BiblioAudit

BiblioAudit is a command-line tool that checks and verifies citations in BibTeX (.bib) files. It cross-checks references across more than 12 academic databases in parallel to confirm that entries exist, identify missing or broken DOIs, correct publication years and journal names, and locate legal open-access PDFs.

It also includes an automated re-check loop (--loop) that repairs metadata, generates a corrected BibTeX library via doi2bib, and verifies that all corrections pass validation.

---

Example Output

========================================================================
  BiblioAudit v2.2.1 - Reference Integrity & Verification System
========================================================================
  Cascading Triangulation Matrix: 12 Academic Databases

[*] Reading bibliography: References/sn-bibliography.bib
[*] Found 42 entries. Starting AUDIT...
[*] Processing 42 references (Concurrency limit: 6)...

  [VERIFIED]  (100%) Moskal2014          -> Crossref (DOI)
  [VERIFIED]  (100%) Tiwari2025          -> Crossref (DOI)
  [VERIFIED]  (100%) Spencer2021         -> Crossref (Search)
  [VERIFIED]  (100%) Tiwari_BiblioAudit  -> DataCite (DOI)
  [ATTENTION] ( 75%) Liu2021             -> Crossref (Search)
        └─ Year: BibTeX='2021' vs Record='2025'
  [NOT FOUND] (  0%) NEMA2018            -> Missing in all queried databases

============================================================
                      AUDIT SUMMARY
============================================================
  Total References Audited : 42
  Verified Clean (>90%)    :   35  ( 83.3%)
  Needs Attention          :    4  (  9.5%)
  Not Found                :    3  (  7.1%)
  Total Time Elapsed       : 12.59s
============================================================
[+] Audit CSV report saved to: References/sn-bibliography_report.csv
[+] Corrected BibTeX library saved to: References/sn-bibliography_fixed.bib

---

Supported Academic Databases

BiblioAudit queries the following scientific databases in parallel:

Database	Coverage and Specialization	Query Method
Crossref	Primary registry for journal articles, conference proceedings, and books	DOI, Title, Author
DataCite	Research datasets, software records, and Zenodo deposits	DOI, Title, Keyword
OpenAlex	Global index covering 250M+ scientific publications	DOI, Title, Author
Semantic Scholar	Multidisciplinary literature and computer science graph	DOI, Title, Author
arXiv	Preprints in physics, mathematics, and computer science	Title, Author
PubMed	Biomedical and clinical literature (NLM/NIH)	E-utilities Search
Europe PMC	Medical physics, clinical research, and European open access	REST Search, DOI
DOAJ	Peer-reviewed open access journals worldwide	Title Search
HAL Open Science	European multidisciplinary open archive (CNRS, INRIA, CEA)	Field Search
Zenodo	Open research code, software citations, and data archives	REST Records API
INSPIRE HEP	High-energy physics, astrophysics, and quantum field theory	Literature Search
NASA ADS	Astrophysics, astronomy, and planetary physics literature	Search API (optional token)
Unpaywall	Legal open-access PDF discovery	DOI lookup
OpenCitations	Citation metrics and citation network counts	COCI DOI index

---

Installation

From PyPI

pip install biblioaudit

From Source

1. Clone the repository:

   git clone https://github.com/Official-Satyam-Tiwari/Reference-Architect.git
   cd Reference-Architect
   

2. Install dependencies:

   pip install -r requirements.txt
   

Optional Local Package Install

To install in editable mode for local development:

pip install -e .

---

Usage

Run the tool from your terminal:

python biblioaudit.py [path-to-bib-file-or-folder] [options]

If you do not specify a path, BiblioAudit automatically checks files in the References/ directory.

Options

Option	Description	Default
-o, --output	File path for the CSV report	<filename>_audit_report.csv
--fix-bib	File path to export corrected BibTeX entries	None
--loop	Run automated feedback loop: re-audits the corrected file and reports convergence metrics	False
-r, --rate	Maximum number of concurrent requests	4
-v, --verbose	Print field-by-field differences for mismatched citations	False
--version	Show program version
-h, --help	Show command-line help message

---

Examples

1. Basic Check

Check a bibliography file and print progress to the console:

python biblioaudit.py References/sn-bibliography.bib

2. Automated Re-Check Loop

Audits the original file, repairs citations via doi2bib, immediately re-checks the repaired file, and prints a before-and-after convergence report:

python biblioaudit.py References/sn-bibliography.bib --loop

3. Generate Reports and Corrected BibTeX

Save a detailed CSV report and write official publisher BibTeX records:

python biblioaudit.py References/sn-bibliography.bib -o report.csv --fix-bib fixed.bib

4. Fast Processing with Detailed Field Differences

Increase the request rate and print field-by-field differences for mismatched citations:

python biblioaudit.py References/sn-bibliography.bib --rate 6 --verbose

---

Project Structure

Reference-Architect/
├── References/              # Directory for your .bib files (ignored by git)
│   └── sn-bibliography.bib
├── biblioaudit.py           # Main runner, engines, and verification logic
├── utility.py               # Text similarity, BibTeX parsing, and CSV exports
├── requirements.txt         # Package dependencies
├── pyproject.toml           # Packaging metadata
├── CITATION.cff             # Citation metadata for GitHub
├── LICENSE                  # MIT License
└── README.md                # Documentation

---

Citation

If you use BiblioAudit in your research, please cite:

Tiwari, S. (2025). BiblioAudit: Automated Citation Integrity & Verification Tool (Version 2.2.1) [Software]. Zenodo. https://doi.org/10.5281/zenodo.18155557

@software{Tiwari_BiblioAudit_2025,
  author = {Tiwari, Satyam},
  title = {{BiblioAudit: Automated Citation Integrity & Verification Tool}},
  version = {2.2.0},
  year = {2025},
  publisher = {Zenodo},
  doi = {10.5281/zenodo.18155557},
  url = {https://doi.org/10.5281/zenodo.18155557}
}

---

License

This project is licensed under the MIT License. See the LICENSE file for details.