v2.132.0
tambo-ai/tambov2.132.0Oct 1, 2026by docling-ops[bot]
Upgrading?
AI Summary
An update to the PDF reading-order algorithm and a collection of fixes for parsing issues across various document formats including CSV, DOCX, Markdown, LaTeX, and Video.
Key Highlights
- Updated reading-order algorithm for PDFs
- Preservation of fidelity for external VLM output
- CSV validation and quote unescaping fixes
- Recognition of .gif files as images
New Features
- Updated reading-order algorithm
- GIF file recognition
- Improved CSV parsing (validation and unescaping)
- Enhanced DOCX bold detection
Full Release Notes
### Feature * **pdf, reading-order:** Updating the reading-order algoirthm ([#4276](https://github.com/docling-project/docling/issues/4276)) ([`42de1e2`](https://github.com/docling-project/docling/commit/42de1e20fe2783c0c30c5b3c0f9c60ab5eef9e16)) ### Fix * Preserve fidelity when mapping external VLM output (Chandra, dots, MinerU) ([#4470](https://github.com/docling-project/docling/issues/4470)) ([`20ba847`](https://github.com/docling-project/docling/commit/20ba84719e1c9e29f7ddfb731eca7ad00488180d)) * **csv:** Validate first-line dialect before accepting a sniff ([#4453](https://github.com/docling-project/docling/issues/4453)) ([`d6f0307`](https://github.com/docling-project/docling/commit/d6f03078ad364108df3e7e82e8f0dcc3fd7f39ea)) * **docx:** Exclude '<w:bCs>' from bold detection ([#4366](https://github.com/docling-project/docling/issues/4366)) ([`bf836f2`](https://github.com/docling-project/docling/commit/bf836f244d92dc001d68230a44a4afa23926df17)) * **image:** Recognize .gif files as images ([#4326](https://github.com/docling-project/docling/issues/4326)) ([`cd05bef`](https://github.com/docling-project/docling/commit/cd05bef2837c879cb636b5e8e61b03405bfdf7b8)) * **cli:** Skip Excel and PowerPoint lock files in directories ([#4324](https://github.com/docling-project/docling/issues/4324)) ([`6864fb3`](https://github.com/docling-project/docling/commit/6864fb314475b3aab8acb254e9c5f2c795fd5b13)) * **csv:** Unescape doubled quotes in quoted fields ([#4362](https://github.com/docling-project/docling/issues/4362)) ([`e53ce5d`](https://github.com/docling-project/docling/commit/e53ce5dfd2e7705dd06855eefa62e404ac59f0ac)) * **md:** Keep inline code spans inside GFM table cells ([#4327](https://github.com/docling-project/docling/issues/4327)) ([`ae8db9a`](https://github.com/docling-project/docling/commit/ae8db9a2f1379875f1a74ad7935f58a70b5183c9)) * **latex:** Run Tectonic without shell escape and skip TikZ that references outside files ([#4419](https://github.com/docling-project/docling/issues/4419)) ([`38b6fa0`](https://github.com/docling-project/docling/commit/38b6fa0a465d46fdacbbec333f50fa19c4f6b342)) * **html:** Validate every resolved address and scope fetch headers to the source origin ([#4420](https://github.com/docling-project/docling/issues/4420)) ([`5e46913`](https://github.com/docling-project/docling/commit/5e469137f275ffc443306a30d12a3a45bceb80fb)) * **video:** Limit ffmpeg input protocols and add subprocess timeouts ([#4421](https://github.com/docling-project/docling/issues/4421)) ([`efaf9ed`](https://github.com/docling-project/docling/commit/efaf9edd61363ef7239a49ba8e6b349ee219268f))
Related Repositories
openclaw/openclaw★ 392,749
The AI that really does things. Any OS. Any Platform. The lobster way. 🦞
nilbuild/developer-roadmap★ 370,126
Interactive roadmaps, guides and other educational content to help developers grow in their careers.
morluto/rea★ 363,548
Reverse engineer anything with agents, from app behavior down to native binaries.
anomalyco/opencode★ 220,031
The open source coding agent.
n8n-io/n8n★ 208,331
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.