Linked Claims Extractor

The Linked Claims Extractor is a tool that uses AI language models to extract claims from text, URLs, or PDFs. It supports multiple LLMs, including Anthropic and OpenAI, and allows users to define custom schemas for claim extraction.

Installation

To install the Linked Claims Extractor, run:

pip install linked-claims-extractor

You will need to set the ANTHROPIC_API_KEY environment variable if you are using the Anthropic model:

export ANTHROPIC_API_KEY=your_api_key_here

Usage

Extracting Claims from Text

from claim_extractor import ClaimExtractor

extractor = ClaimExtractor()
result = extractor.extract_claims('some text')

import pprint
pprint.pprint(result)

Extracting Claims from a URL

result = extractor.extract_claims_from_url('https://example.com')
pprint.pprint(result)

Development and Testing

To set up the development environment, run:

python -m venv .venv
source .venv/bin/activate
pip install -e .

To run tests, use:

pytest -s --pdb

Contributing

Contributions are welcome! Please open an issue or submit a pull request with your changes.

License

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