DS4SD / docling-ibm-models

MIT License
44 stars 10 forks source link

fix: TableFormer raises IndexError: too many indices for array #17

Closed maxmnemonic closed 2 months ago

maxmnemonic commented 2 months ago

Fix for "TableFormer IndexError: too many indices for array" BUG. This would happen when PDF backend for whatever reason doesn't return PDF cells for a given table (broken PDF, or non empty PDF without text cells, etc.), and match with PDF cells is requested. Bug fixed and added test case to catch this.