macrocosm-os / folding

Decentralized Protein Folding Bittensor Subnet
https://www.macrocosmos.ai/sn25
MIT License
14 stars 16 forks source link

Parsing Error tracking: gro_hash #86

Open schampoux opened 4 months ago

schampoux commented 4 months ago

The gro_hash method in folding/utils/ops.py encounters errors for certain proteins/DNA. This method is responsible for generating the hash for a specific gro file. It does this by parsing the .gro file, connecting the residue name, atom name, and residue number from each line together into a single string. The error arises because some of the atoms contain apostrophe's.

The goal for this issue is to increase the robustness of the gro_hash generator so that it can accommodate for cases like this. Another goal for this issue is to track other errors.

schampoux commented 4 months ago

87 PR from @mccrindlebrian addressing the gro_hash with provided testing files.