6 Pull requests merged by 6 people

• fix: mime error in document streams

  #1523 merged May 6, 2025

• fix: usage of hashlib for FIPS

  #1512 merged May 2, 2025

• chore: format JSON test files to enable comparison

  #1511 merged May 2, 2025

• chore: restore typing hint for self.script_readers

  #1500 merged Apr 30, 2025

• fix: Guard against attribute errors in TesseractOcrModel __del__

  #1494 merged Apr 30, 2025

• fix: enable cuda_use_flash_attention2 for PictureDescriptionVlmModel

  #1496 merged Apr 30, 2025

5 Pull requests opened by 5 people

• Added support for xlsm files

  #1520 opened May 5, 2025

• docs(enrichment): add enrichments for tables and figures

  #1525 opened May 6, 2025

• fix: Fix issue withe detecting docx files, and files with upper case extentions

  #1528 opened May 6, 2025

• fix(HTML): handle row spans in header rows

  #1536 opened May 6, 2025

• feat: add textbox content extraction in msword_backend

  #1538 opened May 7, 2025

5 Issues closed by 3 people

• Bounding Boxes of Docling Document children inconsistently wrong

  #1398 closed May 6, 2025

• How to convert the Docling document generated from VLLM inference of SmolDocling to Markdown in Golang.

  #1514 closed May 5, 2025

• Option to include page number in document conversion output

  #1506 closed May 2, 2025

• TesseractOcrModel can throw an exception during garbage collection

  #1493 closed Apr 30, 2025

• Flash attention is enabled regardless of `cuda_use_flash_attention_2` flag model for `pipeline_options.do_picture_description`

  #1495 closed Apr 30, 2025

21 Issues opened by 19 people

• using SmolVLM-256M-Instruct

  #1537 opened May 7, 2025

• docling.exceptions.ConversionError: File format not allowed

  #1535 opened May 6, 2025

• processing docx, pptx,... will fail if the zip-container has a folder named "[trash]"

  #1534 opened May 6, 2025

• HTML document conversion fails to extract any content

  #1533 opened May 6, 2025

• Upgrading from 2.28.4 to 2.29.0 breaks docx to markdown pipeline

  #1532 opened May 6, 2025

• Conversion fails with munmap_chunk(): invalid pointer

  #1531 opened May 6, 2025

• Add Textbox content extraction in DOCX backend

  #1529 opened May 6, 2025

• Improve parallelization for remote services

  #1522 opened May 5, 2025

• Performance Issue. 28.1% less inference time on demo case with a simple change.

  #1521 opened May 5, 2025

• Docs for hardware specs

  #1517 opened May 4, 2025

• Smoldocling generation output gets stuck in a loop for certain images.

  #1515 opened May 3, 2025

• export functions not working for smoldocling vlm pipelines

  #1513 opened May 2, 2025

• Cache intermediate results with diskcache

  #1509 opened May 2, 2025

• SmolDocling on Windows (was a 245KB file takes forever in the UI)

  #1510 opened May 2, 2025

• Documentation Improvement: Add Descriptions and Context for Features

  #1505 opened May 1, 2025

• Feature Request: Support for .docx and .pptx similar to pdfPipeline

  #1504 opened May 1, 2025

• Include page images as REFERENCED when saving a docling document as json

  #1503 opened May 1, 2025

• Custom OCR

  #1502 opened May 1, 2025

• Error with blank png: Make sure that the channel dimension of the pixel values match with the one set in the configuration.

  #1501 opened Apr 30, 2025

• Usage of force_full_page_ocr breaks with larger documents

  #1499 opened Apr 30, 2025

• Issue about two images are placed side by side in a DOCX file

  #1498 opened Apr 30, 2025

14 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Stuck when using with Celery

  #1277 commented on Apr 30, 2025 • 0 new comments

• Got Error: Javascript is off when trying to process some PDF files

  #1491 commented on Apr 30, 2025 • 0 new comments

• raises_on_error = False should catch all exceptions, including HTTP errors

  #770 commented on Apr 30, 2025 • 0 new comments

• Converter seems to get stuck on very large pdfs

  #1283 commented on May 5, 2025 • 0 new comments

• Make easyocr optional

  #1462 commented on May 5, 2025 • 0 new comments

• How can I extract the content of headers and footers from a DOCX file?

  #1394 commented on May 5, 2025 • 0 new comments

• Export to markdown only contains H2 headers

  #1023 commented on May 6, 2025 • 0 new comments

• Unable to parse this docx file

  #1403 commented on May 6, 2025 • 0 new comments

• HTML Pipeline IndexError: list index out of range

  #1309 commented on May 6, 2025 • 0 new comments

• TableStructureModel initialization fails: "Cannot copy out of meta tensor" when using CPU device

  #1447 commented on May 7, 2025 • 0 new comments

• feat: Establish confidence estimation for document and pages

  #1313 commented on May 6, 2025 • 0 new comments

• fix: incorrect force_backend_text behaviour for VLM DocTag pipelines

  #1371 commented on May 1, 2025 • 0 new comments

• feat(html): add anchor tag support in HTML conversion

  #1402 commented on May 4, 2025 • 0 new comments

• feat: new HTML backend that handles styled html as well as images

  #1411 commented on May 6, 2025 • 0 new comments