Tables and text, oh my! #13

Merged

dhdaines merged 20 commits from tables_and_text_oh_my into main

2025-08-01 00:20:54 -04:00

dhdaines commented

2025-07-31 09:47:04 -04:00

(Migrated from github.com)

Adds

paves.text which mostly just has the function words for now. This gives you whitespace-separated text cells with metadata that can be easily used to visualize or what have you
paves.tables finds tables with logical structure if present, with RT-DETR if not and if transformers is installed

Not here yet:

Efficiency
Table structure prediction (which we will also do with logical structure if possible, then backing off to TableFormer if not)

Adds - `paves.text` which mostly just has the function `words` for now. This gives you whitespace-separated text cells with metadata that can be easily used to visualize or what have you - `paves.tables` finds tables with logical structure if present, with RT-DETR if not and if transformers is installed Not here yet: - Efficiency - Table structure prediction (which we will *also* do with logical structure if possible, then backing off to TableFormer if not)