Starstruk Logo MarkStarstruk

v2.91.0

docling-project/doclingv2.91.0Apr 23, 2026by docling-ops[bot]

AI Summary

Enhancements include VML image extraction from DOCX, strengthened input validation for METS-GBS processing, and various fixes to VLM chat templates, HTML handling, and PPTX notes assignment.

Key Highlights

  • Extract VML images with v:imagedata elements
  • Strengthened METS-GBS input validation
  • VLM chat template preamble fix
  • PPTX notes assigned to ContentLayer.NOTES

New Features

  • VML image extraction
  • METS-GBS validation
  • VLM chat template fix
  • HTML image URL and size handling
  • PPTX notes content layer

Full Release Notes

### Feature

* **docx:** Extract VML images with v:imagedata elements ([#3343](https://github.com/docling-project/docling/issues/3343)) ([`2ddaa3b`](https://github.com/docling-project/docling/commit/2ddaa3be97dbf6d34614ccd5e33f510ba6102033))

### Fix

* Strengthen input validation for METS‑GBS processing ([#3336](https://github.com/docling-project/docling/issues/3336)) ([`c1dbac2`](https://github.com/docling-project/docling/commit/c1dbac22c7a0d55f22d82af5262c9e60df915083))
* EasyOCR model downloading ([#3339](https://github.com/docling-project/docling/issues/3339)) ([`5e161ac`](https://github.com/docling-project/docling/commit/5e161ac1854847f14144310702e336ffeed542fd))
* **vlm:** Remove bogus preamble from VLM chat template ([#3351](https://github.com/docling-project/docling/issues/3351)) ([`c190ba2`](https://github.com/docling-project/docling/commit/c190ba2636ff364a915723256b328237b33ec036))
* **html:** Refine image URL and size handling ([#3348](https://github.com/docling-project/docling/issues/3348)) ([`cd0cb69`](https://github.com/docling-project/docling/commit/cd0cb695303d8ce1b3c9fe620b182b0e22d8c53f))
* Fixes to html_backend ([#3342](https://github.com/docling-project/docling/issues/3342)) ([`9813190`](https://github.com/docling-project/docling/commit/9813190ab4126c1ff2fde1e3e72322821530390b))
* **pptx:** Assign pptx notes to ContentLayer.NOTES ([#3341](https://github.com/docling-project/docling/issues/3341)) ([`3a3c8f6`](https://github.com/docling-project/docling/commit/3a3c8f68dd1f7d3f191a78663ae13445d3643dd0))
* Prevent path traversal in LaTeX macro handlers ([#3330](https://github.com/docling-project/docling/issues/3330)) ([`65ef180`](https://github.com/docling-project/docling/commit/65ef18075fa8720a66aa0a2608313beb0844e2b4))
* **service:** Add explicit usage exceeded exception handling ([#3325](https://github.com/docling-project/docling/issues/3325)) ([`075fa69`](https://github.com/docling-project/docling/commit/075fa69491826a0b295a3022f64c568822a3f68d))

### Documentation

* **uspto:** Improve documentation of USPTO XML parser security config ([#3338](https://github.com/docling-project/docling/issues/3338)) ([`09de7f9`](https://github.com/docling-project/docling/commit/09de7f99df62e0008ee69eeb52b76a2347505d69))