microsoft / evodiff

Generation of protein sequences and evolutionary alignments via discrete diffusion models
MIT License
516 stars 73 forks source link

Missing uniref50_aa_ref_test.csv #12

Closed andrew0901 closed 1 year ago

andrew0901 commented 1 year ago

Hi, I successfully generated sequences for the unconditional situation. However, at the end as main() tried to call aa_reconstruction_parity_plot() (line 148 in evodiff/generate.py) to plot the reconstruction parity, the reference file seems to be missing, giving:

FileNotFoundError: [Errno 2] No such file or directory: '/wynton/home/rotation/fhy/Desktop/DMs/ref/uniref50_aa_ref_test.csv'

I'm wondering where am I supposed to find the file, or is there anything I did wrong?

sarahalamdari commented 1 year ago

Hi! I've added this file to data/ in a recent push.

Nothing is wrong, this file just contains the pre-computed counts of each AA in the test set.

yangkky commented 1 year ago

Thanks @andrew0901 for pointing this out, and @sarahalamdari for the PR.

Fixed by PR #13