Tools
ContextGem v0.27.0 Adds ID-Keyed Paragraph Lookups So Duplicate Contract Clauses Resolve to the Right Occurrence
The document-extraction library (1,982 stars) released v0.27.0 on 2026-08-13 adding `Document.get_paragraph_index()` and `get_sentence_index()`, keyed by unique ID rather than text equality — so duplicate clauses in a contract resolve to their specific occurrence, including after serialization round-trips that preserve those IDs. Lookups are O(1) average via lazily built caches validated against live document state, without affecting equality, serialization or cloning. It also re-pins litellm 1.96.2, openai 2.54.0 and genai-prices 0.1.2, with the maintainer explicitly citing breaking changes observed under unpinned versions.
Source
↳ Follow the thread