fetchbib 0.4.0

Resolve DOIs and search queries into formatted BibTeX entries.

fetchbib

A command-line tool to resolve DOIs and free-text search queries into formatted BibTeX entries. Powered by doi.org, arXiv, and the Crossref API.

Installation

pip

pip install fetchbib

uv

uv tool install fetchbib

Requires Python 3.9+.

Quick start

Fetch BibTeX by DOI (bare or full URL, including arXiv):

fbib 10.1073/pnas.2322823121
fbib https://doi.org/10.1073/pnas.2322823121
fbib 10.48550/arXiv.2410.21554

@article{DeVerna_2024,
  author = {DeVerna, Matthew R. and Yan, Harry Yaojun and Yang, Kai-Cheng and Menczer, Filippo},
  DOI = {10.1073/pnas.2322823121},
  ISSN = {1091-6490},
  journal = {Proceedings of the National Academy of Sciences},
  month = dec,
  number = {50},
  publisher = {Proceedings of the National Academy of Sciences},
  title = {Fact-checking information from large language models can decrease headline discernment},
  url = {http://dx.doi.org/10.1073/pnas.2322823121},
  volume = {121},
  year = {2024}
}

Search by free text by including text between quotes:

fbib "DeVerna Fact-checking information from large language models"

Note: Search works best with full paper titles (or partial titles with an author last name). Use -n to return multiple results if needed. See Specify number of free-text matches section below.

Usage

fbib [-h] [-f FILE] [-o OUTPUT] [-a] [-n MAX_RESULTS]
     [--config-email EMAIL] [--config-protect-titles]
     [--config-exclude-issn] [--config-exclude-doi]
     [inputs ...]

Flexible input

fbib accepts DOIs in any format — bare, full URL, or free-text search queries — and you can mix them freely. Inputs are comma-separated, so all of the following work:

# Multiple positional arguments
fbib 10.1609/icwsm.v5i1.14126 10.1093/jcmc/zmz022

# Comma-separated string
fbib "10.1609/icwsm.v5i1.14126, 10.1093/jcmc/zmz022"

# Full DOI URLs
fbib "https://doi.org/10.1609/icwsm.v5i1.14126, https://doi.org/10.1093/jcmc/zmz022"

# Mix DOIs, URLs, and search queries
fbib 10.1609/icwsm.v5i1.14126 "DeVerna Fact-checking information from large language models"

From a file (--file), each line is treated the same way — one entry per line, or comma-separated on a single line:

fbib --file dois.txt

Duplicate inputs are automatically removed.

Write to a file

Overwrite (default):

fbib --output refs.bib 10.1609/icwsm.v5i1.14126

Append to an existing .bib file:

fbib --append --output refs.bib 10.1093/jcmc/zmz022

Specify number of free-text matches

Free-text searches return 1 result by default. Use -n to control the limit (1-100):

# Get 3 results (see the correct result in position #2)
fbib -n 3 "Fact-checking information from large language models can decrease"

Configuration

Settings are stored in ~/.config/fetchbib/config.json and persist across sessions.

Email

Crossref gives better rate limits to requests that include a contact email. Set yours once:

fbib --config-email [e-mail]

Protect titles

BibTeX processors may change title capitalization. Enable title protection to wrap titles in double braces, preserving the original case:

fbib --config-protect-titles

This transforms title = {A {GPU} Study} into title = {{A GPU Study}}. Run again to toggle off.

Exclude ISSN

Remove the ISSN field from returned BibTeX entries:

fbib --config-exclude-issn

Run again to toggle off.

Exclude DOI

Remove the DOI field from returned BibTeX entries (the DOI is already included in the URL field):

fbib --config-exclude-doi

Run again to toggle off.

Development

Clone the repo and sync dependencies with uv:

git clone https://github.com/mr-devs/fetchbib.git
cd fetchbib
uv tool install -e . # OR pip install -e .

The -e flag (uv tool install -e .), it performs an editable installation. This means any changes you make to the local source code are immediately reflected when you run the command, without needing to reinstall.

Run unit tests:

uv run pytest

Run integration tests (hits live APIs):

uv run pytest -m integration