Christoph Auer comments

Results 170 comments of


                                            Christoph Auer

fix: Add page image injection in Layout+VLM pipeline

This PR will be superseded by another one, since it introduced unnecessary code duplication for this purpose.

Improve backend resolution logic

Another sample of a word document not detected as such is seen in issue https://github.com/DS4SD/docling/issues/476.

Fix page_range stopping at page 32 when start >= 30

@copilot this fix is nonsense, therefore closing.

RuntimeError: [json.exception.type_error.302] type must be string, but is object

This appears to be addressed. Closing.

OCR Extracted Information

@maliktalha370 Can you please elaborate on what your expectations would be? The text information we output in JSON or Markdown is both the programmatic text embedded in PDFs, and the...

Using Docling with costume layout and table recognition models

@ALIYoussef We would be excited to see alternative layout or table structure models implementations from the community. The example above posted by @dolfim-ibm is a good way to understand the...

Issue when locale is set to french

@yannistml I can confirm docling starts to hang on our standard test PDF in `tests/data/pdf/2206.01062.pdf` and produces garbage output in the end. The problem appears to be rooted in the...

feat: support xlsm files

@ShiroYasha18 Thanks for the updates! To get it finalized, can you please get the CI checks green? - Run the pre-commit toolchain: `poetry run pre-commit run --all-files -v` - Ensure...

feat: support xlsm files

@ShiroYasha18 I updated the tests now. Let's see if CI passes, then it should be ok.

fix(layout_utils): correct conditional logic in adapt_bbox function

@Raphilanthrope I can not find any logical difference between the original code and your proposal. Do you have a practical case where this change makes a difference? If yes, please...