Currently, the proof and meta data is kept separate.
This seems like it could cause a lot of problems down the road,
as well as it is currently making the data less manageable.
With this issue we want to fix:
Download meta data and proof data in the same script
Store the data together in the same dict
Update train test split
Update SINE extraction to just process all problems (independent of existing proofs due to E messing the attempts up)
Currently, the proof and meta data is kept separate. This seems like it could cause a lot of problems down the road, as well as it is currently making the data less manageable.
With this issue we want to fix: