diff options
Diffstat (limited to '.agents/skills/citation-verification/scripts')
4 files changed, 1906 insertions, 0 deletions
diff --git a/.agents/skills/citation-verification/scripts/README.md b/.agents/skills/citation-verification/scripts/README.md new file mode 100644 index 0000000..5f1bd38 --- /dev/null +++ b/.agents/skills/citation-verification/scripts/README.md @@ -0,0 +1,81 @@ +# Citation Verification Scripts + +## 状态说明 + +**这些脚本是参考实现，不是主要工作流的一部分。** + +本目录中的Python脚本提供了基于API的引用验证实现，但**实际的引用验证工作流使用WebSearch和Google Scholar**，而不是这些脚本。 + +## 为什么保留这些脚本？ + +这些脚本作为**参考实现**保留，用于： + +1. **理解验证逻辑** - 展示引用验证的完整逻辑和步骤 +2. **学习API使用** - 了解如何使用CrossRef、arXiv、Semantic Scholar等API +3. **高级用例** - 对于需要批量验证或自动化的场景，可以参考这些实现 + +## 主要工作流 + +**实际的引用验证应该使用 `ml-paper-writing` skill 中的 Citation Workflow：** + +1. 使用 WebSearch 查找论文 +2. 在 Google Scholar 上验证 +3. 从 Google Scholar 获取 BibTeX +4. 验证声明（如需要） +5. 添加到 bibliography + +详见 `ml-paper-writing` skill 的 "Citation Workflow (Hallucination Prevention)" 部分。 + +## 脚本说明 + +### verify-citations.py + +完整的引用验证脚本，包含： +- 四层验证机制（格式、存在性、信息匹配、内容验证） +- 多API支持（CrossRef、arXiv、Semantic Scholar） +- 报告生成 + +**用途**: 参考实现，了解完整的验证逻辑 + +### api-clients.py + +API客户端库，包含： +- CrossRefClient - DOI验证 +- ArXivClient - arXiv论文验证 +- SemanticScholarClient - 通用学术搜索 +- CitationAPIManager - 统一API管理 + +**用途**: 参考实现，了解如何使用学术API + +### format-checker.py + +BibTeX和LaTeX格式检查工具，包含： +- BibTeX格式验证 +- LaTeX引用检查 +- 格式错误报告 + +**用途**: 参考实现，了解格式检查逻辑 + +## 使用建议 + +**对于日常论文写作**: +- ✅ 使用 `ml-paper-writing` skill 的 Citation Workflow +- ✅ 使用 WebSearch 和 Google Scholar +- ❌ 不要使用这些Python脚本 + +**对于批量验证或自动化**: +- 可以参考这些脚本的实现 +- 根据需要修改和使用 +- 注意API速率限制 + +## 依赖安装 + +如果需要运行这些脚本（仅用于参考或高级用例）： + +```bash +pip install bibtexparser requests semanticscholar arxiv +``` + +## 更多信息 + +详见 `citation-verification` skill 的 SKILL.md 文件。 diff --git a/.agents/skills/citation-verification/scripts/api-clients.py b/.agents/skills/citation-verification/scripts/api-clients.py new file mode 100644 index 0000000..6528ad6 --- /dev/null +++ b/.agents/skills/citation-verification/scripts/api-clients.py @@ -0,0 +1,542 @@ +#!/usr/bin/env python3 +""" +API Clients for Citation Verification + +提供三个主要 API 客户端: +1. CrossRefClient - DOI 验证 +2. ArXivClient - arXiv 论文验证 +3. SemanticScholarClient - 通用学术搜索 + +每个客户端都包含: +- 错误处理 +- 重试机制 +- 速率限制 +- 结果标准化 +""" + +import time +import requests +from typing import Dict, List, Optional +from abc import ABC, abstractmethod + + +class RateLimiter: + """速率限制器""" + + def __init__(self, calls_per_minute: int): + self.calls_per_minute = calls_per_minute + self.last_call = 0 + self.min_interval = 60.0 / calls_per_minute + + def wait_if_needed(self): + """如果需要,等待以满足速率限制""" + elapsed = time.time() - self.last_call + if elapsed < self.min_interval: + time.sleep(self.min_interval - elapsed) + self.last_call = time.time() + + +class APIClient(ABC): + """API 客户端基类""" + + def __init__(self, rate_limit: int = 20): + """ + Args: + rate_limit: 每分钟最大请求数 + """ + self.rate_limiter = RateLimiter(rate_limit) + + @abstractmethod + def search(self, **kwargs) -> Optional[Dict]: + """搜索论文""" + pass + + def _retry_request(self, func, max_retries: int = 3): + """带重试的请求""" + for i in range(max_retries): + try: + self.rate_limiter.wait_if_needed() + return func() + except requests.exceptions.RequestException as e: + if i == max_retries - 1: + raise + time.sleep(2 ** i) # 指数退避 + return None + + +class CrossRefClient(APIClient): + """CrossRef API 客户端 + + 用于通过 DOI 验证论文信息 + + API 文档: https://api.crossref.org/ + """ + + def __init__(self, rate_limit: int = 50): + """ + Args: + rate_limit: 每分钟最大请求数 (CrossRef 限制较宽松) + """ + super().__init__(rate_limit) + self.base_url = "https://api.crossref.org" + + def search_by_doi(self, doi: str) -> Optional[Dict]: + """通过 DOI 搜索论文 + + Args: + doi: DOI 标识符 (如 10.1038/nature12345) + + Returns: + 标准化的论文信息字典,如果未找到则返回 None + """ + def request(): + url = f"{self.base_url}/works/{doi}" + response = requests.get(url, timeout=10) + response.raise_for_status() + return response.json() + + try: + data = self._retry_request(request) + if data and 'message' in data: + return self._normalize_result(data['message']) + return None + except Exception as e: + print(f"CrossRef API 错误: {e}") + return None + + def search(self, doi: str = None, **kwargs) -> Optional[Dict]: + """搜索论文 (统一接口)""" + if doi: + return self.search_by_doi(doi) + return None + + def _normalize_result(self, data: Dict) -> Dict: + """标准化 CrossRef 返回结果""" + # 提取标题 + title = data.get('title', [''])[0] if 'title' in data else '' + + # 提取作者 + authors = [] + if 'author' in data: + for author in data['author']: + given = author.get('given', '') + family = author.get('family', '') + if given and family: + authors.append(f"{given} {family}") + elif family: + authors.append(family) + + # 提取年份 + year = None + if 'published' in data: + date_parts = data['published'].get('date-parts', [[]])[0] + if date_parts: + year = date_parts[0] + elif 'created' in data: + date_parts = data['created'].get('date-parts', [[]])[0] + if date_parts: + year = date_parts[0] + + # 提取期刊/会议名称 + venue = '' + if 'container-title' in data: + venue = data['container-title'][0] if data['container-title'] else '' + + return { + 'title': title, + 'authors': authors, + 'year': year, + 'venue': venue, + 'doi': data.get('DOI', ''), + 'type': data.get('type', ''), + 'source': 'crossref' + } + + def get_bibtex(self, doi: str) -> Optional[str]: + """通过 DOI 获取 BibTeX + + Args: + doi: DOI 标识符 + + Returns: + BibTeX 字符串,如果失败则返回 None + """ + def request(): + url = f"https://doi.org/{doi}" + headers = {"Accept": "application/x-bibtex"} + response = requests.get(url, headers=headers, timeout=10) + response.raise_for_status() + return response.text + + try: + return self._retry_request(request) + except Exception as e: + print(f"获取 BibTeX 失败: {e}") + return None + + +class ArXivClient(APIClient): + """arXiv API 客户端 + + 用于验证 arXiv 预印本论文 + + API 文档: https://info.arxiv.org/help/api/ + """ + + def __init__(self, rate_limit: int = 20): + """ + Args: + rate_limit: 每分钟最大请求数 + """ + super().__init__(rate_limit) + try: + import arxiv + self.arxiv = arxiv + except ImportError: + raise ImportError("需要安装 arxiv 库: pip install arxiv") + + def search_by_id(self, arxiv_id: str) -> Optional[Dict]: + """通过 arXiv ID 搜索论文 + + Args: + arxiv_id: arXiv 标识符 (如 2301.12345 或 cs/0703001) + + Returns: + 标准化的论文信息字典,如果未找到则返回 None + """ + def request(): + search = self.arxiv.Search(id_list=[arxiv_id]) + paper = next(search.results()) + return paper + + try: + self.rate_limiter.wait_if_needed() + paper = request() + return self._normalize_result(paper) + except StopIteration: + print(f"arXiv 论文未找到: {arxiv_id}") + return None + except Exception as e: + print(f"arXiv API 错误: {e}") + return None + + def search_by_title(self, title: str, max_results: int = 5) -> Optional[Dict]: + """通过标题搜索论文 + + Args: + title: 论文标题 + max_results: 最大返回结果数 + + Returns: + 标准化的论文信息字典(第一个结果),如果未找到则返回 None + """ + def request(): + search = self.arxiv.Search( + query=f'ti:"{title}"', + max_results=max_results, + sort_by=self.arxiv.SortCriterion.Relevance + ) + results = list(search.results()) + return results[0] if results else None + + try: + self.rate_limiter.wait_if_needed() + paper = request() + if paper: + return self._normalize_result(paper) + return None + except Exception as e: + print(f"arXiv API 错误: {e}") + return None + + def search(self, arxiv_id: str = None, title: str = None, **kwargs) -> Optional[Dict]: + """搜索论文 (统一接口)""" + if arxiv_id: + return self.search_by_id(arxiv_id) + elif title: + return self.search_by_title(title) + return None + + def _normalize_result(self, paper) -> Dict: + """标准化 arXiv 返回结果""" + # 提取 arXiv ID + arxiv_id = paper.entry_id.split('/')[-1] + + return { + 'title': paper.title, + 'authors': [a.name for a in paper.authors], + 'year': paper.published.year, + 'venue': 'arXiv', + 'arxiv_id': arxiv_id, + 'doi': paper.doi if hasattr(paper, 'doi') else None, + 'abstract': paper.summary, + 'pdf_url': paper.pdf_url, + 'source': 'arxiv' + } + + @staticmethod + def extract_arxiv_id(text: str) -> Optional[str]: + """从文本中提取 arXiv ID + + Args: + text: 包含 arXiv ID 的文本 + + Returns: + arXiv ID,如果未找到则返回 None + """ + import re + + # 匹配新格式: YYMM.NNNNN + match = re.search(r'\d{4}\.\d{4,5}', text) + if match: + return match.group() + + # 匹配旧格式: arch-ive/YYMMNNN + match = re.search(r'[a-z-]+/\d{7}', text) + if match: + return match.group() + + return None + + +class SemanticScholarClient(APIClient): + """Semantic Scholar API 客户端 + + 用于通用学术论文搜索和验证 + + API 文档: https://api.semanticscholar.org/api-docs/ + """ + + def __init__(self, rate_limit: int = 20): + """ + Args: + rate_limit: 每分钟最大请求数 (Semantic Scholar 限制: 100 requests/5min) + """ + super().__init__(rate_limit) + try: + from semanticscholar import SemanticScholar + self.sch = SemanticScholar() + except ImportError: + raise ImportError("需要安装 semanticscholar 库: pip install semanticscholar") + + def search_by_title(self, title: str, max_results: int = 5) -> Optional[Dict]: + """通过标题搜索论文 + + Args: + title: 论文标题 + max_results: 最大返回结果数 + + Returns: + 标准化的论文信息字典(第一个结果),如果未找到则返回 None + """ + try: + self.rate_limiter.wait_if_needed() + results = self.sch.search_paper(title, limit=max_results) + + if not results: + return None + + # 返回第一个结果 + paper = results[0] + return self._normalize_result(paper) + except Exception as e: + print(f"Semantic Scholar API 错误: {e}") + return None + + def search_by_doi(self, doi: str) -> Optional[Dict]: + """通过 DOI 搜索论文 + + Args: + doi: DOI 标识符 + + Returns: + 标准化的论文信息字典,如果未找到则返回 None + """ + try: + self.rate_limiter.wait_if_needed() + paper = self.sch.get_paper(f"DOI:{doi}") + if paper: + return self._normalize_result(paper) + return None + except Exception as e: + print(f"Semantic Scholar API 错误: {e}") + return None + + def search(self, title: str = None, doi: str = None, **kwargs) -> Optional[Dict]: + """搜索论文 (统一接口)""" + if doi: + return self.search_by_doi(doi) + elif title: + return self.search_by_title(title) + return None + + def _normalize_result(self, paper) -> Dict: + """标准化 Semantic Scholar 返回结果""" + # 提取作者 + authors = [] + if paper.authors: + authors = [a.name for a in paper.authors] + + # 提取外部 ID + external_ids = paper.externalIds if hasattr(paper, 'externalIds') else {} + doi = external_ids.get('DOI') if external_ids else None + arxiv_id = external_ids.get('ArXiv') if external_ids else None + + return { + 'title': paper.title, + 'authors': authors, + 'year': paper.year, + 'venue': paper.venue if hasattr(paper, 'venue') else '', + 'paperId': paper.paperId, + 'doi': doi, + 'arxiv_id': arxiv_id, + 'citationCount': paper.citationCount if hasattr(paper, 'citationCount') else 0, + 'abstract': paper.abstract if hasattr(paper, 'abstract') else '', + 'source': 'semantic_scholar' + } + + +class CitationAPIManager: + """统一的 API 管理器 + + 协调三个 API 客户端,实现智能的 API 选择策略 + """ + + def __init__(self): + """初始化所有 API 客户端""" + self.crossref = None + self.arxiv = None + self.semantic_scholar = None + + # 尝试初始化各个客户端 + try: + self.crossref = CrossRefClient() + except Exception as e: + print(f"警告: CrossRef 客户端初始化失败: {e}") + + try: + self.arxiv = ArXivClient() + except Exception as e: + print(f"警告: arXiv 客户端初始化失败: {e}") + + try: + self.semantic_scholar = SemanticScholarClient() + except Exception as e: + print(f"警告: Semantic Scholar 客户端初始化失败: {e}") + + def verify_citation(self, citation_info: Dict) -> tuple[bool, Optional[str], Optional[Dict]]: + """验证引用 + + 实现 API 选择策略: + 1. DOI 优先 → CrossRef + 2. arXiv ID → arXiv + 3. 标题搜索 → Semantic Scholar + + Args: + citation_info: 引用信息字典,可能包含 doi, arxiv_id, title, authors 等字段 + + Returns: + (exists, api_source, api_data) + - exists: 论文是否存在 + - api_source: 验证来源 ('crossref', 'arxiv', 'semantic_scholar') + - api_data: API 返回的标准化数据 + """ + # 策略 1: DOI 优先 + if 'doi' in citation_info and self.crossref: + data = self.crossref.search_by_doi(citation_info['doi']) + if data: + return True, 'crossref', data + + # 策略 2: arXiv ID + arxiv_id = citation_info.get('arxiv_id') + if not arxiv_id and 'note' in citation_info: + # 尝试从 note 字段提取 arXiv ID + arxiv_id = ArXivClient.extract_arxiv_id(citation_info['note']) + + if arxiv_id and self.arxiv: + data = self.arxiv.search_by_id(arxiv_id) + if data: + return True, 'arxiv', data + + # 策略 3: 通用搜索 (Semantic Scholar) + if 'title' in citation_info and self.semantic_scholar: + data = self.semantic_scholar.search_by_title(citation_info['title']) + if data: + return True, 'semantic_scholar', data + + return False, None, None + + def get_bibtex(self, doi: str) -> Optional[str]: + """通过 DOI 获取 BibTeX + + Args: + doi: DOI 标识符 + + Returns: + BibTeX 字符串,如果失败则返回 None + """ + if self.crossref: + return self.crossref.get_bibtex(doi) + return None + + +# ============================================================================ +# 使用示例 +# ============================================================================ + +if __name__ == '__main__': + # 示例 1: 使用 CrossRef 客户端 + print("示例 1: CrossRef 客户端") + print("-" * 60) + crossref = CrossRefClient() + result = crossref.search_by_doi("10.48550/arXiv.1706.03762") + if result: + print(f"标题: {result['title']}") + print(f"作者: {', '.join(result['authors'][:3])}") + print(f"年份: {result['year']}") + print() + + # 示例 2: 使用 arXiv 客户端 + print("示例 2: arXiv 客户端") + print("-" * 60) + try: + arxiv_client = ArXivClient() + result = arxiv_client.search_by_id("1706.03762") + if result: + print(f"标题: {result['title']}") + print(f"作者: {', '.join(result['authors'][:3])}") + print(f"年份: {result['year']}") + except ImportError as e: + print(f"跳过: {e}") + print() + + # 示例 3: 使用 Semantic Scholar 客户端 + print("示例 3: Semantic Scholar 客户端") + print("-" * 60) + try: + ss_client = SemanticScholarClient() + result = ss_client.search_by_title("Attention is All You Need") + if result: + print(f"标题: {result['title']}") + print(f"作者: {', '.join(result['authors'][:3])}") + print(f"年份: {result['year']}") + print(f"引用数: {result['citationCount']}") + except ImportError as e: + print(f"跳过: {e}") + print() + + # 示例 4: 使用统一管理器 + print("示例 4: 统一 API 管理器") + print("-" * 60) + manager = CitationAPIManager() + citation_info = { + 'title': 'Attention is All You Need', + 'authors': ['Vaswani', 'Shazeer'], + 'year': '2017' + } + exists, source, data = manager.verify_citation(citation_info) + if exists: + print(f"验证成功!") + print(f"来源: {source}") + print(f"标题: {data['title']}") + print(f"年份: {data['year']}") diff --git a/.agents/skills/citation-verification/scripts/format-checker.py b/.agents/skills/citation-verification/scripts/format-checker.py new file mode 100644 index 0000000..c222702 --- /dev/null +++ b/.agents/skills/citation-verification/scripts/format-checker.py @@ -0,0 +1,605 @@ +#!/usr/bin/env python3 +""" +BibTeX and LaTeX Format Checker + +独立的格式检查工具,用于验证 BibTeX 和 LaTeX 引用格式。 + +功能: +1. BibTeX 格式检查 - 验证条目结构、必填字段、字段格式 +2. LaTeX 引用检查 - 提取引用、检查一致性 +3. 快速格式验证 - 无需 API 调用的快速检查 + +使用方法: + python format-checker.py references.bib + python format-checker.py paper.tex --check-latex + python format-checker.py references.bib --strict +""" + +import argparse +import sys +import re +from pathlib import Path +from typing import Dict, List, Tuple, Optional +from dataclasses import dataclass +from enum import Enum + +# 尝试导入 bibtexparser +try: + import bibtexparser + from bibtexparser.bparser import BibTexParser + BIBTEX_AVAILABLE = True +except ImportError: + print("警告: bibtexparser 未安装,BibTeX 解析功能受限") + print("运行: pip install bibtexparser") + BIBTEX_AVAILABLE = False + + +class ErrorLevel(Enum): + """错误级别""" + ERROR = "error" # 严重错误,必须修复 + WARNING = "warning" # 警告,建议修复 + INFO = "info" # 信息,可选修复 + + +@dataclass +class FormatError: + """格式错误数据类""" + level: ErrorLevel + location: str # 文件位置 (如 "entry:smith2020" 或 "line:42") + field: Optional[str] # 字段名 (如 "author", "year") + message: str # 错误描述 + suggestion: Optional[str] = None # 修复建议 + + +def parse_arguments(): + """解析命令行参数""" + parser = argparse.ArgumentParser( + description='检查 BibTeX 和 LaTeX 引用格式', + formatter_class=argparse.RawDescriptionHelpFormatter, + epilog=""" +示例: + %(prog)s references.bib + %(prog)s paper.tex --check-latex + %(prog)s references.bib --strict --output report.txt + %(prog)s references.bib --fix-common + """ + ) + + parser.add_argument( + 'input_file', + type=str, + help='BibTeX 文件(.bib)或 LaTeX 文件(.tex)' + ) + + parser.add_argument( + '--check-latex', + action='store_true', + help='检查 LaTeX 引用(需要提供 .tex 文件)' + ) + + parser.add_argument( + '--strict', + action='store_true', + help='严格模式 - 将警告视为错误' + ) + + parser.add_argument( + '--output', + type=str, + help='输出报告文件路径' + ) + + parser.add_argument( + '--fix-common', + action='store_true', + help='自动修复常见格式问题' + ) + + parser.add_argument( + '--verbose', + action='store_true', + help='显示详细信息' + ) + + parser.add_argument( + '--entry-type', + type=str, + help='只检查特定类型的条目(如 article, inproceedings)' + ) + + return parser.parse_args() + + +def load_bibtex_file(file_path: str) -> List[Dict]: + """加载 BibTeX 文件 + + Args: + file_path: BibTeX 文件路径 + + Returns: + BibTeX 条目列表 + + Raises: + FileNotFoundError: 文件不存在 + ValueError: 文件格式错误 + """ + if not BIBTEX_AVAILABLE: + raise ImportError("需要安装 bibtexparser: pip install bibtexparser") + + try: + with open(file_path, 'r', encoding='utf-8') as f: + parser = BibTexParser(common_strings=True) + bib_database = bibtexparser.load(f, parser) + return bib_database.entries + except FileNotFoundError: + raise FileNotFoundError(f"文件不存在: {file_path}") + except Exception as e: + raise ValueError(f"无法解析 BibTeX 文件: {e}") + + +def load_latex_file(file_path: str) -> str: + """加载 LaTeX 文件 + + Args: + file_path: LaTeX 文件路径 + + Returns: + 文件内容 + + Raises: + FileNotFoundError: 文件不存在 + """ + try: + with open(file_path, 'r', encoding='utf-8') as f: + return f.read() + except FileNotFoundError: + raise FileNotFoundError(f"文件不存在: {file_path}") + except Exception as e: + raise ValueError(f"无法读取 LaTeX 文件: {e}") + + +# ============================================================================ +# BibTeX 格式检查函数 +# ============================================================================ + +def get_required_fields(entry_type: str) -> List[str]: + """获取 BibTeX 条目类型的必填字段 + + Args: + entry_type: 条目类型 (如 'article', 'inproceedings') + + Returns: + 必填字段列表 + """ + required_fields = { + 'article': ['author', 'title', 'journal', 'year'], + 'inproceedings': ['author', 'title', 'booktitle', 'year'], + 'book': ['title', 'publisher', 'year'], + 'incollection': ['author', 'title', 'booktitle', 'publisher', 'year'], + 'inbook': ['author', 'title', 'chapter', 'publisher', 'year'], + 'proceedings': ['title', 'year'], + 'phdthesis': ['author', 'title', 'school', 'year'], + 'mastersthesis': ['author', 'title', 'school', 'year'], + 'techreport': ['author', 'title', 'institution', 'year'], + 'manual': ['title'], + 'misc': ['title'], + 'unpublished': ['author', 'title', 'note'], + } + return required_fields.get(entry_type.lower(), ['title']) + + +def get_optional_fields(entry_type: str) -> List[str]: + """获取 BibTeX 条目类型的可选字段 + + Args: + entry_type: 条目类型 + + Returns: + 可选字段列表 + """ + optional_fields = { + 'article': ['volume', 'number', 'pages', 'month', 'doi', 'url'], + 'inproceedings': ['editor', 'volume', 'series', 'pages', 'address', + 'month', 'organization', 'publisher', 'doi', 'url'], + 'book': ['author', 'editor', 'volume', 'series', 'address', + 'edition', 'month', 'isbn', 'doi', 'url'], + } + return optional_fields.get(entry_type.lower(), []) + + +def check_entry_structure(entry: Dict) -> List[FormatError]: + """检查 BibTeX 条目基本结构 + + Args: + entry: BibTeX 条目字典 + + Returns: + 错误列表 + """ + errors = [] + + # 检查条目类型 + if 'ENTRYTYPE' not in entry: + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"entry:{entry.get('ID', 'unknown')}", + field='ENTRYTYPE', + message="缺少条目类型", + suggestion="添加条目类型,如 @article, @inproceedings" + )) + return errors + + # 检查 ID + if 'ID' not in entry or not entry['ID'].strip(): + errors.append(FormatError( + level=ErrorLevel.ERROR, + location="entry:unknown", + field='ID', + message="缺少 citation key", + suggestion="添加唯一的 citation key" + )) + + # 检查必填字段 + entry_type = entry.get('ENTRYTYPE', '') + required = get_required_fields(entry_type) + for field in required: + if field not in entry or not entry[field].strip(): + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"entry:{entry.get('ID', 'unknown')}", + field=field, + message=f"缺少必填字段: {field}", + suggestion=f"添加 {field} 字段" + )) + + return errors + + +def check_field_formats(entry: Dict) -> List[FormatError]: + """检查字段格式 + + Args: + entry: BibTeX 条目字典 + + Returns: + 错误列表 + """ + errors = [] + entry_id = entry.get('ID', 'unknown') + + # 年份格式检查 + if 'year' in entry: + year = entry['year'].strip() + if not year.isdigit(): + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"entry:{entry_id}", + field='year', + message=f"年份格式错误: {year} (应为4位数字)", + suggestion="使用4位数字年份,如 2023" + )) + elif len(year) != 4: + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"entry:{entry_id}", + field='year', + message=f"年份格式错误: {year} (应为4位数字)", + suggestion="使用4位数字年份,如 2023" + )) + else: + year_int = int(year) + if year_int < 1900 or year_int > 2030: + errors.append(FormatError( + level=ErrorLevel.WARNING, + location=f"entry:{entry_id}", + field='year', + message=f"年份超出合理范围: {year}", + suggestion="检查年份是否正确" + )) + + # DOI 格式检查 + if 'doi' in entry: + doi = entry['doi'].strip() + if not doi.startswith('10.'): + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"entry:{entry_id}", + field='doi', + message=f"DOI 格式错误: {doi}", + suggestion="DOI 应以 '10.' 开头,如 10.1038/nature12345" + )) + # 检查是否包含 URL 前缀 + if 'doi.org' in doi or 'dx.doi.org' in doi: + errors.append(FormatError( + level=ErrorLevel.WARNING, + location=f"entry:{entry_id}", + field='doi', + message=f"DOI 包含 URL 前缀: {doi}", + suggestion="只保留 DOI 本身,移除 https://doi.org/ 前缀" + )) + + # 作者名格式检查 + if 'author' in entry: + author = entry['author'].strip() + # 检查是否为空 + if not author: + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"entry:{entry_id}", + field='author', + message="作者字段为空", + suggestion="添加作者信息" + )) + # 检查格式一致性 + elif ' and ' in author: + authors = author.split(' and ') + formats = [] + for a in authors: + if ',' in a: + formats.append('last_first') # "Last, First" + else: + formats.append('first_last') # "First Last" + + if len(set(formats)) > 1: + errors.append(FormatError( + level=ErrorLevel.WARNING, + location=f"entry:{entry_id}", + field='author', + message="作者名格式不一致", + suggestion="统一使用 'Last, First' 或 'First Last' 格式" + )) + + # 页码格式检查 + if 'pages' in entry: + pages = entry['pages'].strip() + # 检查是否使用了正确的分隔符 + if '-' in pages and '--' not in pages: + errors.append(FormatError( + level=ErrorLevel.INFO, + location=f"entry:{entry_id}", + field='pages', + message=f"页码使用单连字符: {pages}", + suggestion="建议使用双连字符 '--',如 123--145" + )) + + # URL 格式检查 + if 'url' in entry: + url = entry['url'].strip() + if not url.startswith(('http://', 'https://')): + errors.append(FormatError( + level=ErrorLevel.WARNING, + location=f"entry:{entry_id}", + field='url', + message=f"URL 缺少协议前缀: {url}", + suggestion="添加 http:// 或 https:// 前缀" + )) + + return errors + + +def check_consistency(entries: List[Dict]) -> List[FormatError]: + """检查条目间的一致性 + + Args: + entries: BibTeX 条目列表 + + Returns: + 错误列表 + """ + errors = [] + + # 检查重复的 citation key + ids = [e.get('ID', '') for e in entries] + duplicates = [id for id in ids if ids.count(id) > 1] + if duplicates: + for dup_id in set(duplicates): + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"entry:{dup_id}", + field='ID', + message=f"重复的 citation key: {dup_id}", + suggestion="使用唯一的 citation key" + )) + + # 检查作者名格式一致性 + author_formats = {} + for entry in entries: + if 'author' in entry and ' and ' in entry['author']: + entry_id = entry.get('ID', 'unknown') + authors = entry['author'].split(' and ') + for author in authors: + if ',' in author: + author_formats[entry_id] = 'last_first' + else: + author_formats[entry_id] = 'first_last' + break + + if len(set(author_formats.values())) > 1: + errors.append(FormatError( + level=ErrorLevel.WARNING, + location="global", + field='author', + message="不同条目使用了不同的作者名格式", + suggestion="统一使用 'Last, First' 或 'First Last' 格式" + )) + + return errors + + +# ============================================================================ +# LaTeX 引用检查函数 +# ============================================================================ + +def extract_latex_citations(tex_content: str) -> List[str]: + """从 LaTeX 文件中提取引用 + + Args: + tex_content: LaTeX 文件内容 + + Returns: + 引用 key 列表 + """ + # 匹配 \cite{...} 命令 + cite_pattern = r'\\cite(?:\[[^\]]*\])?(?:\[[^\]]*\])?\{([^}]+)\}' + citations = re.findall(cite_pattern, tex_content) + + # 展开多个引用 + all_keys = [] + for cite in citations: + keys = [k.strip() for k in cite.split(',')] + all_keys.extend(keys) + + return list(set(all_keys)) # 去重 + + +def check_latex_consistency(tex_keys: List[str], bib_keys: List[str]) -> List[FormatError]: + """检查 LaTeX 引用与 BibTeX 的一致性 + + Args: + tex_keys: LaTeX 中的引用 key 列表 + bib_keys: BibTeX 中的 key 列表 + + Returns: + 错误列表 + """ + errors = [] + + tex_set = set(tex_keys) + bib_set = set(bib_keys) + + # 未定义的引用 + undefined = tex_set - bib_set + if undefined: + for key in sorted(undefined): + errors.append(FormatError( + level=ErrorLevel.ERROR, + location=f"latex:cite", + field=key, + message=f"未定义的引用: {key}", + suggestion=f"在 BibTeX 文件中添加 {key} 条目" + )) + + # 未使用的引用 + unused = bib_set - tex_set + if unused: + for key in sorted(unused): + errors.append(FormatError( + level=ErrorLevel.WARNING, + location=f"bibtex:entry", + field=key, + message=f"未使用的引用: {key}", + suggestion=f"在 LaTeX 文件中引用 {key} 或从 BibTeX 中删除" + )) + + return errors + + +# ============================================================================ +# 报告生成函数 +# ============================================================================ + +def print_errors(errors: List[FormatError], verbose: bool = False): + """打印错误列表 + + Args: + errors: 错误列表 + verbose: 是否显示详细信息 + """ + if not errors: + print("✅ 未发现格式错误") + return + + # 按级别分组 + errors_by_level = { + ErrorLevel.ERROR: [], + ErrorLevel.WARNING: [], + ErrorLevel.INFO: [] + } + + for error in errors: + errors_by_level[error.level].append(error) + + # 打印统计 + print("\n" + "="*60) + print("格式检查结果") + print("="*60) + print(f"❌ 错误: {len(errors_by_level[ErrorLevel.ERROR])}") + print(f"⚠️ 警告: {len(errors_by_level[ErrorLevel.WARNING])}") + print(f"ℹ️ 信息: {len(errors_by_level[ErrorLevel.INFO])}") + print("="*60) + + # 打印详细错误 + for level in [ErrorLevel.ERROR, ErrorLevel.WARNING, ErrorLevel.INFO]: + level_errors = errors_by_level[level] + if not level_errors: + continue + + level_symbol = { + ErrorLevel.ERROR: "❌", + ErrorLevel.WARNING: "⚠️", + ErrorLevel.INFO: "ℹ️" + }[level] + + print(f"\n{level_symbol} {level.value.upper()} ({len(level_errors)}):\n") + + for error in level_errors: + print(f" [{error.location}]", end="") + if error.field: + print(f" {error.field}:", end="") + print(f" {error.message}") + + if verbose and error.suggestion: + print(f" 💡 建议: {error.suggestion}") + print() + + +def generate_report(errors: List[FormatError], output_file: str): + """生成文本格式的检查报告 + + Args: + errors: 错误列表 + output_file: 输出文件路径 + """ + # 按级别分组 + errors_by_level = { + ErrorLevel.ERROR: [], + ErrorLevel.WARNING: [], + ErrorLevel.INFO: [] + } + + for error in errors: + errors_by_level[error.level].append(error) + + with open(output_file, 'w', encoding='utf-8') as f: + f.write("# BibTeX/LaTeX 格式检查报告\n\n") + + # 总体统计 + f.write("## 总体统计\n\n") + f.write(f"- **错误**: {len(errors_by_level[ErrorLevel.ERROR])}\n") + f.write(f"- **警告**: {len(errors_by_level[ErrorLevel.WARNING])}\n") + f.write(f"- **信息**: {len(errors_by_level[ErrorLevel.INFO])}\n\n") + + # 详细错误 + for level in [ErrorLevel.ERROR, ErrorLevel.WARNING, ErrorLevel.INFO]: + level_errors = errors_by_level[level] + if not level_errors: + continue + + level_name = { + ErrorLevel.ERROR: "错误", + ErrorLevel.WARNING: "警告", + ErrorLevel.INFO: "信息" + }[level] + + f.write(f"## {level_name} ({len(level_errors)})\n\n") + + for error in level_errors: + f.write(f"### [{error.location}]") + if error.field: + f.write(f" {error.field}") + f.write("\n\n") + f.write(f"**问题**: {error.message}\n\n") + if error.suggestion: + f.write(f"**建议**: {error.suggestion}\n\n") + + print(f"\n报告已保存到: {output_file}") diff --git a/.agents/skills/citation-verification/scripts/verify-citations.py b/.agents/skills/citation-verification/scripts/verify-citations.py new file mode 100644 index 0000000..ea092be --- /dev/null +++ b/.agents/skills/citation-verification/scripts/verify-citations.py @@ -0,0 +1,678 @@ +#!/usr/bin/env python3 +""" +Citation Verification Script + +四层验证机制: +1. Format Validation - BibTeX 格式检查 +2. Existence Verification - API 验证论文存在性 +3. Information Matching - 信息匹配(标题、作者、年份) +4. Content Validation - 综合评分和判定 + +使用方法: + python verify-citations.py references.bib + python verify-citations.py paper.tex --check-latex + python verify-citations.py references.bib --verbose --output report.md +""" + +import argparse +import sys +import json +from pathlib import Path +from typing import Dict, List, Optional, Tuple +from dataclasses import dataclass, asdict +import re +from difflib import SequenceMatcher + +# 尝试导入 bibtexparser +try: + import bibtexparser + from bibtexparser.bparser import BibTexParser +except ImportError: + print("错误: 需要安装 bibtexparser") + print("运行: pip install bibtexparser") + sys.exit(1) + +# 尝试导入 API 客户端库 +try: + from semanticscholar import SemanticScholar +except ImportError: + print("警告: semanticscholar 未安装,将跳过 Semantic Scholar 验证") + print("运行: pip install semanticscholar") + +try: + import arxiv +except ImportError: + print("警告: arxiv 未安装,将跳过 arXiv 验证") + print("运行: pip install arxiv") + +try: + import requests +except ImportError: + print("错误: 需要安装 requests") + print("运行: pip install requests") + sys.exit(1) + + +@dataclass +class VerificationResult: + """验证结果数据类""" + citation_key: str + status: str # verified, partial_match, low_match, failed, not_found + confidence: str # high_confidence, medium_confidence, low_confidence, no_confidence + match_score: float + format_errors: List[str] + api_source: Optional[str] # crossref, arxiv, semantic_scholar + message: str + + +def parse_arguments(): + """解析命令行参数""" + parser = argparse.ArgumentParser( + description='验证 BibTeX 引用的准确性和完整性', + formatter_class=argparse.RawDescriptionHelpFormatter, + epilog=""" +示例: + %(prog)s references.bib + %(prog)s paper.tex --check-latex + %(prog)s references.bib --verbose --output report.md + %(prog)s references.bib --api-only + """ + ) + + parser.add_argument( + 'input_file', + type=str, + help='BibTeX 文件(.bib)或 LaTeX 文件(.tex)' + ) + + parser.add_argument( + '--check-latex', + action='store_true', + help='检查 LaTeX 引用一致性(需要提供 .tex 文件)' + ) + + parser.add_argument( + '--verbose', + action='store_true', + help='显示详细验证信息' + ) + + parser.add_argument( + '--output', + type=str, + help='输出报告文件路径(Markdown 格式)' + ) + + parser.add_argument( + '--api-only', + action='store_true', + help='仅进行 API 验证,跳过格式检查' + ) + + parser.add_argument( + '--format-only', + action='store_true', + help='仅进行格式检查,跳过 API 验证' + ) + + parser.add_argument( + '--threshold', + type=float, + default=0.85, + help='匹配阈值(0.0-1.0),默认 0.85' + ) + + return parser.parse_args() + + +def load_bibtex(file_path: str) -> List[Dict]: + """加载 BibTeX 文件""" + try: + with open(file_path, 'r', encoding='utf-8') as f: + parser = BibTexParser(common_strings=True) + bib_database = bibtexparser.load(f, parser) + return bib_database.entries + except FileNotFoundError: + print(f"错误: 文件不存在: {file_path}") + sys.exit(1) + except Exception as e: + print(f"错误: 无法解析 BibTeX 文件: {e}") + sys.exit(1) + + +def extract_latex_citations(tex_file: str) -> List[str]: + """从 LaTeX 文件中提取引用""" + try: + with open(tex_file, 'r', encoding='utf-8') as f: + content = f.read() + + # 匹配 \cite{...} 命令 + cite_pattern = r'\\cite(?:\[[^\]]*\])?\{([^}]+)\}' + citations = re.findall(cite_pattern, content) + + # 展开多个引用 + all_keys = [] + for cite in citations: + keys = [k.strip() for k in cite.split(',')] + all_keys.extend(keys) + + return list(set(all_keys)) # 去重 + except FileNotFoundError: + print(f"错误: 文件不存在: {tex_file}") + sys.exit(1) + except Exception as e: + print(f"错误: 无法解析 LaTeX 文件: {e}") + sys.exit(1) + + +# ============================================================================ +# Layer 1: Format Validation (格式验证) +# ============================================================================ + +def get_required_fields(entry_type: str) -> List[str]: + """获取 BibTeX 条目类型的必填字段""" + required_fields = { + 'article': ['author', 'title', 'journal', 'year'], + 'inproceedings': ['author', 'title', 'booktitle', 'year'], + 'book': ['title', 'publisher', 'year'], + 'misc': ['title'], + 'phdthesis': ['author', 'title', 'school', 'year'], + 'mastersthesis': ['author', 'title', 'school', 'year'], + 'techreport': ['author', 'title', 'institution', 'year'], + } + return required_fields.get(entry_type.lower(), ['title']) + + +def check_bibtex_format(entry: Dict) -> List[str]: + """检查 BibTeX 条目格式 + + Returns: + 错误列表 + """ + errors = [] + + # 检查条目类型 + if 'ENTRYTYPE' not in entry: + errors.append("缺少条目类型") + return errors + + # 检查 ID + if 'ID' not in entry: + errors.append("缺少 citation key") + + # 检查必填字段 + entry_type = entry.get('ENTRYTYPE', '') + required = get_required_fields(entry_type) + for field in required: + if field not in entry or not entry[field].strip(): + errors.append(f"缺少必填字段: {field}") + + # 年份格式检查 + if 'year' in entry: + year = entry['year'].strip() + if not year.isdigit() or len(year) != 4: + errors.append(f"年份格式错误: {year}") + else: + year_int = int(year) + if year_int < 1900 or year_int > 2030: + errors.append(f"年份超出合理范围: {year}") + + # DOI 格式检查 + if 'doi' in entry: + doi = entry['doi'].strip() + if not doi.startswith('10.'): + errors.append(f"DOI 格式错误: {doi}") + + return errors + + +def check_citation_consistency(tex_keys: List[str], bib_keys: List[str]) -> Dict: + """检查 LaTeX 引用与 BibTeX 的一致性 + + Returns: + {'undefined': [...], 'unused': [...]} + """ + tex_set = set(tex_keys) + bib_set = set(bib_keys) + + return { + 'undefined': list(tex_set - bib_set), + 'unused': list(bib_set - tex_set) + } + + +# ============================================================================ +# Layer 2: Existence Verification (存在性验证) +# ============================================================================ + +def verify_with_crossref(doi: str) -> Optional[Dict]: + """通过 CrossRef API 验证 DOI""" + try: + url = f"https://api.crossref.org/works/{doi}" + response = requests.get(url, timeout=10) + if response.status_code == 200: + data = response.json() + return data.get('message') + return None + except Exception as e: + print(f"CrossRef API 错误: {e}") + return None + + +def verify_with_arxiv(arxiv_id: str) -> Optional[Dict]: + """通过 arXiv API 验证""" + try: + search = arxiv.Search(id_list=[arxiv_id]) + paper = next(search.results()) + return { + 'title': paper.title, + 'authors': [a.name for a in paper.authors], + 'year': paper.published.year, + 'arxiv_id': arxiv_id + } + except Exception as e: + print(f"arXiv API 错误: {e}") + return None + + +def verify_with_semantic_scholar(title: str, authors: Optional[List[str]] = None) -> Optional[Dict]: + """通过 Semantic Scholar API 验证""" + try: + sch = SemanticScholar() + results = sch.search_paper(title, limit=5) + + if not results: + return None + + # 返回第一个结果 + paper = results[0] + return { + 'title': paper.title, + 'authors': [a.name for a in paper.authors] if paper.authors else [], + 'year': paper.year, + 'paperId': paper.paperId + } + except Exception as e: + print(f"Semantic Scholar API 错误: {e}") + return None + + +def verify_existence(entry: Dict) -> Tuple[bool, Optional[str], Optional[Dict]]: + """验证论文存在性 + + Returns: + (exists, api_source, api_data) + """ + # 策略 1: DOI 优先 + if 'doi' in entry: + data = verify_with_crossref(entry['doi']) + if data: + return True, 'crossref', data + + # 策略 2: arXiv ID + if 'eprint' in entry or 'arxiv' in entry.get('note', '').lower(): + arxiv_id = entry.get('eprint', '') + if not arxiv_id: + # 尝试从 note 中提取 + match = re.search(r'arXiv:(\d{4}\.\d{4,5})', entry.get('note', '')) + if match: + arxiv_id = match.group(1) + + if arxiv_id: + data = verify_with_arxiv(arxiv_id) + if data: + return True, 'arxiv', data + + # 策略 3: 通用搜索 + if 'title' in entry: + authors = entry.get('author', '').split(' and ') if 'author' in entry else None + data = verify_with_semantic_scholar(entry['title'], authors) + if data: + return True, 'semantic_scholar', data + + return False, None, None + + +# ============================================================================ +# Layer 3 & 4: Information Matching & Content Validation (信息匹配和内容验证) +# ============================================================================ + +def normalize_text(text: str) -> str: + """标准化文本用于匹配""" + text = text.lower() + text = re.sub(r'[^\w\s]', '', text) + return ' '.join(text.split()) + + +def match_title(title1: str, title2: str, threshold: float = 0.85) -> Dict: + """标题匹配""" + t1 = normalize_text(title1) + t2 = normalize_text(title2) + + ratio = SequenceMatcher(None, t1, t2).ratio() + + return { + 'match': ratio >= threshold, + 'similarity': ratio + } + + +def normalize_author_name(name: str) -> str: + """标准化作者名""" + parts = name.replace(',', '').split() + return ' '.join(sorted(parts)).lower() + + +def match_authors(authors1: List[str], authors2: List[str], threshold: float = 0.7) -> Dict: + """作者匹配""" + names1 = [normalize_author_name(a) for a in authors1] + names2 = [normalize_author_name(a) for a in authors2] + + set1 = set(names1) + set2 = set(names2) + intersection = len(set1 & set2) + union = len(set1 | set2) + + if union == 0: + return {'match': False, 'similarity': 0} + + ratio = intersection / union + + return { + 'match': ratio >= threshold, + 'similarity': ratio + } + + +def match_year(year1: str, year2: int, tolerance: int = 1) -> Dict: + """年份匹配""" + try: + y1 = int(year1) + y2 = int(year2) + diff = abs(y1 - y2) + return { + 'match': diff <= tolerance, + 'difference': diff + } + except (ValueError, TypeError): + return {'match': False, 'difference': None} + + +def calculate_match_score(entry: Dict, api_data: Dict, threshold: float) -> float: + """计算综合匹配分数""" + scores = {} + weights = { + 'title': 0.4, + 'authors': 0.3, + 'year': 0.2, + 'venue': 0.1 + } + + # 标题匹配 + if 'title' in entry and 'title' in api_data: + result = match_title(entry['title'], api_data['title'], threshold) + scores['title'] = result['similarity'] + + # 作者匹配 + if 'author' in entry and 'authors' in api_data: + entry_authors = entry['author'].split(' and ') + result = match_authors(entry_authors, api_data['authors']) + scores['authors'] = result['similarity'] + + # 年份匹配 + if 'year' in entry and 'year' in api_data: + result = match_year(entry['year'], api_data['year']) + scores['year'] = 1.0 if result['match'] else 0.0 + + # 计算加权总分 + total_score = 0 + total_weight = 0 + for key, weight in weights.items(): + if key in scores: + total_score += scores[key] * weight + total_weight += weight + + if total_weight == 0: + return 0 + + return total_score / total_weight + + +def judge_verification_result(match_score: float) -> Dict: + """判定验证结果""" + if match_score >= 0.9: + return { + 'status': 'verified', + 'level': 'high_confidence', + 'message': '✅ 验证通过 - 信息完全匹配' + } + elif match_score >= 0.7: + return { + 'status': 'partial_match', + 'level': 'medium_confidence', + 'message': '⚠️ 部分匹配 - 信息有轻微差异,建议人工确认' + } + elif match_score >= 0.5: + return { + 'status': 'low_match', + 'level': 'low_confidence', + 'message': '❌ 匹配度低 - 信息差异较大,需要人工验证' + } + else: + return { + 'status': 'failed', + 'level': 'no_confidence', + 'message': '❌ 验证失败 - 信息严重不匹配或论文不存在' + } + + +def verify_citation(entry: Dict, args) -> VerificationResult: + """完整的引用验证流程""" + citation_key = entry.get('ID', 'unknown') + + # Layer 1: 格式验证 + format_errors = [] + if not args.api_only: + format_errors = check_bibtex_format(entry) + + # Layer 2: 存在性验证 + if args.format_only: + return VerificationResult( + citation_key=citation_key, + status='format_checked', + confidence='n/a', + match_score=0.0, + format_errors=format_errors, + api_source=None, + message='仅格式检查' + ) + + exists, api_source, api_data = verify_existence(entry) + + if not exists: + return VerificationResult( + citation_key=citation_key, + status='not_found', + confidence='no_confidence', + match_score=0.0, + format_errors=format_errors, + api_source=None, + message='❌ 论文不存在 - 无法通过任何 API 验证' + ) + + # Layer 3 & 4: 信息匹配和内容验证 + match_score = calculate_match_score(entry, api_data, args.threshold) + judgment = judge_verification_result(match_score) + + return VerificationResult( + citation_key=citation_key, + status=judgment['status'], + confidence=judgment['level'], + match_score=match_score, + format_errors=format_errors, + api_source=api_source, + message=judgment['message'] + ) + + +# ============================================================================ +# Report Generation (报告生成) +# ============================================================================ + +def print_summary(results: List[VerificationResult], verbose: bool = False): + """打印验证摘要""" + total = len(results) + verified = sum(1 for r in results if r.status == 'verified') + partial = sum(1 for r in results if r.status == 'partial_match') + low = sum(1 for r in results if r.status == 'low_match') + failed = sum(1 for r in results if r.status in ['failed', 'not_found']) + + print("\n" + "="*60) + print("验证摘要") + print("="*60) + print(f"总引用数: {total}") + print(f"✅ 验证通过: {verified} ({verified/total*100:.1f}%)") + print(f"⚠️ 部分匹配: {partial} ({partial/total*100:.1f}%)") + print(f"❌ 匹配度低: {low} ({low/total*100:.1f}%)") + print(f"❌ 验证失败: {failed} ({failed/total*100:.1f}%)") + print("="*60) + + if verbose: + print("\n详细结果:\n") + for result in results: + print(f"[{result.citation_key}]") + print(f" 状态: {result.message}") + print(f" 匹配分数: {result.match_score:.2f}") + if result.api_source: + print(f" 验证源: {result.api_source}") + if result.format_errors: + print(f" 格式错误: {', '.join(result.format_errors)}") + print() + + +def generate_markdown_report(results: List[VerificationResult], output_file: str): + """生成 Markdown 格式的验证报告""" + total = len(results) + verified = sum(1 for r in results if r.status == 'verified') + partial = sum(1 for r in results if r.status == 'partial_match') + low = sum(1 for r in results if r.status == 'low_match') + failed = sum(1 for r in results if r.status in ['failed', 'not_found']) + + with open(output_file, 'w', encoding='utf-8') as f: + f.write("# Citation Verification Report\n\n") + + # 总体统计 + f.write("## 总体统计\n\n") + f.write(f"- **总引用数**: {total}\n") + f.write(f"- **✅ 验证通过**: {verified} ({verified/total*100:.1f}%)\n") + f.write(f"- **⚠️ 部分匹配**: {partial} ({partial/total*100:.1f}%)\n") + f.write(f"- **❌ 匹配度低**: {low} ({low/total*100:.1f}%)\n") + f.write(f"- **❌ 验证失败**: {failed} ({failed/total*100:.1f}%)\n\n") + + # 详细结果 + f.write("## 详细结果\n\n") + + # 按状态分组 + for status, emoji, title in [ + ('verified', '✅', '验证通过'), + ('partial_match', '⚠️', '部分匹配'), + ('low_match', '❌', '匹配度低'), + ('failed', '❌', '验证失败'), + ('not_found', '❌', '论文不存在') + ]: + status_results = [r for r in results if r.status == status] + if status_results: + f.write(f"### {emoji} {title} ({len(status_results)})\n\n") + for result in status_results: + f.write(f"#### `{result.citation_key}`\n\n") + f.write(f"- **状态**: {result.message}\n") + f.write(f"- **匹配分数**: {result.match_score:.2f}\n") + f.write(f"- **置信度**: {result.confidence}\n") + if result.api_source: + f.write(f"- **验证源**: {result.api_source}\n") + if result.format_errors: + f.write(f"- **格式错误**:\n") + for error in result.format_errors: + f.write(f" - {error}\n") + f.write("\n") + + # 建议操作 + f.write("## 建议操作\n\n") + if failed > 0: + f.write("### 需要修正的引用\n\n") + failed_results = [r for r in results if r.status in ['failed', 'not_found']] + for result in failed_results: + f.write(f"- `{result.citation_key}`: {result.message}\n") + f.write("\n") + + if partial > 0 or low > 0: + f.write("### 需要人工确认的引用\n\n") + check_results = [r for r in results if r.status in ['partial_match', 'low_match']] + for result in check_results: + f.write(f"- `{result.citation_key}`: {result.message}\n") + f.write("\n") + + print(f"\n报告已保存到: {output_file}") + + +# ============================================================================ +# Main Function (主函数) +# ============================================================================ + +def main(): + """主函数""" + args = parse_arguments() + + # 加载 BibTeX 文件 + print(f"正在加载 BibTeX 文件: {args.input_file}") + entries = load_bibtex(args.input_file) + print(f"找到 {len(entries)} 个引用条目") + + # LaTeX 一致性检查 + if args.check_latex: + tex_file = args.input_file.replace('.bib', '.tex') + if Path(tex_file).exists(): + print(f"\n正在检查 LaTeX 引用一致性: {tex_file}") + tex_keys = extract_latex_citations(tex_file) + bib_keys = [e['ID'] for e in entries] + consistency = check_citation_consistency(tex_keys, bib_keys) + + if consistency['undefined']: + print(f"⚠️ 未定义的引用 ({len(consistency['undefined'])}): {', '.join(consistency['undefined'])}") + if consistency['unused']: + print(f"⚠️ 未使用的引用 ({len(consistency['unused'])}): {', '.join(consistency['unused'])}") + if not consistency['undefined'] and not consistency['unused']: + print("✅ LaTeX 引用与 BibTeX 完全一致") + + # 验证每个引用 + print("\n开始验证引用...") + results = [] + for i, entry in enumerate(entries, 1): + citation_key = entry.get('ID', 'unknown') + print(f"[{i}/{len(entries)}] 验证 {citation_key}...", end=' ') + + result = verify_citation(entry, args) + results.append(result) + + # 简短状态输出 + if result.status == 'verified': + print("✅") + elif result.status == 'partial_match': + print("⚠️") + else: + print("❌") + + # 打印摘要 + print_summary(results, args.verbose) + + # 生成报告 + if args.output: + generate_markdown_report(results, args.output) + + # 返回退出码 + failed_count = sum(1 for r in results if r.status in ['failed', 'not_found']) + return 0 if failed_count == 0 else 1 + + +if __name__ == '__main__': + sys.exit(main()) |