Conversation
…ar to protenix, so not necessary
…flag, and remove abc_config script as this supercedes it
… using mmseqs2, also hopefully have added a fix to the issue in ipsae scoring
… added alphapulldown option to ipsae for shah
…ound ipsae and edited how it's displayed on the output page
…p for chai and changed use_paired option from OF3 as it was breaking things
abcfold/alphafold3/run_alphafold3.py
Outdated
| msa = None | ||
| for sequence in json_dict['sequences']: | ||
| protein = sequence.get("protein") | ||
| if protein is not None: |
There was a problem hiding this comment.
I like not nesting too much for readability. Instead of if protein is not None: You can instead do
if protein is None:
continue
Rest of code
This reduces nesting.
There was a problem hiding this comment.
I didn't realise this was
markdown
- --rest of code--
abcfold/boltz/af3_to_boltz.py
Outdated
| return yaml_string | ||
|
|
||
| def add_key_and_value(self, key: str, value: str): | ||
| def add_key_and_value(self, key: str, value): |
There was a problem hiding this comment.
Why was the typing removed from here?
There was a problem hiding this comment.
the value isn't always a string, so it was breaking for bonded pairs iirc
Could potentially replace with Union[str, bool]
| except ValueError: | ||
| logger.error("ValueError when calculating ipSAE score, bypassing ipSAE") | ||
| pass | ||
| ipsae = Ipsae(model, pae_obj) |
abcfold/output/protenix.py
Outdated
| for seed in self.seeds | ||
| } | ||
|
|
||
| def update_chain_labels(self, cif_file) -> CifFile: |
|
Looking good. I think we should add mypy with type-checking to ensure we have all our type-hints in order. Also, what's your local test coverage getting. we have 67% in actions but we don't run the programs here. |
|
I've made all the changes you suggested, included the addition of mypy type-checking (this has been the most labour intensive part) The local test coverage is sitting at 77% when running the programs on the cluster. |
Updated a few aspects of ABCFold, most notably: