LangExtract is a Python library that extracts structured information from text using language models. Its main parts are the annotation and resolver modules that turn model output into extractions aligned to the source text, chunking and prompting modules that prepare inputs, provider modules that talk to model backends, and an io module. Callers use the io module to load a CSV dataset of documents to process and to save annotated results as JSON Lines, then reload them later for review or visualization.
What the team is seeing
incident log
After processing documents written in Persian, Arabic or Chinese, the saved results show unreadable garbled characters on some computers, while English documents look fine. Loading the exported results or the input spreadsheet gives back different text than the original documents.
Acceptance checks 1
Check 1
Discussion is locked — log in and complete this problem to join the discussion and rate your experience.