Closed DimaMolod closed 6 months ago
I also tried to run with explicit parameters in config.yaml for --data_directory and --features_directory like this:
...
create_feature_arguments:
--save_msa_files: False
--use_precomputed_msas: True
--max_template_date: 2050-01-01
--skip_existing: False
structure_inference_arguments:
--num_predictions_per_model : 10
--num_cycle : 24
--data_directory : /scratch/AlphaFold_DBs/2.3.2
--features_directory : /scratch/dima/AP/Z-disc/ap20/features
...
...and it did not help...
The v comes from missing brackets here:
cli_parameters = " ".join( [f"{k}=v" for k, v in config["structure_inference_arguments"].items()] ) Fixed in 4477b8e187a50df4cee45ad791e523ac38182411 features_directory should not be specifiable by the user and data_directory should be defined outside structure_inference_arguments. Perhaps there is a way to ensure that in future versions
3 minutes to fix a bug, love you @maurerv
I tried to run the latest version from main but got this error:
`FATAL Flags parsing error: flag --data_directory=None: Flag --data_directory must have a value other than None. flag --features_directory=None: Flag --features_directory must have a value other than None. Pass --helpshort or --helpfull to see help on flags. [Wed May 29 14:31:08 2024] Error in rule structure_inference: jobid: 0 input: /scratch/dima/AP/Z-disc/ap20/features/Q702N8-2.pkl, /scratch/dima/AP/Z-disc/ap20/features/Q86TC9.pkl output: /scratch/dima/AP/Z-disc/ap20/predictions/Q702N8-2:837-978_Q86TC9/completed_fold.txt shell:
MAXRAM=$(bc <<< "$(ulimit -m) / 1024.0")
Could it be because there is a white space after --protein_delimiter and not '='? Where did these 'v' letters come from? Also, what is the proper way of debugging code with snake deploy? Do I need to create a new branch on git and re-deploy to debug or there is a easier way?