BibCheck

Bibliography verification tool for academic papers. Cross-checks BibTeX entries against DBLP, Crossref, and OpenAlex databases.

Unlike Google Scholar (which provides no stable API), this tool uses three free, open-access bibliographic sources to verify and correct bibliography information.

Features

• Verify bibliography entries against multiple databases (DBLP, Crossref, OpenAlex)
• Correct inconsistent author names and ordering
• Validate journal/conference names and publication years
• Preserve citation keys by default (optional rewriting to canonical format)
• Automatic LaTeX file updates when keys are rewritten
• Generate detailed reports (JSON and HTML) showing all changes
• Workspace batch processing via VS Code extension
• Strict publication-grade verification (exact title/author matching)

Installation

pip install -r requirements.txt

---

## CLI Usage

### Basic Commands

```bash
# Refine BibTeX file (preserves keys, in-place)
python bibtex_refiner.py refs.bib

# Save to new file
python bibtex_refiner.py refs.bib -o refs_refined.bib

# Rewrite citation keys
python bibtex_refiner.py refs.bib --rewrite-keys

# Generate report
python bibtex_refiner.py refs.bib --report report.json

# Verbose output for debugging
python bibtex_refiner.py refs.bib --verbose

HTML Report

python generate_report.py report.json report.html

VS Code Extension

Installation

1. Download .vsix from Releases
2. VS Code: Ctrl+Shift+P → "Install from VSIX"

Commands

• Refine Active BibTeX File: Right-click on .bib file
• Scan Workspace BibTeX Files: Click status bar button
• Scan Folder BibTeX Files: Right-click folder

Verification Algorithm

Entries are verified against strict publication standards:

1. Title: Must match exactly (100% after normalization)
2. Authors: Must be in correct order with exact names
3. Year: Must match exactly
4. DOI: Must match if present
5. Journal/Booktitle: Secondary confirmation

If any critical field (title/authors/year) mismatches, the entry is rejected. This conservative approach ensures bibliography quality suitable for submission.

Author handling: If input has "et al.", database list must have at least the same author prefix. If database has "et al.", it means the author list is abbreviated and remaining authors can be safely omitted.

Report Format

JSON Report

Contains:

• Metadata (generation time, threshold settings)
• Summary (total entries, changed count)
• Per-entry changes (before/after for each field)
• Confidence scores
• Low-confidence flagging

HTML Report

Displays:

• Summary statistics
• Color-coded confidence levels
• Before/after field comparisons
• Per-entry change notes