合并多个bibtex

总说

有时候，你写毕业论文之类的，你需要合并多篇论文中的bibtex到一个，从而可以使得你论文的所有bibtex的引用正常。
存在以下四种情况：

1. 标题相同，label不同 -> 二者都保持（这是因为，你直接复制正文时，不需要更改标签了。两种标签都可以用，都引用相同的论文）

2. 标题相同，label相同 -> 保留一个

3. 标题不同，label相同 -> 二者保留，记录log中，人为后续处理

4. 标题不同，label相同 -> 二者保留，记录到log中，人为后续处理

5. 其他情况（这是为了防止不小心略过的论文）-> 检查原本所有bib文件中的论文，进行更新

import os
import re
from collections import defaultdict

def read_bib_file(file_path):
    with open(file_path, 'r', encoding='utf-8') as file:
        content = file.read()
    entries = re.split(r'\n@', content)
    entries = ['@' + entry.strip() for entry in entries if entry.strip()]
    return entries

def extract_title(entry):
    title_match = re.search(r'title\s*=\s*{(.+?)}', entry, re.IGNORECASE)
    if title_match:
        return title_match.group(1).strip()
    return None

def extract_citation_key(entry):
    key_match = re.search(r'@\w+{(.+?),', entry)
    if key_match:
        return key_match.group(1).strip()
    return None

def combine_bib_files(file_paths):
    combined_entries = defaultdict(list)
    citation_key_entries = defaultdict(list)
    duplicates = []
    errors = []
    all_entries = []

    for file_path in file_paths:
        entries = read_bib_file(file_path)
        all_entries.extend(entries)  # Keep track of all entries for summary
        for entry in entries:
            title = extract_title(entry)
            citation_key = extract_citation_key(entry)
            if title and citation_key:
                combined_entries[title].append(entry)
                citation_key_entries[citation_key].append(entry)

    final_entries = []

    for title, entries in combined_entries.items():
        if len(entries) > 1:
            keys = set(extract_citation_key(entry) for entry in entries)
            if len(keys) > 1:
                duplicates.append((title, entries))
                final_entries.extend(entries)
            else:
                final_entries.append(entries[0])
        else:
            final_entries.append(entries[0])

    for key, entries in citation_key_entries.items():
        if len(entries) > 1:
            titles = set(extract_title(entry) for entry in entries)
            if len(titles) > 1:
                errors.append((key, entries))
    
    return final_entries, duplicates, errors, all_entries

def write_combined_bib_file(entries, output_file):
    with open(output_file, 'w', encoding='utf-8') as file:
        for entry in entries:
            file.write(entry + '\n\n')

def log_duplicates(duplicates, log_file):
    with open(log_file, 'w', encoding='utf-8') as file:
        for title, entries in duplicates:
            file.write(f"Title: {title}\n")
            for entry in entries:
                file.write(entry + '\n\n')
            file.write('-' * 80 + '\n')

def log_errors(errors, log_file):
    with open(log_file, 'w', encoding='utf-8') as file:
        for key, entries in errors:
            file.write(f"Citation Key: {key}\n")
            for entry in entries:
                file.write(entry + '\n\n')
            file.write('-' * 80 + '\n')

if __name__ == "__main__":
    # List of paths to your bib files
    bib_files = ['ref1.bib', 'ref2.bib', 'ref3.bib']

    # Combine entries from all files
    combined_entries, duplicates, errors, all_entries = combine_bib_files(bib_files)

    # Write the combined entries to a new file
    output_file = 'combined_ref.bib'
    write_combined_bib_file(combined_entries, output_file)
    
    # Log duplicates for manual merging
    duplicates_log_file = 'duplicates_log.txt'
    log_duplicates(duplicates, duplicates_log_file)

    # Log errors for manual checking
    error_log_file = 'error_log.txt'
    log_errors(errors, error_log_file)

    # Print summary information
    print(f"Combined bib file created: {output_file} with {len(combined_entries)} entries")
    print(f"Duplicates log created: {duplicates_log_file} with {len(duplicates)} entries")
    print(f"Error log created: {error_log_file} with {len(errors)} entries")

    for file in bib_files:
        entries = read_bib_file(file)
        print(f"{file}: {len(entries)} entries processed")
    print(f"Total unique entries combined: {len(combined_entries)}")
    print(f"Total entries processed: {len(all_entries)}")