Address technical debt #41
Description
The technical debt is creeping in
- There are scripts in src that are completely obsolete. Example: https://github.com/swerik-project/scripts/blob/main/src/cur-prot/reclassify.py which is even part of https://github.com/swerik-project/scripts/blob/main/src/cur-prot/post-pipeline.py. I deleted two more a few days ago. Potentially more left still.
- There is probably stuff on other branches that should have been merged months or years ago
- The folder structure is all over the place, some things are neatly in subfolders (eg. cur-prot, cur-mot), some randomly just in src (eg. sampling stuff), and some in src despite a subfolder existing
- Documentation can be out of date, but IMO it's a consequence of technical debt, not a core issue
Tasks:
- Make bert-riksdagen-classifier a submodule of scripts called src-bert
- Delete all obsolete scripts in src
- Organize everything in src into one level of subfolders
- Make sure each script has a docstring, with a "validated 2026-02-09" tag so that we can follow when they have been checked
- Automatically test docstring existance and that everything has been validated in the past year
- Make a new script or adjust post-pipeline.py for the borne-digital era
- Go through all branches, merge or delete