zfill(2)\n day = date[2].zfill(2)\n\n # 检查日期的有效性\n try:\n formatted_date = f\"{year}-{month}-{day}\"\n _ = datetime.datetime.strptime(formatted_date, \"%Y-%m-%d\")\n formatted_dates.append(formatted_date)\n except ValueError:\n pass\n\n if not formatted_dates:\n return \"未找到有效日期\"\n\n return formatted_dates\n\n```"} 12 | {"_id": 142, "response": "```python\nclass Process:\n def __init__(self, pid, code):\n self.pid = pid\n self.code = code\n self.memory = None\n\ndef allocate_memory(memory_map, process):\n if process.memory is not None:\n raise ValueError(\"Process already has memory\")\n\n for i in range(len(memory_map)):\n if memory_map[i] == 0:\n process.memory = i\n for j in range(len(process.code)):\n memory_map[i + j] = 1\n return True\n\n return False\n\ndef deallocate_memory(memory_map, process):\n if process.memory is not None:\n for i in range(len(process.code)):\n memory_map[process.memory + i] = 0\n process.memory = None\n\ndef run_processes(processes):\n for process in processes:\n print(f\"Running Process {process.pid} with code: {process.code}\")\n\n# 创建一个简单的操作系统\nmemory_map = [0] * 100\nprocesses = []\n\n# 创建两个进程\nprocess1 = Process(pid=1, code=[1, 2, 3])\nprocess2 = Process(pid=2, code=[4, 5, 6])\n\n# 将进程添加到列表\nprocesses.append(process1)\nprocesses.append(process2)\n\n# 分配内存\nallocate_memory(memory_map, process1)\nallocate_memory(memory_map, process2)\n\n# 运行进程\nrun_processes(processes)\n\n# 释放内存\ndeallocate_memory(memory_map, process1)\ndeallocate_memory(memory_map, process2)\n\n```"} 13 | {"_id": 143, "response": "```python\nimport re\nfrom collections import Counter\nfrom bs4 import BeautifulSoup\n\n\ndef get_top_k_words_from_html(html, k):\n # 定义需要排除的单词列表\n exclude_list = ['this','is','are','a', 'an', 'the', 'and', 'or', 'in', 'on', 'at', 'to', 'of', 'for', 'with', 'by', 'from']\n\n try:\n # 解析HTML文本\n soup = BeautifulSoup(html, 'lxml')\n\n # 获取所有文本内容\n text = soup.get_text()\n\n # 将文本转换为小写，并使用正则表达式提取单词\n\n lower=text.lower()\n words = re.findall(r'\\b(? Tuple[List[str], List[float]]:\n cities = []\n temperatures = []\n lines = text.strip().split('\\n')\n for line in lines:\n city_temperature = line.split('，')\n if len(city_temperature) == 2:\n city, temperature = city_temperature\n temperature = float(temperature.replace('℃', ''))\n cities.append(city)\n temperatures.append(temperature)\n else:\n return [], []\n return cities, temperatures\n\ndef calculate_average_temperature(temperatures: List[float]) -> float:\n total_temperature = sum(temperatures)\n average_temperature = total_temperature / len(temperatures)\n return average_temperature\n\ndef find_highest_temperature(cities: List[str], temperatures: List[float]) -> Tuple[float, str]:\n if not temperatures:\n return 0, \"无\"\n index = temperatures.index(max(temperatures))\n highest_temperature = temperatures[index]\n city = cities[index]\n return highest_temperature, city\n\ndef find_lowest_temperature(cities: List[str], temperatures: List[float]) -> Tuple[float, str]:\n if not temperatures:\n return 0, \"无\"\n index = temperatures.index(min(temperatures))\n lowest_temperature = temperatures[index]\n city = cities[index]\n return lowest_temperature, city\n\n\n```"} 15 | {"_id": 145, "response": "```python\nimport re\n\n\ndef count_html_tags(html):\n pattern = r'<\\s*([a-zA-Z]+)[^>]*>|<\\s*/\\s*([a-zA-Z]+)\\s*>'\n matches = re.findall(pattern, html)\n\n tag_count = 0\n nested_tags = []\n max_nested_depth = 0\n\n for opening_tag, closing_tag in matches:\n if opening_tag:\n nested_tags.append(opening_tag.lower())\n tag_count += 1\n max_nested_depth = max(max_nested_depth, len(nested_tags))\n elif nested_tags and nested_tags[-1] == closing_tag.lower():\n nested_tags.pop()\n tag_count += 1\n\n return tag_count, max_nested_depth\n\n\n\n```"} 16 | {"_id": 146, "response": "```python\nfrom cryptography.hazmat.primitives.ciphers import Cipher, algorithms, modes\nfrom cryptography.hazmat.backends import default_backend\nimport base64\n\n\ndef encrypt(text, key):\n backend = default_backend()\n cipher = Cipher(algorithms.AES(key), modes.ECB(), backend=backend)\n encryptor = cipher.encryptor()\n\n # 补全文本长度至AES块大小的倍数\n text = text + ' ' * (16 - len(text) % 16)\n\n # 加密\n ciphertext = encryptor.update(text.encode()) + encryptor.finalize()\n return base64.b64encode(ciphertext).decode()\n\n\ndef decrypt(encrypted_text, key):\n backend = default_backend()\n cipher = Cipher(algorithms.AES(key), modes.ECB(), backend=backend)\n decryptor = cipher.decryptor()\n\n # 解密\n ciphertext = base64.b64decode(encrypted_text.encode())\n decrypted_text = decryptor.update(ciphertext) + decryptor.finalize()\n return decrypted_text.decode().rstrip()\n\n\ndef save_to_file(data, filename):\n with open(filename, 'w') as file:\n file.write(data)\n\n\ndef read_from_file(filename):\n with open(filename, 'r') as file:\n return file.read()\n\n\n\ntext_to_encrypt = \"Hello, World!\"\nencryption_key = b'Sixteen byte key'\n\n# # 加密并保存到文件\n# encrypted_text = encrypt(text_to_encrypt, encryption_key)\n# save_to_file(encrypted_text, \"encrypted_data.txt\")\n#\n# # 从文件中读取并解密\n# loaded_encrypted_text = read_from_file(\"encrypted_data.txt\")\n# decrypted_text = decrypt(loaded_encrypted_text, encryption_key)\n\n\n\n```"} 17 | {"_id": 147, "response": "```python\nimport os\nimport shutil\n\ndef backup_folder(source_folder, target_folder):\n try:\n # 检查源文件夹是否存在\n if not os.path.exists(source_folder):\n raise Exception(\"Source folder does not exist.\")\n\n # 检查目标文件夹是否存在，如果不存在则创建\n if not os.path.exists(target_folder):\n os.makedirs(target_folder)\n\n # 遍历源文件夹中的所有文件和子文件夹\n for root, dirs, files in os.walk(source_folder):\n # 构建目标文件夹中的相对路径\n relative_path = os.path.relpath(root, source_folder)\n target_path = os.path.join(target_folder, relative_path)\n\n # 创建目标文件夹中的子文件夹\n if not os.path.exists(target_path):\n os.makedirs(target_path)\n\n # 复制源文件夹中的所有文件到目标文件夹中\n for file in files:\n source_file = os.path.join(root, file)\n target_file = os.path.join(target_path, file)\n shutil.copy2(source_file, target_file)\n\n print(\"Backup completed successfully.\")\n\n except Exception as e:\n print(f\"An error occurred: {e}\")\n\n```"} 18 | {"_id": 148, "response": "```python\nimport os\nimport shutil\n\ndef create_folder(folder_name):\n if not os.path.exists(folder_name):\n os.makedirs(folder_name)\n print(f'Folder {folder_name} created.')\n\ndef create_and_write_file(file_path, content):\n with open(file_path, 'w', encoding='utf-8') as file:\n file.write(content)\n print(f'File {file_path} created and written.')\n\ndef read_and_print_file(file_path):\n with open(file_path, 'r', encoding='utf-8') as file:\n content = file.read()\n print(f'Contents of {file_path}:')\n print(content)\n\ndef copy_file(source_path, destination_path):\n shutil.copy(source_path, destination_path)\n print(f'File {source_path} copied to {destination_path}.')\n\ndef move_file(source_path, destination_path):\n shutil.move(source_path, destination_path)\n print(f'File {source_path} moved to {destination_path}.')\n\ndef delete_folder(folder_name):\n shutil.rmtree(folder_name)\n print(f'Folder {folder_name} and its contents deleted.')\n\n```"} 19 | {"_id": 149, "response": "```python\nimport re\n\n\ndef extract_phone_numbers(text):\n # 电话号码的正则表达式模式\n pattern = re.compile(r'\\(?\\b(\\d{3})\\)?[-.\\s]?(\\d{3})[-.\\s]?(\\d{4})\\b')\n\n # 使用 finditer 方法提取所有匹配的电话号码\n matches = pattern.finditer(text)\n\n # 存储提取到的电话号码\n phone_numbers = []\n\n for match in matches:\n # 获取匹配对象的匹配组\n groups = match.groups()\n\n # 格式化提取到的电话号码\n formatted_number = f\"({groups[0]}) {groups[1]}-{groups[2]}\"\n\n # 添加到结果列表\n phone_numbers.append(formatted_number)\n\n return phone_numbers\n\n\n```"} 20 | {"_id": 150, "response": "```python\nimport numpy as np\nfrom scipy.linalg import eig\n\n\ndef calculate_eigen(matrix, num_largest=5):\n # 计算特征值和特征向量\n eigenvalues, eigenvectors = eig(matrix)\n\n # 找到最大的几个特征值及其对应的索引\n largest_indices = np.argsort(eigenvalues)[-num_largest:]\n\n # 提取最大的几个特征值和对应的特征向量\n largest_eigenvalues = eigenvalues[largest_indices]\n largest_eigenvectors = eigenvectors[:, largest_indices]\n\n return largest_eigenvalues, largest_eigenvectors\n\n```"} 21 | {"_id": 151, "response": "```python\ndef flatten_list(nested_list):\n flattened_list = []\n stack = [nested_list]\n\n while stack:\n current = stack.pop()\n\n for element in current:\n if isinstance(element, list):\n stack.append(element)\n else:\n flattened_list.append(element)\n\n return flattened_list\n\n\n```"} 22 | {"_id": 152, "response": "```python\nimport openpyxl\n# 打开txt文件\n\ndef txt2excle(file_path):\n with open(file_path, 'r') as file:\n lines = file.readlines()\n\n # 创建Excel工作簿\n workbook = openpyxl.Workbook()\n sheet = workbook.active\n\n # 将列表数据写入Excel工作表\n for row, line in enumerate(lines, start=1):\n # 假设列表中的元素使用逗号进行分隔\n elements = line.strip().split('\\t')\n for col, element in enumerate(elements, start=1):\n sheet.cell(row=row, column=col).value = element\n\n # 保存为Excel文件\n workbook.save('output.xlsx')\n\n```"} 23 | {"_id": 153, "response": "```python\nimport xml.etree.ElementTree as ET\n \ndef calculate_iou(box1, box2):\n \"\"\"\n 计算两个边界框之间的IoU\n \"\"\"\n x1, y1, x2, y2 = box1\n x3, y3, x4, y4 = box2\n\n # 计算相交矩形的左上角和右下角坐标\n inter_x1 = max(x1, x3)\n inter_y1 = max(y1, y3)\n inter_x2 = min(x2, x4)\n inter_y2 = min(y2, y4)\n\n # 计算相交矩形的面积\n inter_area = max(0, inter_x2 - inter_x1 + 1) * max(0, inter_y2 - inter_y1 + 1)\n\n # 计算并集面积和IoU\n box1_area = (x2 - x1 + 1) * (y2 - y1 + 1)\n box2_area = (x4 - x3 + 1) * (y4 - y3 + 1)\n union_area = box1_area + box2_area - inter_area\n iou = inter_area / union_area\n\n return iou\n\ndef convert_labels(label_file, target_region):\n \"\"\"\n 将在目标区域内与真实框标签的IoU大于0.2的所有框标签划分为类别0\n \"\"\"\n tree = ET.parse(label_file)\n root = tree.getroot()\n\n for object_elem in root.findall('object'):\n bbox = object_elem.find('bndbox')\n xmin = int(bbox.find('xmin').text)\n ymin = int(bbox.find('ymin').text)\n xmax = int(bbox.find('xmax').text)\n ymax = int(bbox.find('ymax').text)\n\n # 计算当前边界框与目标区域的IoU\n iou = calculate_iou(target_region, (xmin, ymin, xmax, ymax))\n\n # 如果IoU大于0.2，则将类别设置为0\n if iou > 0.2:\n object_elem.find('name').text = '0'\n\n tree.write(label_file)\n\n```"} 24 | {"_id": 154, "response": "```python\nimport os\nimport random\nimport shutil\n\n\ndef split_data(dataset_path,train_ratio,test_ratio,val_ratio):\n # 设置随机种子\n random_seed = 42\n random.seed(random_seed)\n train_path = os.path.join(dataset_path,'train')\n test_path = os.path.join(dataset_path,'test')\n val_path = os.path.join(dataset_path,'val')\n\n if not os.path.exists(train_path):\n os.makedirs(train_path)\n\n if not os.path.exists(test_path):\n os.makedirs(test_path)\n\n if not os.path.exists(val_path):\n os.makedirs(val_path)\n\n image_files = [f for f in os.listdir(dataset_path) if f.endswith('.jpg')]\n\n # 随机打乱图像文件列表\n random.shuffle(image_files)\n\n # 计算划分的索引位置\n train_split = int(len(image_files) * train_ratio)\n test_split = int(len(image_files) * (train_ratio + test_ratio))\n\n # 划分训练集\n train_files = image_files[:train_split]\n for file in train_files:\n src_path = os.path.join(dataset_path, file)\n dest_path = os.path.join(train_path, file)\n shutil.copy(src_path, dest_path)\n\n # 划分测试集\n test_files = image_files[train_split:test_split]\n for file in test_files:\n src_path = os.path.join(dataset_path, file)\n dest_path = os.path.join(test_path, file)\n shutil.copy(src_path, dest_path)\n\n # 划分验证集\n val_files = image_files[test_split:]\n for file in val_files:\n src_path = os.path.join(dataset_path, file)\n dest_path = os.path.join(val_path, file)\n shutil.copy(src_path, dest_path)\n\n\n# import os\n# import shutil\n# from sklearn.model_selection import train_test_split\n\n\n# def split_data(dataset_path, train_ratio, test_ratio, val_ratio):\n# # 获取所有文件的路径\n# files = [os.path.join(dataset_path, f) for f in os.listdir(dataset_path) if\n# os.path.isfile(os.path.join(dataset_path, f))]\n#\n# # 划分训练集和其余部分\n# train_files, rest_files = train_test_split(files, train_size=train_ratio, random_state=42)\n#\n# # 计算测试集的比例\n# test_ratio_adjusted = test_ratio / (1 - train_ratio)\n#\n# # 划分测试集和验证集\n# test_files, val_files = train_test_split(rest_files, test_size=test_ratio_adjusted, random_state=42)\n#\n# # 创建训练集、测试集和验证集的文件夹\n# train_dir = os.path.join(dataset_path, 'train')\n# test_dir = os.path.join(dataset_path, 'test')\n# val_dir = os.path.join(dataset_path, 'val')\n# os.makedirs(train_dir, exist_ok=True)\n# os.makedirs(test_dir, exist_ok=True)\n# os.makedirs(val_dir, exist_ok=True)\n#\n# # 将文件移动到对应的文件夹\n# for f in train_files:\n# shutil.move(f, os.path.join(train_dir, os.path.basename(f)))\n# for f in test_files:\n# shutil.move(f, os.path.join(test_dir, os.path.basename(f)))\n# for f in val_files:\n# shutil.move(f, os.path.join(val_dir, os.path.basename(f)))\n```"} 25 | {"_id": 155, "response": "```python\nimport numpy as np\n\ndef hinge_loss(y_true, y_pred):\n return np.mean(np.maximum(0, 1 - y_true * y_pred))\n\n```"} 26 | {"_id": 156, "response": "```python\nimport os\n\n\ndef remove_empty_lines_folder(folder_path):\n for file_name in os.listdir(folder_path):\n if file_name.endswith('.txt'):\n file_path = os.path.join(folder_path, file_name)\n remove_empty_lines(file_path)\n\n\ndef remove_empty_lines(file_path):\n with open(file_path, 'r+') as file:\n lines = file.readlines()\n file.seek(0)\n file.truncate()\n for line in lines:\n if line.strip(): # 检查是否为空行\n file.write(line)\n\n# folder_path=r'C:\\Users\\Administrator\\Desktop\\work\\1207-Q17\\yolo\\labels'\n# remove_empty_lines_folder(folder_path)\n\n```"} 27 | {"_id": 157, "response": "```python\nimport os\n\n\ndef label_filter(folder_path,Label_category):\n\n txt_files = [f for f in os.listdir(folder_path) if f.endswith('.txt')]\n for file_name in txt_files:\n file_path = os.path.join(folder_path, file_name)\n with open(file_path, 'r') as file:\n lines = file.readlines()\n\n new_lines = []\n\n\n for line in lines:\n parts = line.split()\n\n if int(parts[0]) in Label_category :\n continue\n new_line = ' '.join(parts) + '\\n'\n new_lines.append(new_line)\n\n # 将处理后的行写回文件\n with open(file_path, 'w') as file:\n file.writelines(new_lines)\n\n```"} 28 | {"_id": 158, "response": "```python\nimport os\n\n\ndef w_h_filter(folder_path,limit_wideth,limit_height):\n txt_files = [f for f in os.listdir(folder_path) if f.endswith('.txt')]\n for file_name in txt_files:\n file_path = os.path.join(folder_path, file_name)\n with open(file_path, 'r') as file:\n lines = file.readlines()\n\n new_lines = []\n\n\n for line in lines:\n parts = line.split()\n\n if (float(parts[3]) < limit_wideth or float(parts[4]) < limit_height):\n continue\n new_line = ' '.join(parts) + '\\n'\n new_lines.append(new_line)\n\n # 将处理后的行写回文件\n with open(file_path, 'w') as file:\n file.writelines(new_lines)\n\n\n\n\n# if __name__ == \"__main__\":\n#\n# folder_path = r\"C:\\Users\\Administrator\\Desktop\\work\\1207-Q14\\labels\"\n# limit_wideth=0.1\n# limit_height=0.02\n# w_h_filter(folder_path,limit_wideth,limit_height)\n\n```"} 29 | {"_id": 159, "response": "```python\nimport numpy as np\nimport torch\nimport torch.nn as nn\nimport torch.optim as optim\n\ndef mean_squared_error(y_true, y_pred):\n return np.mean(np.square(y_true - y_pred))\n```"} 30 | {"_id": 160, "response": "```python\nimport nltk\nnltk.data.path.append('nltk_package')\nfrom nltk.tokenize import word_tokenize\nfrom nltk.stem import PorterStemmer\nfrom nltk import pos_tag\n\n\ndef process_sentence(sentence):\n tokens = word_tokenize(sentence)\n pos_tags = pos_tag(tokens)\n # stop_words = set(stopwords.words('english'))\n # filtered_tokens = [word.lower() for word in tokens if word.lower() not in stop_words]\n filtered_tokens = [word.lower() for word in tokens]\n stemmer = PorterStemmer()\n stemmed_tokens = [stemmer.stem(word) for word in filtered_tokens]\n # 输出结果\n print(\"Original Sentence:\\n\", sentence)\n print(\"\\nTokenized Sentence:\\n\", tokens)\n print(\"\\nPOS Tagged Sentence:\\n\", pos_tags)\n print(\"\\nSentence without Stopwords:\\n\", filtered_tokens)\n print(\"\\nStemmed Sentence:\\n\", stemmed_tokens)\n return tokens, pos_tags, filtered_tokens, stemmed_tokens\n\nprint(process_sentence(\"!@#$%^&*()\"))\n\n```"} 31 | {"_id": 161, "response": "```python\nimport pandas as pd\nfrom sklearn.preprocessing import OneHotEncoder\n\n\ndef calculate_and_save_probabilities(input_file, output_file):\n # 读取数据集\n df = pd.read_csv(input_file)\n\n # 计算概率\n probabilities = {}\n for column_value in df['Sex'].unique():\n prob = df[(df['Sex'] == column_value) & (df[\"Survived\"] == 1)].shape[0] / df[df['Sex'] == column_value].shape[0]\n probabilities[f\"Probability({'Sex'}={column_value}, {'Survived'}=1)\"] = prob\n\n # 打印计算结果\n for key, value in probabilities.items():\n print(key + \":\", value)\n\n # 进行独热编码\n non_numeric_columns = df.select_dtypes(include=['object']).columns\n encoder = OneHotEncoder(sparse=False)\n encoded_features = pd.DataFrame(encoder.fit_transform(df[non_numeric_columns]))\n encoded_features.columns = encoder.get_feature_names_out(non_numeric_columns)\n\n # 替换原始数据集中的非数值列\n df.drop(non_numeric_columns, axis=1, inplace=True)\n df = pd.concat([df, encoded_features], axis=1)\n\n # 将数据集保存到新文件\n df.to_csv(output_file, index=False)\n\n return probabilities[f\"Probability(Sex=male, Survived=1)\"], probabilities[\"Probability(Sex=female, Survived=1)\"]\n\n\nprint(calculate_and_save_probabilities(\"test6/dataset.csv\", \"test6/test_label.csv\"))\n```"} 32 | {"_id": 162, "response": "```python\nimport numpy as np\n\n\ndef linear_fit(x, y):\n x = np.array(x)\n y = np.array(y)\n\n mean_x = np.mean(x)\n mean_y = np.mean(y)\n\n SS_xy = np.dot(x - mean_x, y - mean_y)\n SS_xx = np.dot(x - mean_x, x - mean_x)\n a = SS_xy / SS_xx\n b = mean_y - a * mean_x\n\n r = np.corrcoef(x, y)[0, 1]\n\n return a, b, r\n\n\n```"} 33 | {"_id": 163, "response": "```python\nimport os\n\ndef list_files(directory_path):\n # 检查目录是否存在\n if not os.path.exists(directory_path) or not os.path.isdir(directory_path):\n print(\"Invalid directory path.\")\n return\n\n list_files_recursive(directory_path)\n\ndef list_files_recursive(directory):\n # 获取目录中的所有文件和子目录\n files = os.listdir(directory)\n for file in files:\n file_path = os.path.join(directory, file)\n if os.path.isfile(file_path):\n # 如果是文件，则输出文件名\n print(file)\n elif os.path.isdir(file_path):\n # 如果是目录，则输出目录名，并递归调用列出子目录\n list_files_recursive(file_path)\n\n```"} 34 | {"_id": 164, "response": "```python\nimport os\n\ndef find_file(file_name, directory_path):\n exist = False\n # 使用os.walk遍历指定目录及其子目录中的文件\n for root, dirs, files in os.walk(directory_path):\n for file in files:\n if file == file_name:\n exist = True\n # 如果找到文件，输出文件的绝对路径\n file_path = os.path.join(root, file)\n print(file_path)\n \n if not exist:\n print(\"The file is not on disk\")\n\n```"} 35 | {"_id": 165, "response": "```python\nimport csv\n\ndef calculate_average(filename, column_index):\n with open(filename, 'r') as file:\n reader = csv.reader(file)\n next(reader) \n total, count = 0, 0\n for row in reader:\n if row[column_index].isdigit():\n total += int(row[column_index])\n count += 1\n return total / count if count != 0 else 0\n\n\n```"} 36 | {"_id": 166, "response": "```python\nimport os\nimport pandas as pd\n\n\ndef find_max_temperature(folder_path, output_file):\n csv_files = [f for f in os.listdir(folder_path) if f.endswith('.csv')]\n \n result_df = pd.DataFrame(columns=['City', 'Max_Temperature', 'Date'])\n \n for csv_file in csv_files:\n file_path = os.path.join(folder_path, csv_file)\n df = pd.read_csv(file_path)\n \n city_name = csv_file[:-4]\n \n # 找到最高气温的值\n max_temp = df['Temperature'].max()\n \n # 找到所有最高气温的日期\n max_temp_dates = df.loc[df['Temperature'] == max_temp, 'Date'].tolist()\n \n # 将结果添加到result_df\n for date in max_temp_dates:\n result_df = result_df._append({'City': city_name,\n 'Max_Temperature': max_temp,\n 'Date': date}, ignore_index=True)\n \n result_df.to_csv(output_file, index=False)\n\n```"} 37 | {"_id": 167, "response": "```python\nimport pandas as pd\nimport torch\n\ndef merge_and_smooth_data(filenames, window_size):\n dfs = [pd.read_csv(filename) for filename in filenames]\n merged_df = pd.concat(dfs).sort_values(by='timestamp')\n\n tensor = torch.tensor(merged_df.drop(columns=['timestamp']).values)\n\n cumsum = tensor.cumsum(dim=0)\n sliding_avg = (cumsum[window_size:] - cumsum[:-window_size]) / float(window_size)\n return sliding_avg\n\n\n```"} 38 | {"_id": 168, "response": "```python\nfrom PIL import Image\nimport os\n\n\ndef resize_and_rename_images(input_folder, output_folder, target_resolution=(800, 600)):\n if not os.path.exists(output_folder):\n os.makedirs(output_folder)\n image_files = [f for f in os.listdir(input_folder) if f.endswith(('.png', '.jpg', '.jpeg'))]\n image_files.sort() # 按文件名排序\n for i, image_file in enumerate(image_files):\n input_path = os.path.join(input_folder, image_file)\n output_path = os.path.join(output_folder, f\"{i + 1:03d}.jpg\") # 格式化文件名\n\n # 打开图片\n img = Image.open(input_path)\n\n # 转换为RGB模式\n img = img.convert(\"RGB\")\n\n # 统一分辨率\n img = img.resize(target_resolution, Image.LANCZOS)\n\n # 保存图片\n img.save(output_path)\n\n print(\"Modified images resolutions:\")\n print_resolutions(output_folder)\n return output_folder\n\n\ndef print_resolutions(folder_path):\n image_files = [f for f in os.listdir(folder_path) if f.endswith(('.png', '.jpg', '.jpeg'))]\n image_files.sort()\n\n for image_file in image_files:\n image_path = os.path.join(folder_path, image_file)\n img = Image.open(image_path)\n resolution = img.size\n print(f\"{image_file}: {resolution[0]} x {resolution[1]}\")\n\n\nresize_and_rename_images('test6/input', 'test6/output', (2560, 1440))\n```"} 39 | {"_id": 169, "response": "```python\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\n\nclass FocalLoss(nn.Module):\n def __init__(self, alpha=0.25, gamma=2.0):\n super(FocalLoss, self).__init__()\n self.alpha = alpha\n self.gamma = gamma\n\n def forward(self, inputs, targets):\n \n BCE_loss = F.binary_cross_entropy_with_logits(inputs, targets, reduction='none')\n pt = torch.exp(-BCE_loss) \n focal_loss = self.alpha * (1-pt)**self.gamma * BCE_loss\n return focal_loss.mean()\n\n\ncriterion = FocalLoss(alpha=0.25, gamma=2.0)\ninputs = torch.tensor([0.1, 0.2, 0.4, 0.8, 1.2, -0.1, -0.2, -0.4, -0.8, -1.2], requires_grad=True)\ntargets = torch.tensor([1., 0., 1., 0., 1., 0., 1., 0., 1., 0.])\n\nloss = criterion(inputs, targets).detach()\nprint(loss)\n\n\n```"} 40 | {"_id": 170, "response": "```python\nimport os\nimport shutil\n\n\ndef copy_and_move_file(source_path, destination_directory):\n # 获取源文件的文件名\n file_name = os.path.basename(source_path)\n\n # 构建目标路径\n destination_path = os.path.join(destination_directory, file_name)\n\n # 检查源文件是否存在\n if os.path.exists(source_path):\n # 复制文件\n shutil.copy(source_path, destination_path)\n else:\n print(\"File not exist.\")\n\n```"} 41 | {"_id": 171, "response": "```python\nfrom docx import Document\n\n\ndef count_keywords_in_docx(docx_path, keywords):\n doc = Document(docx_path)\n text = \" \".join([paragraph.text for paragraph in doc.paragraphs])\n\n keyword_counts = {}\n for keyword in keywords:\n count = text.lower().count(keyword.lower())\n keyword_counts[keyword] = count\n\n return keyword_counts\n\n```"} 42 | {"_id": 172, "response": "```python\nimport shutil\nimport os\n\ndef delete_folder(folder_path):\n try:\n if os.path.exists(folder_path):\n if os.path.isdir(folder_path):\n shutil.rmtree(folder_path)\n print(f\"Successfully delete the folder.\")\n else:\n print(f\"{folder_path} is not a directory\")\n else:\n print(f\"Path not exist: {folder_path}\")\n except Exception as e:\n print(f\"Can't delete the folder.\")\n\n```"} 43 | {"_id": 173, "response": "```python\nimport numpy as np\nimport scipy.sparse as sp\n\n\ndef csr_to_numpy(input_file, output_file):\n input_arr_csr = sp.load_npz(input_file)\n input_arr = input_arr_csr.toarray()\n np.save(output_file, input_arr)\n\n```"} 44 | {"_id": 174, "response": "```python\nimport shutil\nimport os\nimport threading\nimport time\n\n\ndef copy_folder(source_folder, dest_folders):\n for dest_folder in dest_folders:\n try:\n if os.path.exists(dest_folder):\n shutil.rmtree(dest_folder)\n shutil.copytree(source_folder, dest_folder)\n print(\"success from {source_folder} to {dest_folder}\")\n except Exception as e:\n print(\"copy {source_folder} to {dest_folder} is err:{e}\")\ndef create_thread(source_folder, dest_folders):\n # 计算每个线程需要拷贝的目标文件夹数量\n threads_num = len(dest_folders)\n folders_per_thread = len(dest_folders) // threads_num\n\n threads = []\n for i in range(threads_num):\n start = i * folders_per_thread\n end = start + folders_per_thread\n if i == threads_num - 1:\n end = len(dest_folders)\n thread = threading.Thread(target=copy_folder, args=(source_folder, dest_folders[start:end]))\n threads.append(thread)\n \n return threads\n \ndef run(threads):\n # 记录开始时间\n start_time = time.time()\n # 启动所有线程\n for thread in threads:\n thread.start()\n # 等待所有线程结束\n for thread in threads:\n thread.join()\n # 记录结束时间并输出用时\n end_time = time.time()\n print(f\"time: {end_time - start_time} 秒\")\n \n \ndef multithread_copy(source_folder, dest_folders):\n threads = create_thread(source_folder, dest_folders)\n run(threads)\n\n```"} 45 | {"_id": 175, "response": "```python\nimport numpy as np\nimport scipy.sparse as sp\n\ndef numpy_to_csr(input_file, output_file):\n input_arr = np.load(input_file)\n input_arr_csr = sp.csr_matrix(input_arr)\n sp.save_npz(output_file, input_arr_csr)\n```"} 46 | {"_id": 176, "response": "```python\nimport numpy as np\nfrom sklearn.metrics.pairwise import cosine_similarity\n\n\ndef calculate_similarity(file_A, file_B, target_file):\n A = np.load(file_A)\n B = np.load(file_B)\n cos_sim = cosine_similarity(A, B)\n np.save(target_file, cos_sim)\n\n\n```"} 47 | {"_id": 177, "response": "```python\nimport torch\nimport numpy as np\n\n\ndef process_tensor(array1, array2):\n tensor1 = torch.FloatTensor(array1)\n tensor2 = torch.FloatTensor(array2)\n tensor2 = tensor2.permute(0, 2, 1)\n attention_scores = torch.bmm(tensor1, tensor2)\n with open('result.txt', 'w') as file:\n file.write(str(attention_scores))\n return attention_scores\n\n\narray1 = np.ones((100, 130, 80))\narray2 = np.ones((100, 130, 80))\nprint(process_tensor(array1, array2).shape)\n\n\n```"} 48 | {"_id": 178, "response": "```python\nimport math\n\ndef euclidean_distance(point1, point2):\n return math.sqrt(sum((p - q) ** 2 for p, q in zip(point1, point2)))\n\ndef manhattan_distance(point1, point2):\n return sum(abs(p - q) for p, q in zip(point1, point2))\n\ndef chebyshev_distance(point1, point2):\n return max(abs(p - q) for p, q in zip(point1, point2))\n\ndef minkowski_distance(point1, point2, p):\n return math.pow(sum(abs(p - q) ** p for p, q in zip(point1, point2)), 1/p)\n\ndef cosine_similarity(vector1, vector2):\n dot_product = sum(p * q for p, q in zip(vector1, vector2))\n norm_vector1 = math.sqrt(sum(p ** 2 for p in vector1))\n norm_vector2 = math.sqrt(sum(q ** 2 for q in vector2))\n return dot_product / (norm_vector1 * norm_vector2)\n\n\ndef hamming_distance(string1, string2):\n return sum(p != q for p, q in zip(string1, string2))\n\ndef jaccard_similarity(set1, set2):\n intersection_size = len(set1.intersection(set2))\n union_size = len(set1.union(set2))\n return intersection_size / union_size\n\n\n\n\n```"} 49 | {"_id": 179, "response": "```python\nimport numpy as np\n\ndef diag_zero(matrix):\n matrix[np.triu_indices(matrix.shape[0], k=1)] = 0\n sum_remaining_values = np.sum(matrix)\n return sum_remaining_values\n\nmatrix = np.arange(1, 10).reshape(3, 3)\nprint(diag_zero(matrix))\n\n\n```"} 50 | {"_id": 180, "response": "```python\nimport numpy as np\n\n\ndef array_avg(matrix):\n # 计算数组元素的平均值\n mean_value = np.mean(matrix)\n\n # 大于平均值的元素置为零\n matrix[matrix > mean_value] = 0\n\n # 将数组展开为一维列表\n flattened_array = matrix.flatten()\n return mean_value, flattened_array\n\nmatrix = [[2,1,4,3],[4,5,8,9],[10,2,3,4],[1,14,21,2]]\nmatrix = np.array(matrix)\nprint(array_avg(matrix))\n\n```"} 51 | {"_id": 181, "response": "```python\nimport os\n\ndef count_files(dir):\n count = 0\n for root, dirs, files in os.walk(dir):\n count += len(files)\n return count\n\n```"} 52 | {"_id": 182, "response": "```python\nimport pandas as pd\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn.decomposition import PCA\n\ndef pac_fit(input_file):\n characters = pd.read_csv(input_file)\n characters = StandardScaler().fit_transform(characters)\n pca = PCA().fit(characters)\n return pca.explained_variance_ratio_\n\n\ndf = pd.DataFrame({'A': [1, 2], 'B': [2, 3], 'C': [3, 4]})\ndf.to_csv('test.csv', index=False)\nprint(pac_fit('test.csv'))\n```"} 53 | {"_id": 183, "response": "```python\nimport os\nimport shutil\nimport zipfile\nfrom datetime import datetime\nimport glob\n\ndef backup(source_path, backup_folder, max_size_mb=10):\n # 创建备份文件夹\n if not os.path.exists(backup_folder):\n os.makedirs(backup_folder)\n\n # 生成时间戳\n timestamp = datetime.now().strftime('%Y%m%d%H%M%S')\n\n # 构建备份文件名\n backup_filename = f\"backup_{timestamp}.zip\"\n backup_path = os.path.join(backup_folder, backup_filename)\n\n # 备份文件夹\n shutil.make_archive(backup_path[:-4], 'zip', source_path)\n\n # 获取备份文件夹总大小\n total_size_mb = sum(os.path.getsize(f) for f in glob.glob(os.path.join(backup_folder, '*.zip'))) / (1024 ** 2)\n\n # 如果总大小超过设定的阈值，删除最旧的备份文件\n while total_size_mb > max_size_mb:\n oldest_backup = min(glob.glob(os.path.join(backup_folder, '*.zip')), key=os.path.getctime)\n total_size_mb -= os.path.getsize(oldest_backup) / (1024 ** 2)\n os.remove(oldest_backup)\n\n print(f\"Backup completed: {backup_path}\")\n```"} 54 | {"_id": 184, "response": "```python\nimport torch\n\ndef bbox_iou(boxes1, boxes2):\n \"\"\"\n 计算两组对应位置bounding boxes的IoU。\n \n 参数:\n boxes1, boxes2: 维度为[N, 4]的tensor，表示bounding boxes，格式为[x, y, w, h]。\n \n 返回:\n 一个长度为N的tensor，表示boxes1和boxes2中对应box的IoU。\n \"\"\"\n assert boxes1.shape == boxes2.shape\n boxes1_x1 = boxes1[:, 0] - boxes1[:, 2] / 2\n boxes1_y1 = boxes1[:, 1] - boxes1[:, 3] / 2\n boxes1_x2 = boxes1[:, 0] + boxes1[:, 2] / 2\n boxes1_y2 = boxes1[:, 1] + boxes1[:, 3] / 2\n\n boxes2_x1 = boxes2[:, 0] - boxes2[:, 2] / 2\n boxes2_y1 = boxes2[:, 1] - boxes2[:, 3] / 2\n boxes2_x2 = boxes2[:, 0] + boxes2[:, 2] / 2\n boxes2_y2 = boxes2[:, 1] + boxes2[:, 3] / 2\n\n\n inter_x1 = torch.max(boxes1_x1, boxes2_x1)\n inter_y1 = torch.max(boxes1_y1, boxes2_y1)\n inter_x2 = torch.min(boxes1_x2, boxes2_x2)\n inter_y2 = torch.min(boxes1_y2, boxes2_y2)\n\n inter_area = torch.clamp(inter_x2 - inter_x1, min=0) * torch.clamp(inter_y2 - inter_y1, min=0)\n\n\n boxes1_area = (boxes1_x2 - boxes1_x1) * (boxes1_y2 - boxes1_y1)\n boxes2_area = (boxes2_x2 - boxes2_x1) * (boxes2_y2 - boxes2_y1)\n\n\n union_area = boxes1_area + boxes2_area - inter_area\n\n\n iou = inter_area / union_area\n\n return iou\n\n\n\n\n```"} 55 | {"_id": 185, "response": "```python\nimport re\nfrom collections import defaultdict\nimport json\n\ndef nginx_log_analysis(nginx_log_file, statistics_output_file, high_frequency_output_file):\n # 定义一个字典用于存储IP地址访问次数\n ip_count = defaultdict(int)\n\n # 定义一个集合用于存储高频率访问的IP地址\n high_frequency_ips = defaultdict(int)\n \n # 定义正则表达式来匹配IP地址\n ip_pattern = re.compile(r'\\b(?:[0-9]{1,3}\\.){3}[0-9]{1,3}\\b')\n \n # 打开Nginx日志文件并逐行处理\n with open(nginx_log_file, 'r') as file:\n for line in file:\n # 使用正则表达式匹配IP地址\n match = ip_pattern.search(line)\n if match:\n ip_address = match.group(0)\n \n # 增加IP地址访问次数\n ip_count[ip_address] += 1\n \n # 判断是否是高频率访问\n if ip_count[ip_address] > 30:\n high_frequency_ips[ip_address] = ip_count[ip_address]\n\n print(ip_count)\n # 输出IP地址及其访问次数到统计文件\n with open(statistics_output_file, 'w') as output:\n json.dump(ip_count, output, ensure_ascii=False, indent=2)\n \n # 输出高频率访问的IP地址到单独的文件\n with open(high_frequency_output_file, 'w') as output:\n json.dump(high_frequency_ips, output, ensure_ascii=False, indent=2)\n\n```"} 56 | {"_id": 186, "response": "```python\ndef lengthLongestPath(input: str) -> int:\n st = []\n ans, i, n = 0, 0, len(input)\n while i < n:\n # 检测当前文件的深度\n depth = 1\n while i < n and input[i] == '\\t':\n depth += 1\n i += 1\n # 统计当前文件名的长度\n length, isFile = 0, False\n while i < n and input[i] != '\\n':\n if input[i] == '.':\n isFile = True\n length += 1\n i += 1\n i += 1 # 跳过换行符\n while len(st) >= depth:\n st.pop()\n if st:\n length += st[-1] + 1\n if isFile:\n ans = max(ans, length)\n else:\n st.append(length)\n return ans\n\n\n```"} 57 | {"_id": 187, "response": "```python\nfrom sklearn.cluster import KMeans\nimport numpy as np\n\ndef k_means_clustering(data, num_clusters):\n kmeans = KMeans(n_clusters=num_clusters, random_state=42)\n kmeans.fit(data)\n return kmeans.labels_\n\n```"} 58 | {"_id": 188, "response": "```python\nimport cv2\nimport numpy as np\n\ndef gamma_correction(image, gamma):\n # 将图像的亮度值转换为范围在[0, 1]之间的浮点数\n image = image / 255.0\n \n # 对图像进行伽马亮度矫正\n corrected_image = np.power(image, gamma)\n \n # 将亮度值重新缩放到范围[0, 255]\n corrected_image = np.uint8(corrected_image * 255)\n \n return corrected_image\n```"} 59 | {"_id": 189, "response": "```python\n\nimport os\nfrom PyPDF2 import PdfMerger\n\n# 定义一个函数来合并PDF文件\ndef merge_pdfs(root_dir):\n def merge(files, output):\n merger = PdfMerger()\n for pdf in files:\n merger.append(pdf)\n merger.write(output)\n merger.close()\n\n # 遍历文件夹\n for root, dirs, files in os.walk(root_dir):\n pdf_files = [f for f in files if f.endswith(\".pdf\")]\n if pdf_files: # 如果存在PDF文件\n output = os.path.join(root, \"merged.pdf\") # 输出文件名\n full_pdf_files = [os.path.join(root, f) for f in pdf_files] # 获取完整路径\n full_pdf_files = sorted(full_pdf_files)\n merge(full_pdf_files, output) # 合并PDF文件\n\n\n\n\n\n\n\n```"} 60 | {"_id": 190, "response": "```python\n\nimport pandas as pd\nfrom io import StringIO\n\ndef calculate_date_difference(csv_contents):\n # Read the CSV contents into a pandas DataFrame\n df = pd.read_csv(StringIO(csv_contents))\n\n # Convert the 'Date' column to datetime\n df['Date'] = pd.to_datetime(df['Date'])\n\n # Calculate the difference in days between the earliest and latest date\n date_difference = (df['Date'].max() - df['Date'].min()).days\n\n return date_difference\n\n\n```"} 61 | {"_id": 191, "response": "```python\n\nimport os\nimport shutil\n\ndef copy_files(file_path, dest_path, pairs_per_folder=10):\n files = os.listdir(file_path)\n files.sort()\n pairs = zip(files[::2], files[1::2])\n for i, pair in enumerate(pairs):\n sub_folder = os.path.join(dest_path, 'sub{}'.format(i // pairs_per_folder))\n os.makedirs(sub_folder, exist_ok=True)\n for file in pair:\n shutil.copy(os.path.join(file_path, file), sub_folder)\n\n# 使用方法\n# copy_files('/path/to/source/folder', '/path/to/destination/folder')\n\n\n```"} 62 | {"_id": 192, "response": "```python\n\nimport re\n\ndef replace_code_block(text):\n # 定义一个字典，将代码块标记映射到相应的HTML类\n languages = {\n \"html\": \"language-html\",\n \"css\": \"language-css\",\n \"python\": \"language-python\",\n \"javascript\": \"language-javascript\",\n \"golang\": \"language-golang\"\n }\n\n # 对于每种语言，使用正则表达式查找和替换相应的代码块\n for lang, html_class in languages.items():\n # 构建正则表达式模式\n pattern = rf\"```{lang}\\b\\s*(.*?)\\s*```\"\n # 替换文本\n replacement = rf'
\\1
'\n # 应用替换\n text = re.sub(pattern, replacement, text, flags=re.DOTALL)\n\n return text\n\n\n\n```"} 63 | {"_id": 193, "response": "```python\n\ndef calculate_et(delta, rn, g, gamma, t, u2, es, ea):\n numerator = 0.408 * delta * (rn - g) + gamma * (900 / (t + 273)) * u2 * (es - ea)\n denominator = delta + gamma * (1 + 0.34 * u2)\n if denominator == 0:\n return 0\n et = numerator / denominator\n\n print(et)\n return et\n\n\n```"} 64 | {"_id": 194, "response": "```python\n\nfrom lxml import etree\n\nhtml = \"\"\"\n

\n 3131\n AAAA\n BBB\n CCC\n

\n\"\"\"\n\nroot = etree.fromstring(html)\n\ndef has_strong_or_em(tag):\n if tag.find('strong') is not None or tag.find('em') is not None:\n return True\n return False\n\nfor p in root.iter('p'):\n if has_strong_or_em(p):\n print(\"P tag has strong or em tag\")\n else:\n print(\"P tag does not have strong or em tag\")\n\n\n```"} 65 | {"_id": 195, "response": "```python\n\nimport pandas as pd\n\ndef flatten_dict_to_rows(d):\n def flatten_dict(d, parent_key='', sep='.'):\n items = []\n for k, v in d.items():\n new_key = f\"{parent_key}{sep}{k}\" if parent_key else k\n if isinstance(v, dict):\n if not v:\n items.append((new_key, {}))\n else:\n items.extend(flatten_dict(v, new_key, sep=sep).items())\n else:\n items.append((new_key, v))\n return dict(items)\n\n flat_dict = flatten_dict(d)\n\n # Convert the flattened dictionary to a pandas DataFrame\n df = pd.DataFrame(list(flat_dict.items()), columns=['Key', 'Value'])\n print(df.to_dict())\n return df\n\n\n```"} 66 | {"_id": 196, "response": "```python\n\ndef is_mac_in_range(mac, mac_range):\n def mac_to_hex(mac_str):\n return mac_str.replace(':', '').replace('-', '').upper()\n\n # 解析MAC地址范围\n start_mac_str, end_mac_str = mac_range.split('-')\n start_mac = int(mac_to_hex(start_mac_str), 16)\n end_mac = int(mac_to_hex(end_mac_str), 16)\n mac_val = int(mac_to_hex(mac), 16)\n\n # 判断MAC地址是否在范围内\n return start_mac <= mac_val <= end_mac\n\n\n```"} 67 | {"_id": 197, "response": "```python\n\nimport numpy as np\nfrom scipy.optimize import curve_fit\n\n# Define the fractional order model with 2 constant phase elements and 3 resistors\ndef fractional_order_model(t, Q, R1, R2, R3, n1, n2):\n return Q * (1 - (1 / (R1 * (t ** n1) + 1)) - (1 / (R2 * (t ** n2) + 1)) - (1 / R3))\n\n# # Sample time and current data\n# time_data = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10])\n# current_data = np.array([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0])\n#\n# # Fit the model to the data\n# popt, pcov = curve_fit(fractional_order_model, time_data, current_data, bounds=(0, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0]))\n# print(len(popt))\n# # The estimated battery capacity in Amp hour is the first parameter of the model\n# estimated_capacity_Ah = popt[0]\n#\n# print(\"Estimated battery capacity: {:.2f} Ah\".format(estimated_capacity_Ah))\n\n\n```"} 68 | {"_id": 198, "response": "```python\n\nimport numpy as np\n\ndef solve_quadratic(points):\n # Create matrix A using a list comprehension\n A = np.array([[x**2, x, 1] for x, y in points])\n # Create vector b\n b = np.array([y for x, y in points])\n # Use numpy.linalg.solve to solve for [a, b, c]\n coefficients = np.linalg.solve(A, b)\n\n return coefficients\n\n```"} 69 | {"_id": 199, "response": "```python\n\nimport ipaddress\n\ndef check_ip_overlap(ip_range1, ip_range2):\n ip_net1 = ipaddress.ip_network(ip_range1)\n ip_net2 = ipaddress.ip_network(ip_range2)\n\n return ip_net1.overlaps(ip_net2)\n\n\n\n```"} 70 | {"_id": 200, "response": "```python\n\nimport os\nimport re\n\ndef rename_files_in_folder(folder_path):\n for filename in os.listdir(folder_path):\n file_type = filename.split('.')[-1]\n new_filename = re.sub(\"[A-Za-z]\", \"\", filename[:-len(file_type)]) + file_type\n os.rename(os.path.join(folder_path, filename), os.path.join(folder_path, new_filename))\n\n# Call the function with the path to the folder\n# rename_files_in_folder('/path/to/your/folder')\n\n\n```"} 71 | -------------------------------------------------------------------------------- /results/reference/reference_ncb_python_zh.jsonl: -------------------------------------------------------------------------------- 1 | {"_id": 131, "response": "```python\nimport string\nfrom collections import Counter\n\n\ndef word_count(file_path):\n try:\n # 读取文件内容\n with open(file_path, 'r', encoding='utf-8') as file:\n content = file.read()\n\n # 移除标点符号并转换为小写\n translator = str.maketrans(\"\", \"\", string.punctuation)\n content = content.translate(translator).lower()\n\n # 使用 Counter 统计单词出现次数\n words = content.split()\n word_counter = Counter(words)\n\n # 按照出现次数降序排列\n sorted_word_count = sorted(word_counter.items(), key=lambda x: x[1], reverse=True)\n\n for word, count in sorted_word_count:\n print(f\"'{word}': {count}\")\n\n except FileNotFoundError:\n print(f\"Error: File '{file_path}' not found.\")\n```"} 2 | {"_id": 132, "response": "```python\nimport re\n\n\ndef remove_html_tags(s: str) -> str:\n if not isinstance(s, str):\n raise ValueError(\"Input must be a string\")\n\n if not s:\n return \"\"\n\n # 定义新的正则表达式模式\n tag_pattern = r\"<[^>]*>\"\n comment_pattern = r\"\"\n doctype_pattern = r\"]*>\"\n\n # 移除 HTML 注释和 DOCTYPE\n s = re.sub(comment_pattern, \"\", s)\n s = re.sub(doctype_pattern, \"\", s)\n\n # 移除标签\n s = re.sub(tag_pattern, \"\", s)\n\n # 返回处理后的字符串\n return s.strip()\n\n\n\n```"} 3 | {"_id": 133, "response": "```python\nfrom datetime import datetime\n\nclass Schedule:\n def __init__(self):\n self.events = {}\n\n def add_event(self, event_name, event_date):\n if event_date in self.events:\n self.events[event_date].append(event_name)\n else:\n self.events[event_date] = [event_name]\n\n def view_events(self):\n for event_date, event_names in self.events.items():\n print(f\"{event_date.strftime('%Y-%m-%d %H:%M')}:\")\n for event_name in event_names:\n print(f\"- {event_name}\")\n print()\n\n def remove_event(self, event_name, event_date):\n if event_date in self.events:\n if event_name in self.events[event_date]:\n self.events[event_date].remove(event_name)\n if not self.events[event_date]:\n del self.events[event_date]\n print(f\"已删除事件：{event_name}\")\n return\n print(f\"未找到事件'{event_name}'\")\n\n\n```"} 4 | {"_id": 134, "response": "```python\nclass ListNode:\n def __init__(self, value=0, next=None):\n self.value = value\n self.next = next\n\n\ndef add_two_numbers(l1, l2):\n carry = 0\n dummy_head = ListNode()\n current = dummy_head\n\n while l1 or l2 or carry:\n # 获取当前节点的值\n val1 = l1.value if l1 else 0\n val2 = l2.value if l2 else 0\n\n # 计算和以及进位\n total = val1 + val2 + carry\n carry, digit = divmod(total, 10)\n\n # 创建新节点并更新当前节点的next\n current.next = ListNode(digit)\n current = current.next\n\n # 移动到下一个节点\n if l1:\n l1 = l1.next\n if l2:\n l2 = l2.next\n\n return dummy_head.next\n```"} 5 | {"_id": 135, "response": "```python\nfrom typing import List\nimport xml.etree.ElementTree as ET\n\ndef parse_xml(file_path: str, tag_name: str) -> List[str]:\n\n try:\n root = ET.parse(file_path).getroot()\n except ET.ParseError:\n return []\n data_list = []\n for element in root.findall(tag_name):\n data_list.append(element.text)\n return data_list\n\n```"} 6 | {"_id": 136, "response": "```python\nimport re\n\n\ndef is_valid_variable_name(name):\n \"\"\"\n 判断字符串是否是有效的 JavaScript 变量名\n \"\"\"\n if len(name) > 255:\n return False\n if not re.match(r'^[a-zA-Z_$][a-zA-Z0-9_$]*$', name):\n return False\n return True\n\n```"} 7 | {"_id": 137, "response": "```python\nclass Order:\n def __init__(self, order_id):\n self.order_id = order_id\n self.items = []\n self.total_price = 0.0\n self.status = \"未支付\"\n\n def add_item(self, item_name, item_price):\n if not isinstance(item_price, (int, float)) or item_price <= 0:\n return \"商品价格必须是正数\"\n\n item = {\"name\": item_name, \"price\": item_price}\n self.items.append(item)\n\n def remove_item(self, item_name):\n for item in self.items:\n if item[\"name\"] == item_name:\n self.items.remove(item)\n break\n else:\n return \"该商品不存在于订单中\"\n\n def calculate_total_price(self):\n self.total_price = sum(item[\"price\"] for item in self.items)\n\n def pay_order(self):\n if not self.items:\n return \"订单中没有商品\"\n self.status = \"已支付\"\n\n def cancel_order(self):\n self.status = \"已取消\"\n\n def view_order(self):\n print(\"订单编号：\", self.order_id)\n print(\"商品列表：\")\n for item in self.items:\n print(\"- {}: ￥{}\".format(item[\"name\"], item[\"price\"]))\n print(\"总价格：￥{}\".format(self.total_price))\n print(\"状态：\", self.status)\n```"} 8 | {"_id": 138, "response": "```python\ndef is_valid_sequence(s):\n # 判断字符串长度是否在 2 到 10 之间\n if len(s) < 2 or len(s) > 10:\n return False\n\n # 判断字符串中的字符是否都是小写字母\n for char in s:\n if not char.islower():\n return False\n\n # 判断字符串中是否出现连续的重复字符\n for i in range(len(s) - 1):\n if s[i] == s[i + 1]:\n return False\n\n # 判断左右两部分是否都是回文字符串或递增序列\n left, right = s[:len(s) // 2], s[len(s) // 2:]\n if left == right[::-1] or sorted(left) == sorted(right):\n return True\n else:\n return False\n```"} 9 | {"_id": 139, "response": "```python\nimport math\n\nclass RouteSimulator:\n def __init__(self, x, y):\n self.x = x\n self.y = y\n self.move_history = []\n\n def move(self, distance, angle_degrees):\n # 将角度转换为弧度\n angle_radians = math.radians(angle_degrees)\n\n # 计算新的位置\n new_x = self.x + distance * math.cos(angle_radians)\n new_y = self.y + distance * math.sin(angle_radians)\n\n # 更新位置\n self.x, self.y = new_x, new_y\n\n # 记录移动历史\n self.move_history.append({'distance': distance, 'angle_degrees': angle_degrees, 'position': (new_x, new_y)})\n\n def calculate_distance(self, x1, y1, x2, y2):\n # 计算两点之间的距离\n return math.sqrt((x2 - x1)**2 + (y2 - y1)**2)\n\n def print_current_position(self):\n print(f\"当前位置：({self.x:.2f}, {self.y:.2f})\")\n\n def print_move_history(self):\n print(\"移动历史：\")\n for move in self.move_history:\n distance = move['distance']\n angle_degrees = move['angle_degrees']\n position = move['position']\n print(f\"距离: {distance:.2f}, 角度: {angle_degrees:.2f}, 位置: ({position[0]:.2f}, {position[1]:.2f})\")\n\n```"} 10 | {"_id": 140, "response": "```python\ndef utf8_to_ascii(utf8_string):\n # 验证输入字符串是否为空\n if not utf8_string:\n print(\"输入字符串不能为空！\")\n return None\n\n result = []\n\n try:\n # 遍历输入字符串的每个字符\n for i in range(len(utf8_string)):\n if utf8_string[i] == '%' and i + 2 < len(utf8_string) and utf8_string[i + 1:i + 3].isalnum():\n # 如果遇到以%开头的16进制字符编码，则进行解析\n hex_str = utf8_string[i + 1:i + 3]\n ascii_code = int(hex_str, 16)\n result.append(ascii_code)\n i += 2\n else:\n # 普通的UTF-8字符直接转换为ASCII码\n ascii_code = ord(utf8_string[i])\n result.append(ascii_code)\n except IndexError:\n print(\"输入字符串格式错误！\")\n return None\n except ValueError:\n print(\"输入字符串包含非法的字符编码！\")\n return None\n except Exception as e:\n print(\"发生未知错误：\", str(e))\n return None\n\n return result\n\n\n\n```"} 11 | {"_id": 141, "response": "```python\nimport re\nimport datetime\n\n\ndef find_dates(text):\n if not text:\n return \"输入的文本为空，请提供有效的文本内容\"\n\n date_patterns = [\n r\"\\b(\\d{4})-(0[1-9]|1[0-2])-(0[1-9]|[12][0-9]|3[01])\\b\", # YYYY-MM-DD\n r\"\\b(0[1-9]|1[0-2])/(0[1-9]|[12][0-9]|3[01])/(19|20)?\\d{2}\\b\", # MM/DD/YYYY or M/D/YYYY\n r\"\\b(0[1-9]|[12][0-9]|3[01])-(0[1-9]|1[0-2])-(19|20)?\\d{2}\\b\", # DD-MM-YYYY or D-M-YYYY\n r\"\\b(0[1-9]|[12][0-9]|3[01])\\s(January|February|March|April|May|June|July|August|September|October|November|December),?\\s(19|20)?\\d{2}\\b\"\n # DD Month YYYY or D Month YYYY\n ]\n dates = []\n for pattern in date_patterns:\n dates.extend(re.findall(pattern, text))\n\n formatted_dates = []\n for date in dates:\n if date[0]: # 如果匹配到了年份\n year = date[0]\n else:\n year = \"20\" + date[3] if date[3] else \"\" # 如果没有匹配到年份，则根据模式中的分组确定年份\n\n month = date[1].zfill(2)\n day = date[2].zfill(2)\n\n # 检查日期的有效性\n try:\n formatted_date = f\"{year}-{month}-{day}\"\n _ = datetime.datetime.strptime(formatted_date, \"%Y-%m-%d\")\n formatted_dates.append(formatted_date)\n except ValueError:\n pass\n\n if not formatted_dates:\n return \"未找到有效日期\"\n\n return formatted_dates\n\n```"} 12 | {"_id": 142, "response": "```python\nclass Process:\n def __init__(self, pid, code):\n self.pid = pid\n self.code = code\n self.memory = None\n\ndef allocate_memory(memory_map, process):\n if process.memory is not None:\n raise ValueError(\"Process already has memory\")\n\n for i in range(len(memory_map)):\n if memory_map[i] == 0:\n process.memory = i\n for j in range(len(process.code)):\n memory_map[i + j] = 1\n return True\n\n return False\n\ndef deallocate_memory(memory_map, process):\n if process.memory is not None:\n for i in range(len(process.code)):\n memory_map[process.memory + i] = 0\n process.memory = None\n\ndef run_processes(processes):\n for process in processes:\n print(f\"Running Process {process.pid} with code: {process.code}\")\n\n# 创建一个简单的操作系统\nmemory_map = [0] * 100\nprocesses = []\n\n# 创建两个进程\nprocess1 = Process(pid=1, code=[1, 2, 3])\nprocess2 = Process(pid=2, code=[4, 5, 6])\n\n# 将进程添加到列表\nprocesses.append(process1)\nprocesses.append(process2)\n\n# 分配内存\nallocate_memory(memory_map, process1)\nallocate_memory(memory_map, process2)\n\n# 运行进程\nrun_processes(processes)\n\n# 释放内存\ndeallocate_memory(memory_map, process1)\ndeallocate_memory(memory_map, process2)\n\n```"} 13 | {"_id": 143, "response": "```python\nimport re\nfrom collections import Counter\nfrom bs4 import BeautifulSoup\n\n\ndef get_top_k_words_from_html(html, k):\n # 定义需要排除的单词列表\n exclude_list = ['this','is','are','a', 'an', 'the', 'and', 'or', 'in', 'on', 'at', 'to', 'of', 'for', 'with', 'by', 'from']\n\n try:\n # 解析HTML文本\n soup = BeautifulSoup(html, 'lxml')\n\n # 获取所有文本内容\n text = soup.get_text()\n\n # 将文本转换为小写，并使用正则表达式提取单词\n\n lower=text.lower()\n words = re.findall(r'\\b(? Tuple[List[str], List[float]]:\n cities = []\n temperatures = []\n lines = text.strip().split('\\n')\n for line in lines:\n city_temperature = line.split('，')\n if len(city_temperature) == 2:\n city, temperature = city_temperature\n temperature = float(temperature.replace('℃', ''))\n cities.append(city)\n temperatures.append(temperature)\n else:\n return [], []\n return cities, temperatures\n\ndef calculate_average_temperature(temperatures: List[float]) -> float:\n total_temperature = sum(temperatures)\n average_temperature = total_temperature / len(temperatures)\n return average_temperature\n\ndef find_highest_temperature(cities: List[str], temperatures: List[float]) -> Tuple[float, str]:\n if not temperatures:\n return 0, \"无\"\n index = temperatures.index(max(temperatures))\n highest_temperature = temperatures[index]\n city = cities[index]\n return highest_temperature, city\n\ndef find_lowest_temperature(cities: List[str], temperatures: List[float]) -> Tuple[float, str]:\n if not temperatures:\n return 0, \"无\"\n index = temperatures.index(min(temperatures))\n lowest_temperature = temperatures[index]\n city = cities[index]\n return lowest_temperature, city\n\n\n```"} 15 | {"_id": 145, "response": "```python\nimport re\n\n\ndef count_html_tags(html):\n pattern = r'<\\s*([a-zA-Z]+)[^>]*>|<\\s*/\\s*([a-zA-Z]+)\\s*>'\n matches = re.findall(pattern, html)\n\n tag_count = 0\n nested_tags = []\n max_nested_depth = 0\n\n for opening_tag, closing_tag in matches:\n if opening_tag:\n nested_tags.append(opening_tag.lower())\n tag_count += 1\n max_nested_depth = max(max_nested_depth, len(nested_tags))\n elif nested_tags and nested_tags[-1] == closing_tag.lower():\n nested_tags.pop()\n tag_count += 1\n\n return tag_count, max_nested_depth\n\n\n\n```"} 16 | {"_id": 146, "response": "```python\nfrom cryptography.hazmat.primitives.ciphers import Cipher, algorithms, modes\nfrom cryptography.hazmat.backends import default_backend\nimport base64\n\n\ndef encrypt(text, key):\n backend = default_backend()\n cipher = Cipher(algorithms.AES(key), modes.ECB(), backend=backend)\n encryptor = cipher.encryptor()\n\n # 补全文本长度至AES块大小的倍数\n text = text + ' ' * (16 - len(text) % 16)\n\n # 加密\n ciphertext = encryptor.update(text.encode()) + encryptor.finalize()\n return base64.b64encode(ciphertext).decode()\n\n\ndef decrypt(encrypted_text, key):\n backend = default_backend()\n cipher = Cipher(algorithms.AES(key), modes.ECB(), backend=backend)\n decryptor = cipher.decryptor()\n\n # 解密\n ciphertext = base64.b64decode(encrypted_text.encode())\n decrypted_text = decryptor.update(ciphertext) + decryptor.finalize()\n return decrypted_text.decode().rstrip()\n\n\ndef save_to_file(data, filename):\n with open(filename, 'w') as file:\n file.write(data)\n\n\ndef read_from_file(filename):\n with open(filename, 'r') as file:\n return file.read()\n\n\n\ntext_to_encrypt = \"Hello, World!\"\nencryption_key = b'Sixteen byte key'\n\n# # 加密并保存到文件\n# encrypted_text = encrypt(text_to_encrypt, encryption_key)\n# save_to_file(encrypted_text, \"encrypted_data.txt\")\n#\n# # 从文件中读取并解密\n# loaded_encrypted_text = read_from_file(\"encrypted_data.txt\")\n# decrypted_text = decrypt(loaded_encrypted_text, encryption_key)\n\n\n\n```"} 17 | {"_id": 147, "response": "```python\nimport os\nimport shutil\n\ndef backup_folder(source_folder, target_folder):\n try:\n # 检查源文件夹是否存在\n if not os.path.exists(source_folder):\n raise Exception(\"Source folder does not exist.\")\n\n # 检查目标文件夹是否存在，如果不存在则创建\n if not os.path.exists(target_folder):\n os.makedirs(target_folder)\n\n # 遍历源文件夹中的所有文件和子文件夹\n for root, dirs, files in os.walk(source_folder):\n # 构建目标文件夹中的相对路径\n relative_path = os.path.relpath(root, source_folder)\n target_path = os.path.join(target_folder, relative_path)\n\n # 创建目标文件夹中的子文件夹\n if not os.path.exists(target_path):\n os.makedirs(target_path)\n\n # 复制源文件夹中的所有文件到目标文件夹中\n for file in files:\n source_file = os.path.join(root, file)\n target_file = os.path.join(target_path, file)\n shutil.copy2(source_file, target_file)\n\n print(\"Backup completed successfully.\")\n\n except Exception as e:\n print(f\"An error occurred: {e}\")\n\n```"} 18 | {"_id": 148, "response": "```python\nimport os\nimport shutil\n\ndef create_folder(folder_name):\n if not os.path.exists(folder_name):\n os.makedirs(folder_name)\n print(f'Folder {folder_name} created.')\n\ndef create_and_write_file(file_path, content):\n with open(file_path, 'w', encoding='utf-8') as file:\n file.write(content)\n print(f'File {file_path} created and written.')\n\ndef read_and_print_file(file_path):\n with open(file_path, 'r', encoding='utf-8') as file:\n content = file.read()\n print(f'Contents of {file_path}:')\n print(content)\n\ndef copy_file(source_path, destination_path):\n shutil.copy(source_path, destination_path)\n print(f'File {source_path} copied to {destination_path}.')\n\ndef move_file(source_path, destination_path):\n shutil.move(source_path, destination_path)\n print(f'File {source_path} moved to {destination_path}.')\n\ndef delete_folder(folder_name):\n shutil.rmtree(folder_name)\n print(f'Folder {folder_name} and its contents deleted.')\n\n```"} 19 | {"_id": 149, "response": "```python\nimport re\n\n\ndef extract_phone_numbers(text):\n # 电话号码的正则表达式模式\n pattern = re.compile(r'\\(?\\b(\\d{3})\\)?[-.\\s]?(\\d{3})[-.\\s]?(\\d{4})\\b')\n\n # 使用 finditer 方法提取所有匹配的电话号码\n matches = pattern.finditer(text)\n\n # 存储提取到的电话号码\n phone_numbers = []\n\n for match in matches:\n # 获取匹配对象的匹配组\n groups = match.groups()\n\n # 格式化提取到的电话号码\n formatted_number = f\"({groups[0]}) {groups[1]}-{groups[2]}\"\n\n # 添加到结果列表\n phone_numbers.append(formatted_number)\n\n return phone_numbers\n\n\n```"} 20 | {"_id": 150, "response": "```python\nimport numpy as np\nfrom scipy.linalg import eig\n\n\ndef calculate_eigen(matrix, num_largest=5):\n # 计算特征值和特征向量\n eigenvalues, eigenvectors = eig(matrix)\n\n # 找到最大的几个特征值及其对应的索引\n largest_indices = np.argsort(eigenvalues)[-num_largest:]\n\n # 提取最大的几个特征值和对应的特征向量\n largest_eigenvalues = eigenvalues[largest_indices]\n largest_eigenvectors = eigenvectors[:, largest_indices]\n\n return largest_eigenvalues, largest_eigenvectors\n\n```"} 21 | {"_id": 151, "response": "```python\ndef flatten_list(nested_list):\n flattened_list = []\n stack = [nested_list]\n\n while stack:\n current = stack.pop()\n\n for element in current:\n if isinstance(element, list):\n stack.append(element)\n else:\n flattened_list.append(element)\n\n return flattened_list\n\n\n```"} 22 | {"_id": 152, "response": "```python\nimport openpyxl\n# 打开txt文件\n\ndef txt2excle(file_path):\n with open(file_path, 'r') as file:\n lines = file.readlines()\n\n # 创建Excel工作簿\n workbook = openpyxl.Workbook()\n sheet = workbook.active\n\n # 将列表数据写入Excel工作表\n for row, line in enumerate(lines, start=1):\n # 假设列表中的元素使用逗号进行分隔\n elements = line.strip().split('\\t')\n for col, element in enumerate(elements, start=1):\n sheet.cell(row=row, column=col).value = element\n\n # 保存为Excel文件\n workbook.save('output.xlsx')\n\n```"} 23 | {"_id": 153, "response": "```python\nimport xml.etree.ElementTree as ET\n \ndef calculate_iou(box1, box2):\n \"\"\"\n 计算两个边界框之间的IoU\n \"\"\"\n x1, y1, x2, y2 = box1\n x3, y3, x4, y4 = box2\n\n # 计算相交矩形的左上角和右下角坐标\n inter_x1 = max(x1, x3)\n inter_y1 = max(y1, y3)\n inter_x2 = min(x2, x4)\n inter_y2 = min(y2, y4)\n\n # 计算相交矩形的面积\n inter_area = max(0, inter_x2 - inter_x1 + 1) * max(0, inter_y2 - inter_y1 + 1)\n\n # 计算并集面积和IoU\n box1_area = (x2 - x1 + 1) * (y2 - y1 + 1)\n box2_area = (x4 - x3 + 1) * (y4 - y3 + 1)\n union_area = box1_area + box2_area - inter_area\n iou = inter_area / union_area\n\n return iou\n\ndef convert_labels(label_file, target_region):\n \"\"\"\n 将在目标区域内与真实框标签的IoU大于0.2的所有框标签划分为类别0\n \"\"\"\n tree = ET.parse(label_file)\n root = tree.getroot()\n\n for object_elem in root.findall('object'):\n bbox = object_elem.find('bndbox')\n xmin = int(bbox.find('xmin').text)\n ymin = int(bbox.find('ymin').text)\n xmax = int(bbox.find('xmax').text)\n ymax = int(bbox.find('ymax').text)\n\n # 计算当前边界框与目标区域的IoU\n iou = calculate_iou(target_region, (xmin, ymin, xmax, ymax))\n\n # 如果IoU大于0.2，则将类别设置为0\n if iou > 0.2:\n object_elem.find('name').text = '0'\n\n tree.write(label_file)\n\n```"} 24 | {"_id": 154, "response": "```python\nimport os\nimport random\nimport shutil\n\n\ndef split_data(dataset_path,train_ratio,test_ratio,val_ratio):\n # 设置随机种子\n random_seed = 42\n random.seed(random_seed)\n train_path = os.path.join(dataset_path,'train')\n test_path = os.path.join(dataset_path,'test')\n val_path = os.path.join(dataset_path,'val')\n\n if not os.path.exists(train_path):\n os.makedirs(train_path)\n\n if not os.path.exists(test_path):\n os.makedirs(test_path)\n\n if not os.path.exists(val_path):\n os.makedirs(val_path)\n\n image_files = [f for f in os.listdir(dataset_path) if f.endswith('.jpg')]\n\n # 随机打乱图像文件列表\n random.shuffle(image_files)\n\n # 计算划分的索引位置\n train_split = int(len(image_files) * train_ratio)\n test_split = int(len(image_files) * (train_ratio + test_ratio))\n\n # 划分训练集\n train_files = image_files[:train_split]\n for file in train_files:\n src_path = os.path.join(dataset_path, file)\n dest_path = os.path.join(train_path, file)\n shutil.copy(src_path, dest_path)\n\n # 划分测试集\n test_files = image_files[train_split:test_split]\n for file in test_files:\n src_path = os.path.join(dataset_path, file)\n dest_path = os.path.join(test_path, file)\n shutil.copy(src_path, dest_path)\n\n # 划分验证集\n val_files = image_files[test_split:]\n for file in val_files:\n src_path = os.path.join(dataset_path, file)\n dest_path = os.path.join(val_path, file)\n shutil.copy(src_path, dest_path)\n\n\n# import os\n# import shutil\n# from sklearn.model_selection import train_test_split\n\n\n# def split_data(dataset_path, train_ratio, test_ratio, val_ratio):\n# # 获取所有文件的路径\n# files = [os.path.join(dataset_path, f) for f in os.listdir(dataset_path) if\n# os.path.isfile(os.path.join(dataset_path, f))]\n#\n# # 划分训练集和其余部分\n# train_files, rest_files = train_test_split(files, train_size=train_ratio, random_state=42)\n#\n# # 计算测试集的比例\n# test_ratio_adjusted = test_ratio / (1 - train_ratio)\n#\n# # 划分测试集和验证集\n# test_files, val_files = train_test_split(rest_files, test_size=test_ratio_adjusted, random_state=42)\n#\n# # 创建训练集、测试集和验证集的文件夹\n# train_dir = os.path.join(dataset_path, 'train')\n# test_dir = os.path.join(dataset_path, 'test')\n# val_dir = os.path.join(dataset_path, 'val')\n# os.makedirs(train_dir, exist_ok=True)\n# os.makedirs(test_dir, exist_ok=True)\n# os.makedirs(val_dir, exist_ok=True)\n#\n# # 将文件移动到对应的文件夹\n# for f in train_files:\n# shutil.move(f, os.path.join(train_dir, os.path.basename(f)))\n# for f in test_files:\n# shutil.move(f, os.path.join(test_dir, os.path.basename(f)))\n# for f in val_files:\n# shutil.move(f, os.path.join(val_dir, os.path.basename(f)))\n```"} 25 | {"_id": 155, "response": "```python\nimport numpy as np\n\ndef hinge_loss(y_true, y_pred):\n return np.mean(np.maximum(0, 1 - y_true * y_pred))\n\n```"} 26 | {"_id": 156, "response": "```python\nimport os\n\n\ndef remove_empty_lines_folder(folder_path):\n for file_name in os.listdir(folder_path):\n if file_name.endswith('.txt'):\n file_path = os.path.join(folder_path, file_name)\n remove_empty_lines(file_path)\n\n\ndef remove_empty_lines(file_path):\n with open(file_path, 'r+') as file:\n lines = file.readlines()\n file.seek(0)\n file.truncate()\n for line in lines:\n if line.strip(): # 检查是否为空行\n file.write(line)\n\n# folder_path=r'C:\\Users\\Administrator\\Desktop\\work\\1207-Q17\\yolo\\labels'\n# remove_empty_lines_folder(folder_path)\n\n```"} 27 | {"_id": 157, "response": "```python\nimport os\n\n\ndef label_filter(folder_path,Label_category):\n\n txt_files = [f for f in os.listdir(folder_path) if f.endswith('.txt')]\n for file_name in txt_files:\n file_path = os.path.join(folder_path, file_name)\n with open(file_path, 'r') as file:\n lines = file.readlines()\n\n new_lines = []\n\n\n for line in lines:\n parts = line.split()\n\n if int(parts[0]) in Label_category :\n continue\n new_line = ' '.join(parts) + '\\n'\n new_lines.append(new_line)\n\n # 将处理后的行写回文件\n with open(file_path, 'w') as file:\n file.writelines(new_lines)\n\n```"} 28 | {"_id": 158, "response": "```python\nimport os\n\n\ndef w_h_filter(folder_path,limit_wideth,limit_height):\n txt_files = [f for f in os.listdir(folder_path) if f.endswith('.txt')]\n for file_name in txt_files:\n file_path = os.path.join(folder_path, file_name)\n with open(file_path, 'r') as file:\n lines = file.readlines()\n\n new_lines = []\n\n\n for line in lines:\n parts = line.split()\n\n if (float(parts[3]) < limit_wideth or float(parts[4]) < limit_height):\n continue\n new_line = ' '.join(parts) + '\\n'\n new_lines.append(new_line)\n\n # 将处理后的行写回文件\n with open(file_path, 'w') as file:\n file.writelines(new_lines)\n\n\n\n\n# if __name__ == \"__main__\":\n#\n# folder_path = r\"C:\\Users\\Administrator\\Desktop\\work\\1207-Q14\\labels\"\n# limit_wideth=0.1\n# limit_height=0.02\n# w_h_filter(folder_path,limit_wideth,limit_height)\n\n```"} 29 | {"_id": 159, "response": "```python\nimport numpy as np\nimport torch\nimport torch.nn as nn\nimport torch.optim as optim\n\ndef mean_squared_error(y_true, y_pred):\n return np.mean(np.square(y_true - y_pred))\n```"} 30 | {"_id": 160, "response": "```python\nimport nltk\nnltk.data.path.append('nltk_package')\nfrom nltk.tokenize import word_tokenize\nfrom nltk.stem import PorterStemmer\nfrom nltk import pos_tag\n\n\ndef process_sentence(sentence):\n tokens = word_tokenize(sentence)\n pos_tags = pos_tag(tokens)\n # stop_words = set(stopwords.words('english'))\n # filtered_tokens = [word.lower() for word in tokens if word.lower() not in stop_words]\n filtered_tokens = [word.lower() for word in tokens]\n stemmer = PorterStemmer()\n stemmed_tokens = [stemmer.stem(word) for word in filtered_tokens]\n # 输出结果\n print(\"Original Sentence:\\n\", sentence)\n print(\"\\nTokenized Sentence:\\n\", tokens)\n print(\"\\nPOS Tagged Sentence:\\n\", pos_tags)\n print(\"\\nSentence without Stopwords:\\n\", filtered_tokens)\n print(\"\\nStemmed Sentence:\\n\", stemmed_tokens)\n return tokens, pos_tags, filtered_tokens, stemmed_tokens\n\nprint(process_sentence(\"!@#$%^&*()\"))\n\n```"} 31 | {"_id": 161, "response": "```python\nimport pandas as pd\nfrom sklearn.preprocessing import OneHotEncoder\n\n\ndef calculate_and_save_probabilities(input_file, output_file):\n # 读取数据集\n df = pd.read_csv(input_file)\n\n # 计算概率\n probabilities = {}\n for column_value in df['Sex'].unique():\n prob = df[(df['Sex'] == column_value) & (df[\"Survived\"] == 1)].shape[0] / df[df['Sex'] == column_value].shape[0]\n probabilities[f\"Probability({'Sex'}={column_value}, {'Survived'}=1)\"] = prob\n\n # 打印计算结果\n for key, value in probabilities.items():\n print(key + \":\", value)\n\n # 进行独热编码\n non_numeric_columns = df.select_dtypes(include=['object']).columns\n encoder = OneHotEncoder(sparse=False)\n encoded_features = pd.DataFrame(encoder.fit_transform(df[non_numeric_columns]))\n encoded_features.columns = encoder.get_feature_names_out(non_numeric_columns)\n\n # 替换原始数据集中的非数值列\n df.drop(non_numeric_columns, axis=1, inplace=True)\n df = pd.concat([df, encoded_features], axis=1)\n\n # 将数据集保存到新文件\n df.to_csv(output_file, index=False)\n\n return probabilities[f\"Probability(Sex=male, Survived=1)\"], probabilities[\"Probability(Sex=female, Survived=1)\"]\n\n\nprint(calculate_and_save_probabilities(\"test6/dataset.csv\", \"test6/test_label.csv\"))\n```"} 32 | {"_id": 162, "response": "```python\nimport numpy as np\n\n\ndef linear_fit(x, y):\n x = np.array(x)\n y = np.array(y)\n\n mean_x = np.mean(x)\n mean_y = np.mean(y)\n\n SS_xy = np.dot(x - mean_x, y - mean_y)\n SS_xx = np.dot(x - mean_x, x - mean_x)\n a = SS_xy / SS_xx\n b = mean_y - a * mean_x\n\n r = np.corrcoef(x, y)[0, 1]\n\n return a, b, r\n\n\n```"} 33 | {"_id": 163, "response": "```python\nimport os\n\ndef list_files(directory_path):\n # 检查目录是否存在\n if not os.path.exists(directory_path) or not os.path.isdir(directory_path):\n print(\"Invalid directory path.\")\n return\n\n list_files_recursive(directory_path)\n\ndef list_files_recursive(directory):\n # 获取目录中的所有文件和子目录\n files = os.listdir(directory)\n for file in files:\n file_path = os.path.join(directory, file)\n if os.path.isfile(file_path):\n # 如果是文件，则输出文件名\n print(file)\n elif os.path.isdir(file_path):\n # 如果是目录，则输出目录名，并递归调用列出子目录\n list_files_recursive(file_path)\n\n```"} 34 | {"_id": 164, "response": "```python\nimport os\n\ndef find_file(file_name, directory_path):\n exist = False\n # 使用os.walk遍历指定目录及其子目录中的文件\n for root, dirs, files in os.walk(directory_path):\n for file in files:\n if file == file_name:\n exist = True\n # 如果找到文件，输出文件的绝对路径\n file_path = os.path.join(root, file)\n print(file_path)\n \n if not exist:\n print(\"The file is not on disk\")\n\n```"} 35 | {"_id": 165, "response": "```python\nimport csv\n\ndef calculate_average(filename, column_index):\n with open(filename, 'r') as file:\n reader = csv.reader(file)\n next(reader) \n total, count = 0, 0\n for row in reader:\n if row[column_index].isdigit():\n total += int(row[column_index])\n count += 1\n return total / count if count != 0 else 0\n\n\n```"} 36 | {"_id": 166, "response": "```python\nimport os\nimport pandas as pd\n\n\ndef find_max_temperature(folder_path, output_file):\n csv_files = [f for f in os.listdir(folder_path) if f.endswith('.csv')]\n \n result_df = pd.DataFrame(columns=['City', 'Max_Temperature', 'Date'])\n \n for csv_file in csv_files:\n file_path = os.path.join(folder_path, csv_file)\n df = pd.read_csv(file_path)\n \n city_name = csv_file[:-4]\n \n # 找到最高气温的值\n max_temp = df['Temperature'].max()\n \n # 找到所有最高气温的日期\n max_temp_dates = df.loc[df['Temperature'] == max_temp, 'Date'].tolist()\n \n # 将结果添加到result_df\n for date in max_temp_dates:\n result_df = result_df._append({'City': city_name,\n 'Max_Temperature': max_temp,\n 'Date': date}, ignore_index=True)\n \n result_df.to_csv(output_file, index=False)\n\n```"} 37 | {"_id": 167, "response": "```python\nimport pandas as pd\nimport torch\n\ndef merge_and_smooth_data(filenames, window_size):\n dfs = [pd.read_csv(filename) for filename in filenames]\n merged_df = pd.concat(dfs).sort_values(by='timestamp')\n\n tensor = torch.tensor(merged_df.drop(columns=['timestamp']).values)\n\n cumsum = tensor.cumsum(dim=0)\n sliding_avg = (cumsum[window_size:] - cumsum[:-window_size]) / float(window_size)\n return sliding_avg\n\n\n```"} 38 | {"_id": 168, "response": "```python\nfrom PIL import Image\nimport os\n\n\ndef resize_and_rename_images(input_folder, output_folder, target_resolution=(800, 600)):\n if not os.path.exists(output_folder):\n os.makedirs(output_folder)\n image_files = [f for f in os.listdir(input_folder) if f.endswith(('.png', '.jpg', '.jpeg'))]\n image_files.sort() # 按文件名排序\n for i, image_file in enumerate(image_files):\n input_path = os.path.join(input_folder, image_file)\n output_path = os.path.join(output_folder, f\"{i + 1:03d}.jpg\") # 格式化文件名\n\n # 打开图片\n img = Image.open(input_path)\n\n # 转换为RGB模式\n img = img.convert(\"RGB\")\n\n # 统一分辨率\n img = img.resize(target_resolution, Image.LANCZOS)\n\n # 保存图片\n img.save(output_path)\n\n print(\"Modified images resolutions:\")\n print_resolutions(output_folder)\n return output_folder\n\n\ndef print_resolutions(folder_path):\n image_files = [f for f in os.listdir(folder_path) if f.endswith(('.png', '.jpg', '.jpeg'))]\n image_files.sort()\n\n for image_file in image_files:\n image_path = os.path.join(folder_path, image_file)\n img = Image.open(image_path)\n resolution = img.size\n print(f\"{image_file}: {resolution[0]} x {resolution[1]}\")\n\n\nresize_and_rename_images('test6/input', 'test6/output', (2560, 1440))\n```"} 39 | {"_id": 169, "response": "```python\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\n\nclass FocalLoss(nn.Module):\n def __init__(self, alpha=0.25, gamma=2.0):\n super(FocalLoss, self).__init__()\n self.alpha = alpha\n self.gamma = gamma\n\n def forward(self, inputs, targets):\n \n BCE_loss = F.binary_cross_entropy_with_logits(inputs, targets, reduction='none')\n pt = torch.exp(-BCE_loss) \n focal_loss = self.alpha * (1-pt)**self.gamma * BCE_loss\n return focal_loss.mean()\n\n\ncriterion = FocalLoss(alpha=0.25, gamma=2.0)\ninputs = torch.tensor([0.1, 0.2, 0.4, 0.8, 1.2, -0.1, -0.2, -0.4, -0.8, -1.2], requires_grad=True)\ntargets = torch.tensor([1., 0., 1., 0., 1., 0., 1., 0., 1., 0.])\n\nloss = criterion(inputs, targets).detach()\nprint(loss)\n\n\n```"} 40 | {"_id": 170, "response": "```python\nimport os\nimport shutil\n\n\ndef copy_and_move_file(source_path, destination_directory):\n # 获取源文件的文件名\n file_name = os.path.basename(source_path)\n\n # 构建目标路径\n destination_path = os.path.join(destination_directory, file_name)\n\n # 检查源文件是否存在\n if os.path.exists(source_path):\n # 复制文件\n shutil.copy(source_path, destination_path)\n else:\n print(\"File not exist.\")\n\n```"} 41 | {"_id": 171, "response": "```python\nfrom docx import Document\n\n\ndef count_keywords_in_docx(docx_path, keywords):\n doc = Document(docx_path)\n text = \" \".join([paragraph.text for paragraph in doc.paragraphs])\n\n keyword_counts = {}\n for keyword in keywords:\n count = text.lower().count(keyword.lower())\n keyword_counts[keyword] = count\n\n return keyword_counts\n\n```"} 42 | {"_id": 172, "response": "```python\nimport shutil\nimport os\n\ndef delete_folder(folder_path):\n try:\n if os.path.exists(folder_path):\n if os.path.isdir(folder_path):\n shutil.rmtree(folder_path)\n print(f\"Successfully delete the folder.\")\n else:\n print(f\"{folder_path} is not a directory\")\n else:\n print(f\"Path not exist: {folder_path}\")\n except Exception as e:\n print(f\"Can't delete the folder.\")\n\n```"} 43 | {"_id": 173, "response": "```python\nimport numpy as np\nimport scipy.sparse as sp\n\n\ndef csr_to_numpy(input_file, output_file):\n input_arr_csr = sp.load_npz(input_file)\n input_arr = input_arr_csr.toarray()\n np.save(output_file, input_arr)\n\n```"} 44 | {"_id": 174, "response": "```python\nimport shutil\nimport os\nimport threading\nimport time\n\n\ndef copy_folder(source_folder, dest_folders):\n for dest_folder in dest_folders:\n try:\n if os.path.exists(dest_folder):\n shutil.rmtree(dest_folder)\n shutil.copytree(source_folder, dest_folder)\n print(\"success from {source_folder} to {dest_folder}\")\n except Exception as e:\n print(\"copy {source_folder} to {dest_folder} is err:{e}\")\ndef create_thread(source_folder, dest_folders):\n # 计算每个线程需要拷贝的目标文件夹数量\n threads_num = len(dest_folders)\n folders_per_thread = len(dest_folders) // threads_num\n\n threads = []\n for i in range(threads_num):\n start = i * folders_per_thread\n end = start + folders_per_thread\n if i == threads_num - 1:\n end = len(dest_folders)\n thread = threading.Thread(target=copy_folder, args=(source_folder, dest_folders[start:end]))\n threads.append(thread)\n \n return threads\n \ndef run(threads):\n # 记录开始时间\n start_time = time.time()\n # 启动所有线程\n for thread in threads:\n thread.start()\n # 等待所有线程结束\n for thread in threads:\n thread.join()\n # 记录结束时间并输出用时\n end_time = time.time()\n print(f\"time: {end_time - start_time} 秒\")\n \n \ndef multithread_copy(source_folder, dest_folders):\n threads = create_thread(source_folder, dest_folders)\n run(threads)\n\n```"} 45 | {"_id": 175, "response": "```python\nimport numpy as np\nimport scipy.sparse as sp\n\ndef numpy_to_csr(input_file, output_file):\n input_arr = np.load(input_file)\n input_arr_csr = sp.csr_matrix(input_arr)\n sp.save_npz(output_file, input_arr_csr)\n```"} 46 | {"_id": 176, "response": "```python\nimport numpy as np\nfrom sklearn.metrics.pairwise import cosine_similarity\n\n\ndef calculate_similarity(file_A, file_B, target_file):\n A = np.load(file_A)\n B = np.load(file_B)\n cos_sim = cosine_similarity(A, B)\n np.save(target_file, cos_sim)\n\n\n```"} 47 | {"_id": 177, "response": "```python\nimport torch\nimport numpy as np\n\n\ndef process_tensor(array1, array2):\n tensor1 = torch.FloatTensor(array1)\n tensor2 = torch.FloatTensor(array2)\n tensor2 = tensor2.permute(0, 2, 1)\n attention_scores = torch.bmm(tensor1, tensor2)\n with open('result.txt', 'w') as file:\n file.write(str(attention_scores))\n return attention_scores\n\n\narray1 = np.ones((100, 130, 80))\narray2 = np.ones((100, 130, 80))\nprint(process_tensor(array1, array2).shape)\n\n\n```"} 48 | {"_id": 178, "response": "```python\nimport math\n\ndef euclidean_distance(point1, point2):\n return math.sqrt(sum((p - q) ** 2 for p, q in zip(point1, point2)))\n\ndef manhattan_distance(point1, point2):\n return sum(abs(p - q) for p, q in zip(point1, point2))\n\ndef chebyshev_distance(point1, point2):\n return max(abs(p - q) for p, q in zip(point1, point2))\n\ndef minkowski_distance(point1, point2, p):\n return math.pow(sum(abs(p - q) ** p for p, q in zip(point1, point2)), 1/p)\n\ndef cosine_similarity(vector1, vector2):\n dot_product = sum(p * q for p, q in zip(vector1, vector2))\n norm_vector1 = math.sqrt(sum(p ** 2 for p in vector1))\n norm_vector2 = math.sqrt(sum(q ** 2 for q in vector2))\n return dot_product / (norm_vector1 * norm_vector2)\n\n\ndef hamming_distance(string1, string2):\n return sum(p != q for p, q in zip(string1, string2))\n\ndef jaccard_similarity(set1, set2):\n intersection_size = len(set1.intersection(set2))\n union_size = len(set1.union(set2))\n return intersection_size / union_size\n\n\n\n\n```"} 49 | {"_id": 179, "response": "```python\nimport numpy as np\n\ndef diag_zero(matrix):\n matrix[np.triu_indices(matrix.shape[0], k=1)] = 0\n sum_remaining_values = np.sum(matrix)\n return sum_remaining_values\n\nmatrix = np.arange(1, 10).reshape(3, 3)\nprint(diag_zero(matrix))\n\n\n```"} 50 | {"_id": 180, "response": "```python\nimport numpy as np\n\n\ndef array_avg(matrix):\n # 计算数组元素的平均值\n mean_value = np.mean(matrix)\n\n # 大于平均值的元素置为零\n matrix[matrix > mean_value] = 0\n\n # 将数组展开为一维列表\n flattened_array = matrix.flatten()\n return mean_value, flattened_array\n\nmatrix = [[2,1,4,3],[4,5,8,9],[10,2,3,4],[1,14,21,2]]\nmatrix = np.array(matrix)\nprint(array_avg(matrix))\n\n```"} 51 | {"_id": 181, "response": "```python\nimport os\n\ndef count_files(dir):\n count = 0\n for root, dirs, files in os.walk(dir):\n count += len(files)\n return count\n\n```"} 52 | {"_id": 182, "response": "```python\nimport pandas as pd\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn.decomposition import PCA\n\ndef pac_fit(input_file):\n characters = pd.read_csv(input_file)\n characters = StandardScaler().fit_transform(characters)\n pca = PCA().fit(characters)\n return pca.explained_variance_ratio_\n\n\ndf = pd.DataFrame({'A': [1, 2], 'B': [2, 3], 'C': [3, 4]})\ndf.to_csv('test.csv', index=False)\nprint(pac_fit('test.csv'))\n```"} 53 | {"_id": 183, "response": "```python\nimport os\nimport shutil\nimport zipfile\nfrom datetime import datetime\nimport glob\n\ndef backup(source_path, backup_folder, max_size_mb=10):\n # 创建备份文件夹\n if not os.path.exists(backup_folder):\n os.makedirs(backup_folder)\n\n # 生成时间戳\n timestamp = datetime.now().strftime('%Y%m%d%H%M%S')\n\n # 构建备份文件名\n backup_filename = f\"backup_{timestamp}.zip\"\n backup_path = os.path.join(backup_folder, backup_filename)\n\n # 备份文件夹\n shutil.make_archive(backup_path[:-4], 'zip', source_path)\n\n # 获取备份文件夹总大小\n total_size_mb = sum(os.path.getsize(f) for f in glob.glob(os.path.join(backup_folder, '*.zip'))) / (1024 ** 2)\n\n # 如果总大小超过设定的阈值，删除最旧的备份文件\n while total_size_mb > max_size_mb:\n oldest_backup = min(glob.glob(os.path.join(backup_folder, '*.zip')), key=os.path.getctime)\n total_size_mb -= os.path.getsize(oldest_backup) / (1024 ** 2)\n os.remove(oldest_backup)\n\n print(f\"Backup completed: {backup_path}\")\n```"} 54 | {"_id": 184, "response": "```python\nimport torch\n\ndef bbox_iou(boxes1, boxes2):\n \"\"\"\n 计算两组对应位置bounding boxes的IoU。\n \n 参数:\n boxes1, boxes2: 维度为[N, 4]的tensor，表示bounding boxes，格式为[x, y, w, h]。\n \n 返回:\n 一个长度为N的tensor，表示boxes1和boxes2中对应box的IoU。\n \"\"\"\n assert boxes1.shape == boxes2.shape\n boxes1_x1 = boxes1[:, 0] - boxes1[:, 2] / 2\n boxes1_y1 = boxes1[:, 1] - boxes1[:, 3] / 2\n boxes1_x2 = boxes1[:, 0] + boxes1[:, 2] / 2\n boxes1_y2 = boxes1[:, 1] + boxes1[:, 3] / 2\n\n boxes2_x1 = boxes2[:, 0] - boxes2[:, 2] / 2\n boxes2_y1 = boxes2[:, 1] - boxes2[:, 3] / 2\n boxes2_x2 = boxes2[:, 0] + boxes2[:, 2] / 2\n boxes2_y2 = boxes2[:, 1] + boxes2[:, 3] / 2\n\n\n inter_x1 = torch.max(boxes1_x1, boxes2_x1)\n inter_y1 = torch.max(boxes1_y1, boxes2_y1)\n inter_x2 = torch.min(boxes1_x2, boxes2_x2)\n inter_y2 = torch.min(boxes1_y2, boxes2_y2)\n\n inter_area = torch.clamp(inter_x2 - inter_x1, min=0) * torch.clamp(inter_y2 - inter_y1, min=0)\n\n\n boxes1_area = (boxes1_x2 - boxes1_x1) * (boxes1_y2 - boxes1_y1)\n boxes2_area = (boxes2_x2 - boxes2_x1) * (boxes2_y2 - boxes2_y1)\n\n\n union_area = boxes1_area + boxes2_area - inter_area\n\n\n iou = inter_area / union_area\n\n return iou\n\n\n\n\n```"} 55 | {"_id": 185, "response": "```python\nimport re\nfrom collections import defaultdict\nimport json\n\ndef nginx_log_analysis(nginx_log_file, statistics_output_file, high_frequency_output_file):\n # 定义一个字典用于存储IP地址访问次数\n ip_count = defaultdict(int)\n\n # 定义一个集合用于存储高频率访问的IP地址\n high_frequency_ips = defaultdict(int)\n \n # 定义正则表达式来匹配IP地址\n ip_pattern = re.compile(r'\\b(?:[0-9]{1,3}\\.){3}[0-9]{1,3}\\b')\n \n # 打开Nginx日志文件并逐行处理\n with open(nginx_log_file, 'r') as file:\n for line in file:\n # 使用正则表达式匹配IP地址\n match = ip_pattern.search(line)\n if match:\n ip_address = match.group(0)\n \n # 增加IP地址访问次数\n ip_count[ip_address] += 1\n \n # 判断是否是高频率访问\n if ip_count[ip_address] > 30:\n high_frequency_ips[ip_address] = ip_count[ip_address]\n\n print(ip_count)\n # 输出IP地址及其访问次数到统计文件\n with open(statistics_output_file, 'w') as output:\n json.dump(ip_count, output, ensure_ascii=False, indent=2)\n \n # 输出高频率访问的IP地址到单独的文件\n with open(high_frequency_output_file, 'w') as output:\n json.dump(high_frequency_ips, output, ensure_ascii=False, indent=2)\n\n```"} 56 | {"_id": 186, "response": "```python\ndef lengthLongestPath(input: str) -> int:\n st = []\n ans, i, n = 0, 0, len(input)\n while i < n:\n # 检测当前文件的深度\n depth = 1\n while i < n and input[i] == '\\t':\n depth += 1\n i += 1\n # 统计当前文件名的长度\n length, isFile = 0, False\n while i < n and input[i] != '\\n':\n if input[i] == '.':\n isFile = True\n length += 1\n i += 1\n i += 1 # 跳过换行符\n while len(st) >= depth:\n st.pop()\n if st:\n length += st[-1] + 1\n if isFile:\n ans = max(ans, length)\n else:\n st.append(length)\n return ans\n\n\n```"} 57 | {"_id": 187, "response": "```python\nfrom sklearn.cluster import KMeans\nimport numpy as np\n\ndef k_means_clustering(data, num_clusters):\n kmeans = KMeans(n_clusters=num_clusters, random_state=42)\n kmeans.fit(data)\n return kmeans.labels_\n\n```"} 58 | {"_id": 188, "response": "```python\nimport cv2\nimport numpy as np\n\ndef gamma_correction(image, gamma):\n # 将图像的亮度值转换为范围在[0, 1]之间的浮点数\n image = image / 255.0\n \n # 对图像进行伽马亮度矫正\n corrected_image = np.power(image, gamma)\n \n # 将亮度值重新缩放到范围[0, 255]\n corrected_image = np.uint8(corrected_image * 255)\n \n return corrected_image\n```"} 59 | {"_id": 189, "response": "```python\n\nimport os\nfrom PyPDF2 import PdfMerger\n\n# 定义一个函数来合并PDF文件\ndef merge_pdfs(root_dir):\n def merge(files, output):\n merger = PdfMerger()\n for pdf in files:\n merger.append(pdf)\n merger.write(output)\n merger.close()\n\n # 遍历文件夹\n for root, dirs, files in os.walk(root_dir):\n pdf_files = [f for f in files if f.endswith(\".pdf\")]\n if pdf_files: # 如果存在PDF文件\n output = os.path.join(root, \"merged.pdf\") # 输出文件名\n full_pdf_files = [os.path.join(root, f) for f in pdf_files] # 获取完整路径\n full_pdf_files = sorted(full_pdf_files)\n merge(full_pdf_files, output) # 合并PDF文件\n\n\n\n\n\n\n\n```"} 60 | {"_id": 190, "response": "```python\n\nimport pandas as pd\nfrom io import StringIO\n\ndef calculate_date_difference(csv_contents):\n # Read the CSV contents into a pandas DataFrame\n df = pd.read_csv(StringIO(csv_contents))\n\n # Convert the 'Date' column to datetime\n df['Date'] = pd.to_datetime(df['Date'])\n\n # Calculate the difference in days between the earliest and latest date\n date_difference = (df['Date'].max() - df['Date'].min()).days\n\n return date_difference\n\n\n```"} 61 | {"_id": 191, "response": "```python\n\nimport os\nimport shutil\n\ndef copy_files(file_path, dest_path, pairs_per_folder=10):\n files = os.listdir(file_path)\n files.sort()\n pairs = zip(files[::2], files[1::2])\n for i, pair in enumerate(pairs):\n sub_folder = os.path.join(dest_path, 'sub{}'.format(i // pairs_per_folder))\n os.makedirs(sub_folder, exist_ok=True)\n for file in pair:\n shutil.copy(os.path.join(file_path, file), sub_folder)\n\n# 使用方法\n# copy_files('/path/to/source/folder', '/path/to/destination/folder')\n\n\n```"} 62 | {"_id": 192, "response": "```python\n\nimport re\n\ndef replace_code_block(text):\n # 定义一个字典，将代码块标记映射到相应的HTML类\n languages = {\n \"html\": \"language-html\",\n \"css\": \"language-css\",\n \"python\": \"language-python\",\n \"javascript\": \"language-javascript\",\n \"golang\": \"language-golang\"\n }\n\n # 对于每种语言，使用正则表达式查找和替换相应的代码块\n for lang, html_class in languages.items():\n # 构建正则表达式模式\n pattern = rf\"```{lang}\\b\\s*(.*?)\\s*```\"\n # 替换文本\n replacement = rf'
\\1
'\n # 应用替换\n text = re.sub(pattern, replacement, text, flags=re.DOTALL)\n\n return text\n\n\n\n```"} 63 | {"_id": 193, "response": "```python\n\ndef calculate_et(delta, rn, g, gamma, t, u2, es, ea):\n numerator = 0.408 * delta * (rn - g) + gamma * (900 / (t + 273)) * u2 * (es - ea)\n denominator = delta + gamma * (1 + 0.34 * u2)\n if denominator == 0:\n return 0\n et = numerator / denominator\n\n print(et)\n return et\n\n\n```"} 64 | {"_id": 194, "response": "```python\n\nfrom lxml import etree\n\nhtml = \"\"\"\n

\n 3131\n AAAA\n BBB\n CCC\n

\n\"\"\"\n\nroot = etree.fromstring(html)\n\ndef has_strong_or_em(tag):\n if tag.find('strong') is not None or tag.find('em') is not None:\n return True\n return False\n\nfor p in root.iter('p'):\n if has_strong_or_em(p):\n print(\"P tag has strong or em tag\")\n else:\n print(\"P tag does not have strong or em tag\")\n\n\n```"} 65 | {"_id": 195, "response": "```python\n\nimport pandas as pd\n\ndef flatten_dict_to_rows(d):\n def flatten_dict(d, parent_key='', sep='.'):\n items = []\n for k, v in d.items():\n new_key = f\"{parent_key}{sep}{k}\" if parent_key else k\n if isinstance(v, dict):\n if not v:\n items.append((new_key, {}))\n else:\n items.extend(flatten_dict(v, new_key, sep=sep).items())\n else:\n items.append((new_key, v))\n return dict(items)\n\n flat_dict = flatten_dict(d)\n\n # Convert the flattened dictionary to a pandas DataFrame\n df = pd.DataFrame(list(flat_dict.items()), columns=['Key', 'Value'])\n print(df.to_dict())\n return df\n\n\n```"} 66 | {"_id": 196, "response": "```python\n\ndef is_mac_in_range(mac, mac_range):\n def mac_to_hex(mac_str):\n return mac_str.replace(':', '').replace('-', '').upper()\n\n # 解析MAC地址范围\n start_mac_str, end_mac_str = mac_range.split('-')\n start_mac = int(mac_to_hex(start_mac_str), 16)\n end_mac = int(mac_to_hex(end_mac_str), 16)\n mac_val = int(mac_to_hex(mac), 16)\n\n # 判断MAC地址是否在范围内\n return start_mac <= mac_val <= end_mac\n\n\n```"} 67 | {"_id": 197, "response": "```python\n\nimport numpy as np\nfrom scipy.optimize import curve_fit\n\n# Define the fractional order model with 2 constant phase elements and 3 resistors\ndef fractional_order_model(t, Q, R1, R2, R3, n1, n2):\n return Q * (1 - (1 / (R1 * (t ** n1) + 1)) - (1 / (R2 * (t ** n2) + 1)) - (1 / R3))\n\n# # Sample time and current data\n# time_data = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10])\n# current_data = np.array([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0])\n#\n# # Fit the model to the data\n# popt, pcov = curve_fit(fractional_order_model, time_data, current_data, bounds=(0, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0]))\n# print(len(popt))\n# # The estimated battery capacity in Amp hour is the first parameter of the model\n# estimated_capacity_Ah = popt[0]\n#\n# print(\"Estimated battery capacity: {:.2f} Ah\".format(estimated_capacity_Ah))\n\n\n```"} 68 | {"_id": 198, "response": "```python\n\nimport numpy as np\n\ndef solve_quadratic(points):\n # Create matrix A using a list comprehension\n A = np.array([[x**2, x, 1] for x, y in points])\n # Create vector b\n b = np.array([y for x, y in points])\n # Use numpy.linalg.solve to solve for [a, b, c]\n coefficients = np.linalg.solve(A, b)\n\n return coefficients\n\n```"} 69 | {"_id": 199, "response": "```python\n\nimport ipaddress\n\ndef check_ip_overlap(ip_range1, ip_range2):\n ip_net1 = ipaddress.ip_network(ip_range1)\n ip_net2 = ipaddress.ip_network(ip_range2)\n\n return ip_net1.overlaps(ip_net2)\n\n\n\n```"} 70 | {"_id": 200, "response": "```python\n\nimport os\nimport re\n\ndef rename_files_in_folder(folder_path):\n for filename in os.listdir(folder_path):\n file_type = filename.split('.')[-1]\n new_filename = re.sub(\"[A-Za-z]\", \"\", filename[:-len(file_type)]) + file_type\n os.rename(os.path.join(folder_path, filename), os.path.join(folder_path, new_filename))\n\n# Call the function with the path to the folder\n# rename_files_in_folder('/path/to/your/folder')\n\n\n```"} 71 | --------------------------------------------------------------------------------