v0.5.1: 25 MB document ceiling via streaming extraction
25 MB document ceiling via streaming extraction
- Raised the
contentdocument fetch ceiling 5,000,000 → 25,000,000 bytes, reaching the largest enacted bills in either published format (116th Congress H.R. 133: 6,790,482 bytes as Formatted Text, 9,812,888 as XML) (#56) - New incremental extractor streams and retains only the requested character window instead of buffering the whole body, held byte-identical to the existing extractor by a differential test
- Fixed: a body that fails mid-stream now maps to
document_fetch_failed; thedocument_too_largerecovery hint no longer points at a format that fails the same way