COAD proteomics download fails for all sources with NoneType // int error

[isaakovakate]

I’m seeing a reproducible failure when accessing COAD proteomics through CPTAC. My cptac package is up to date, and the issue occurs for COAD from both UMICH and BCM.

import cptac
coad = cptac.Coad()
proteomics = coad.get_proteomics('umich')

cptac error: Failed to download data file for umich coad mapping with error:
unsupported operand type(s) for //: 'NoneType' and 'int'

I see COAD failure for proteomics specifically, and the other datasets load fine.

Could you please help me with the COAD mapping download path and the NoneType // int failure? It looks like, it could be a problem with source mapping for COAD.

[jnktsj]

I've had the same issue, debugged, and found it was a space in the file name causing the error.

Here is my fix - instead of making a pull request, I'll post it here.

# cptac/tools/download_tools.py: modified two lines in two functions
def download(cancer: str, source: str, dtype: str, data_file: str) -> bool:
    ...
    # replace space with underscore for downloading file
    output_file = os.path.join(output_dir, data_file.replace(" ", "_")) 

def get_data(url: str, subfolder: str = '', num_threads: int = 4) -> str:
    ...
    # place back a space to enable file name search
    file_name = url.split('/')[-2].replace("%20", " ") 

# cptac/cancers/umich/umichcoad.py
def load_mapping(self):
    ...
    # replace spaces in the downloaded mapping file name with underscores
    df = pd.read_excel(file_path.replace(" ", "_"), index_col = 'Label', usecols = ['Label', 'Sample Code'])