LexPredict/lexpredict-lexnlp

Extracts structured data from unstructured legal documents using pre-trained NLP models for domain-specific tasks like identifying monetary amounts, dates, courts, regulations, and clause types. Built with Python 3.8, it provides legal-aware tokenization that handles abbreviations (LLC., F.3d.) and includes pre-trained word embeddings, topic models, and document classifiers validated against hundreds of real contract examples. Integrates with ContraxSuite for document analysis workflows and includes extensible tools for building custom clustering and classification pipelines.

768 stars. No commits in the last 6 months.