llm-arxiv

LLM plugin for loading arXiv paper text content using the arxiv API and PyMuPDF.

Installation

Install this plugin in the same environment as LLM.

llm install llm-arxiv

This plugin requires the arxiv and PyMuPDF packages.

Usage

This plugin adds support for the arxiv: fragment prefix. You can use it to load the full text content of a paper from its arXiv ID or URL.

# Load by ID (new format)
llm fragment arxiv:2310.06825

# Load by ID (old format)
llm fragment arxiv:hep-th/0101001

# Load by URL
llm fragment arxiv:https://arxiv.org/abs/2310.06825

You can pipe the output to LLM commands:

llm -f arxiv:2310.06825 -m claude-3-haiku "Summarize this paper"

Development

To set up this plugin locally, first checkout the code. Then create a new virtual environment:

cd llm-arxiv
python -m venv venv
source venv/bin/activate

Now install the dependencies and test dependencies:

# Install in editable mode with test dependencies
python -m pip install -e '.[test]'

To run the tests:

pytest