doi2bib: DOI to BibTeX Citation Fetcher

A Python utility that reads DOI identifiers from a text file, retrieves BibTeX citations through DOI content negotiation, and saves the results to a .bib file.

The program supports both DOI-only files and files that contain local citation keys.

Features

• Accepts DOI URLs such as https://doi.org/10.xxxx/xxxxx.
• Accepts bare DOI identifiers such as 10.xxxx/xxxxx.
• Accepts doi:10.xxxx/xxxxx.
• Supports optional local citation keys.
• Retrieves BibTeX metadata through DOI content negotiation.
• Preserves the publisher's BibTeX key when no local key is provided.
• Replaces the publisher's BibTeX key when a local citation key is provided.
• Reports successful and failed DOI requests.
• Does not guess missing bibliographic information.
• Adds a short delay between requests.

Requirements

• Python 3.8 or newer
• requests

Install requests with:

pip install requests

Input Formats

The program supports three common formats.

1. Citation key + DOI

baker2015structjumper https://doi.org/10.1145/2702123.2702589
baltes2022sampling https://doi.org/10.1007/s10664-021-10072-8
bonaker2022nomon https://doi.org/10.1145/3491102.3517738

The citation key is preserved in the generated BibTeX.

For example:

@inproceedings{baker2015structjumper,
  ...
}

2. DOI URL only

https://doi.org/10.1145/2702123.2702589
https://doi.org/10.1007/s10664-021-10072-8
https://doi.org/10.1145/3491102.3517738

When no local key is provided, the publisher's BibTeX key is preserved.

3. Bare DOI only

10.1145/2702123.2702589
10.1007/s10664-021-10072-8
10.1145/3491102.3517738

The program automatically recognizes the DOI and retrieves the citation.

Mixed input

The formats can also be mixed in the same file:

baker2015structjumper https://doi.org/10.1145/2702123.2702589
https://doi.org/10.1007/s10664-021-10072-8
10.1145/3491102.3517738
doi:10.1038/s41467-024-53873-3

Usage

Place the Python script and input file in the same directory:

project/
├── get_citations.py
└── doi.txt

Run:

python get_citations.py

The program reads:

doi.txt

and creates:

citations.bib

The resulting directory will look like:

project/
├── get_citations.py
├── doi.txt
└── citations.bib

Citation Keys

If the input contains a local citation key:

baker2015structjumper https://doi.org/10.1145/2702123.2702589

the program replaces the publisher's BibTeX key with:

baker2015structjumper

For example:

@inproceedings{baker2015structjumper,
  ...
}

If the input contains only a DOI:

10.1145/2702123.2702589

the program does not modify the publisher's BibTeX key.

This behavior is useful when the DOI list comes from an existing bibliography where citation keys need to be preserved.

Output

Successfully retrieved citations are written to:

citations.bib

For example:

@inproceedings{baker2015structjumper,
  author = {...},
  title = {...},
  year = {...},
  pages = {...},
  doi = {10.1145/2702123.2702589}
}

The exact fields depend on the metadata returned by the DOI provider and publisher.

Failed Requests

The program reports failed requests during execution.

For example:

[3/20] example2024paper
  DOI: 10.1145/example
  ERROR: HTTP 403
  FAILED

At the end, it reports a summary:

============================================================
Successfully retrieved: 18
Failed:                2

Failed entries:
  example2024paper: 10.1145/example
  another2023paper: 10.1007/example

BibTeX written to: citations.bib

Failed entries should be checked manually through the publisher or DOI landing page.

How It Retrieves BibTeX

The program sends a request to the DOI resolver with the following HTTP header:

Accept: application/x-bibtex

This requests the citation in BibTeX format through DOI content negotiation.

The program then verifies that the response begins with a BibTeX entry such as:

@article{

or:

@inproceedings{

before writing it to the output file.

Limitations

DOI content negotiation does not guarantee that every publisher will return complete BibTeX metadata.

Possible issues include:

• HTTP 403 responses
• HTTP 503 responses
• Missing page ranges
• Missing numpages
• Missing articleno
• Incomplete metadata
• Publisher-specific BibTeX formatting
• DOI providers that do not return BibTeX

Therefore, a successful retrieval does not necessarily mean that every bibliographic field is complete.

In particular, page information should not be guessed. If a page range is missing, verify it using the publisher or another authoritative source before adding it to a research paper.

Verification

The generated citations.bib file should be reviewed before being used in a final paper.

Important fields to verify include:

• Authors
• Title
• Venue
• Publication year
• DOI
• Pages
• Article number
• Volume
• Issue

The program is a citation retrieval tool and is not a complete bibliography validation system.

Request Rate

The program waits one second between DOI requests:

time.sleep(1)

This reduces the rate of requests and is intended to avoid unnecessarily sending requests in rapid succession.