Edouard Belval comments

Results 293 comments of


                                            Edouard Belval

Intermittent failure of to_markdown() in lambda

This seems similar to #424 which should be fixed in 1.9.2. Can you test with the latest version?

Landscape tables result in jumbled text when using extractor.start_document_analysis with TextractFeatures.TABLES

Thank you for the thorough and nicely formatted issue. My hunch would be that there is sorting happening somewhere that incorrectly assumes that all text is left to right and...

Incorrect table cell word and line order

I'll try to reproduce the issue on our side and get back to you on this. Thanks!

Incorrect table cell word and line order

I was able to find the asset online, the issue is that the page is rotated and Textractor has a significant amount of logic that does ordering based on x...

JPEG conversion in `analyze_document` significantly impacts table predictions

Issue is mitigated by settign the JPEG compression parameters, will require further discussion for using PNG as we otherwise see a significant latency hit.

Proper way of getting cell content?

Could you provide an example? `cell.text` should be the correct one but I suspect what you are seeing is caused by a merged cell. A table cell that belongs to...

analyze_expense error: 'NoneType' object has no attribute 'spatial_object'

I was able to reproduce the issue. Once #404 is merged and version 1.8.4 is released it should be fixed.

[textractprettyprinter] List contents are duplicated when generating text output using `get_text_from_layout_json`

Regarding that last message do you have an example of `LAYOUT_LIST` containing `LAYOUT_SECTION_HEADER` or `LAYOUT_TITLE`?

Added config parameter

What is the use case here? Isn't this redundant since Config contains region as well?

Mistake a text field above a table as table title

If this happens often you can probably get away with a coarse filtering on the `len(table.title.get_text()) > 50` (or another suitable value). Unfortunately this is a Textract Tables issue and...