Describe the bug
I tried pruning new networks and am getting a value error
ValueError: could not convert string to float: 'A0AVT1'
To Reproduce
Load weighted network that has either an accession site that is NOT named "substrate_acc" and prune from there.
Expected behavior
Since I set accession and site columns, I expected KSTAR to ignore those columns and only try to produce pruned networks for the rest of the columns.
Screenshots
Additional context
My understanding is prune.py line 52 determines non-numeric columns by the parameter "nonweight_cols". It will skip these columns as it iterates thru the df to make pruned networks. I expected nonweight columns to automatically include the columns set by the user as acc column and site column but it seems unless the user ALSO specifies that the site and and acc columns are nonweight columns, KSTAR will automatically look for the default names for these columns ['substrate_acc','site','substrate_id', 'substrate_name', 'pep'] from the NetworKIN matrix from KinPred. Not a big deal once I figured it out but did confuse me for a bit.
Describe the bug
I tried pruning new networks and am getting a value error
ValueError: could not convert string to float: 'A0AVT1'
To Reproduce
Load weighted network that has either an accession site that is NOT named "substrate_acc" and prune from there.
Expected behavior
Since I set accession and site columns, I expected KSTAR to ignore those columns and only try to produce pruned networks for the rest of the columns.
Screenshots
Additional context
My understanding is prune.py line 52 determines non-numeric columns by the parameter "nonweight_cols". It will skip these columns as it iterates thru the df to make pruned networks. I expected nonweight columns to automatically include the columns set by the user as acc column and site column but it seems unless the user ALSO specifies that the site and and acc columns are nonweight columns, KSTAR will automatically look for the default names for these columns ['substrate_acc','site','substrate_id', 'substrate_name', 'pep'] from the NetworKIN matrix from KinPred. Not a big deal once I figured it out but did confuse me for a bit.