bib-optimizer

Oh, sure, because who doesn't love manually cleaning up messy .bib files? bib_optimizer.py heroically steps in to remove those lazy, unused citations and reorder the survivors exactly as they appear in the .tex file—because, clearly, chaos is the default setting for bibliographies.

In layman's terms, it automates bibliography management by:

1. removing unused citations,
2. reordering the remaining ones to match their order of appearance in the .tex file.

Input Files:

• main.tex – The LaTeX source file.
• ref.bib – The original bibliography file.

These input files will remain unchanged.

Output File:

• ref_opt.bib – A placeholder filename for the newly generated, cleaned, and ordered bibliography file.

Installation

Install from PyPI with pip

Create and activate a virtual environment, then install the latest release from PyPI:

python -m venv .venv

# Linux/macOS
source .venv/bin/activate

# Windows
.venv\Scripts\activate

python -m pip install --upgrade bib_optimizer

Alternatively, install the command in an isolated environment with uv:

uv tool install bib_optimizer

After either installation method, the bibopt command is available in your terminal.

Install from source for development

This repository uses uv for project and dependency management.

1. Clone the repository

   git clone https://github.com/kwyip/bib_optimizer.git
   cd bib_optimizer
   

2. Install environment

   # This creates the virtual environment and installs all dependencies
   uv sync
   

3. Activate environment

   # Linux/MacOS
   source .venv/bin/activate
   
   # Windows
   .venv\Scripts\activate
   

🐍 This requires Python 3.8 or newer versions

Steps to Clean Your Bibliography

1. Prepare the input files (e.g., by downloading them from Overleaf).

2. Run the command to generate a new .bib file (for example, you may name it ref_opt.bib):

          bibopt main.tex ref.bib ref_opt.bib

3. Use the Cleaned Bibliography
   Replace ref.bib with ref_opt.bib in your LaTeX project.

Test

You may test the installation using the sample input files (sample_main.tex and sample_ref.bib) located in the test folder.

sample_main.tex and sample_ref.bib

A sample ref_opt.bib created after running bibopt sample_main.tex sample_ref.bib ref_opt.bib

Code-free use with an AI agent

Download SKILL.md, then upload these files to your AI agent:

• SKILL.md,
• the root .tex file,
• the source .bib file, and if any:
• every .tex file referenced through \input or \include.

https://github.com/user-attachments/assets/2b4e3224-2aee-44c1-974c-812dead72e2f

New feature (version 0.4.0)

If the main.tex calls inputs from other .tex (e.g., with \input{...}), the newly generated ref_opt.bib will preserve the order of appearances in the main.tex with each inputted .tex as well.
(The dependent .tex files need to be placed in the same directory as main.tex.)

New feature (version 0.4.1)

On top of version 0.4, skip any \input .tex file if not found.

New feature/Fix (version 0.4.2)

Added support for the \include{...} command in addition to \input{...}, fixing missing citations when a LaTeX project uses \include to split its content across files.

New feature/Fix (version 0.4.3)

Added Python 3.14 support and constrained bibtexparser to the compatible 1.x release series, fixing ModuleNotFoundError: No module named 'bibtexparser.bwriter' during fresh installations.

New feature/Fix (version 0.5.0)

Add code-free SKILL.md

♥ Lastly executed on Python 3.14 on 2026-09-29.