Table cell word order is incorrect in multi line cells

[Hoho5000]

Using this library to extract table data can lead to incorrect cell text when multiple lines of text are involved.

For example, the following table headers:

 ___________      _______________
| RETURN OF |    |     TAX       |
|  CAPITAL  |    | (WITHHOLDING) |
 ‾‾‾‾‾‾‾‾‾‾‾     |    REFUND     |
                  ‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾

becomes "RETURN CAPITAL OF" and "(WITHHOLDING) TAX REFUND" when creating the table cell object.

This is a result of table_cell.py (Line 177).

If I'm not mistaken, it's my understanding that where this is called internally that the word IDs are already sorted in the response.

Removing the sorting and setting the words directly fixes the issue for me.

Is there a particular reason for words that get added to the table cell need to be "sorted" based on bounding boxes, or can that be removed?

[Belval]

I opened a PR that addresses this issue, if you get a chance to look it over and confirm that your fix is similar I can merge and push the new version to PyPI.

[Hoho5000]

Thanks for the quick response and fix!

Yes, that is the same fix, though I didn't realize that a couple of other places were setting the order as well.

I Just tested this and it appears to work correctly.