python-v2.16.0
run-llama/liteparsepython-v2.16.0Oct 12, 2026by github-actions[bot]
Upgrading?
AI Summary
This release focuses on refining text extraction accuracy for PDFs, improving OCR handling regarding clipping and layout preservation, and introduces concurrent OCR support for the WebAssembly build.
Key Highlights
- Improved PDF clipping logic to exclude hidden text and judge glyphs by ink box.
- Enhanced table detection by clipping block boxes to the page and extracting word boxes.
- Added concurrent OCR support with rolling scheduling for the WASM version.
- Fixed rotated text restacking and layout preservation issues.
- Improved handling of hyphens, list markers, and CJK characters.
New Features
- Support concurrent OCR with rolling scheduling in WASM
Full Release Notes
## What's Changed * fix: exclude native text hidden by rectangular PDF clips by @hexapode in https://github.com/run-llama/liteparse/pull/482 * fix: judge clipped glyphs by their ink box, not the loose box by @hexapode in https://github.com/run-llama/liteparse/pull/485 * fix(node): preserve page boundaries in text output by @tgolob in https://github.com/run-llama/liteparse/pull/431 * fix: refill bounded OCR workers as requests complete by @patelvivekdev in https://github.com/run-llama/liteparse/pull/466 * fix: reject invalid output formats in Node and Python bindings by @sanjayy0612 in https://github.com/run-llama/liteparse/pull/451 * fix: stop rotated-text restacking from compounding canvas offsets by @apeshimam-hebbia in https://github.com/run-llama/liteparse/pull/480 * fix: apply content filters in is_complex before the layout pass by @apeshimam-hebbia in https://github.com/run-llama/liteparse/pull/481 * fix: clip table block and cell boxes to the page by @catlan in https://github.com/run-llama/liteparse/pull/483 * fix: extract word boxes when extract_blocks is on by @catlan in https://github.com/run-llama/liteparse/pull/484 * docs(wasm): correct the OCR default and setup instructions by @Choudhry18 in https://github.com/run-llama/liteparse/pull/492 * fix: keep the dash of a hyphen at the end of a line by @BinaryBrain in https://github.com/run-llama/liteparse/pull/501 * fix: recognize dash bullets as list markers by @moah0911 in https://github.com/run-llama/liteparse/pull/491 * fix: don't treat short CJK OCR words as chart labels by @Diego-Romero-NTT in https://github.com/run-llama/liteparse/pull/498 * fix: count accented Latin vowels in the garbled-text check by @ed-thuando in https://github.com/run-llama/liteparse/pull/496 * feat(wasm): support concurrent OCR with rolling scheduling by @Choudhry18 in https://github.com/run-llama/liteparse/pull/502 ## New Contributors * @tgolob made their first contribution in https://github.com/run-llama/liteparse/pull/431 * @patelvivekdev made their first contribution in https://github.com/run-llama/liteparse/pull/466 * @sanjayy0612 made their first contribution in https://github.com/run-llama/liteparse/pull/451 * @apeshimam-hebbia made their first contribution in https://github.com/run-llama/liteparse/pull/480 * @catlan made their first contribution in https://github.com/run-llama/liteparse/pull/483 * @BinaryBrain made their first contribution in https://github.com/run-llama/liteparse/pull/501 * @moah0911 made their first contribution in https://github.com/run-llama/liteparse/pull/491 * @Diego-Romero-NTT made their first contribution in https://github.com/run-llama/liteparse/pull/498 * @ed-thuando made their first contribution in https://github.com/run-llama/liteparse/pull/496 **Full Changelog**: https://github.com/run-llama/liteparse/compare/crates-v2.15.1...python-v2.16.0
Related Repositories
farion1231/cc-switch★ 147,222
A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io
rustdesk/rustdesk★ 126,546
An open-source remote desktop application designed for self-hosting, as an alternative to TeamViewer.
tauri-apps/tauri★ 112,094
Build smaller, faster, and more secure desktop and mobile applications with a web frontend.
ruvnet/RuView★ 99,051
π RuView turns commodity WiFi signals into real-time spatial intelligence, vital sign monitoring, and presence detection — all without a single pixel of video.
oven-sh/bun★ 96,338
Incredibly fast JavaScript runtime, bundler, test runner, and package manager – all in one