This is the first step of handing-over research code internally to our research engineers.
Current MR issues:
- Fix Issue with omitting language information when adding new descriptions, labels, or aliases
- Update logic for Wikidata labels processing
- Update model binary (remove labels dictionary)
- Update model (recollect data + retrain)
Rewrite at least the following scripts:
- parsing_utils.py
- text_utils.py
After that, research engineers will make another round of review. Note - there might be iterations of this reviewing process, and research engineers should be able to merge the code before this task is considered completed.