Hi
I have a local instance of the sequence server and have found that there is a bug in the python section of the code for which I have created a work around for myself, but I recommend this is fixed as it prevents proper completion of the process with a “Finished” code …
File “…/evcouplings/compare/pdb.py”, line 1332, in load_structures
structures[pdb_id] = PDB.from_id(pdb_id)
File “…/evcouplings/compare/pdb.py”, line 657, in from_id
return cls(f, keep_full_data=keep_full_data)
File “…/evcouplings/compare/pdb.py”, line 547, in init
…
TypeError: unhashable type: ‘Series’
Line of code =
sse_table = sse_table.query(
f"conformation_type.str.startswith(‘{sse_filter}’)"
)
The problem is that the sse_table pandas data frame is being filtered (with query) on type but this is an object and not a string and using type.str does coercion and does not force the type to string. For now I have but this code in a try/except as I have the data I need by this point and I don’t know what the down stream affect would be to change the dataframe type to “str” from “object” in your code.
# filter table down to relevant entries for current secondary structure type
if sse_filter is not None:
try: # TOM WAS HERE - crashes out with bad sse_filter
sse_table = sse_table.query(
f"conformation_type.str.startswith('{sse_filter}')"
)
except Exception:
print("Failed to read SSE filter (TOM edit) ")
print(sse_filter)
print(sse_table)
print(sse_table['conformation_type'])
print(type(sse_table))
break
Thank you
Tom