Unify input syntax

[jkosinski]

Just parking it here for future discussion - we have now two syntaxes, new for snakemake mode and legacy for script-based mode, and as a user of both I already find it confusing. We should consider adding the new syntax also for the scripts mode.

[DimaMolod]

Just parking it here for future discussion - we have now two syntaxes, new for snakemake mode and legacy for script-based mode, and as a user of both I already find it confusing. We should consider adding the new syntax also for the scripts mode.

Just a comment: it's actually different syntax for run_multimer_jobs.py and run_structure_prediction.py, has nothing to do with snakemake. The newer syntax is better, but we keep the original one in run_multimer_jobs.py for back compatibility reasons.

[jkosinski]

It might be that we will need to keep run_multimer_jobs.py or implement equivalent job management in run_structure_prediction.py as for single or few jobs snakemake pipeline feels adding too much overhead (in my cases docker pulls and symlink jobs to reuse features taking more than actual prediction, plus all those extra files and storage that for now seems worthwhile). This could be probably as sth as simple as just supporting custom mode (so raw new syntax) in run_structure_prediction.py but just adding a possibility to run over all lines in an input file.

[DimaMolod]

or we can just re-use this function if input is a file in run_structure_prediction.py, similarly to run_multimer_jobs.py

[DimaMolod]

@maurerv but this function seems to be duplicated and different between snakemake and alphapulldown?!