[BUG] pypet2bids collects additional values from spreadsheet that DNE; csv

[bendhouseart]

Describe the bug Given the attached spreadsheet as input to the -m flag, pet2bids raises a flatten series error, it seems to collect two values from the csv where there should only be on, n/a in this example. However, it's returning a series containing [0, 'n/a'] which is believed to create this error, see:

Traceback (most recent call last):
  File "/usr/local/bin/dcm2niix4pet", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.8/dist-packages/pypet2bids/dcm2niix4pet.py", line 1354, in main
    converter = Dcm2niix4PET(
  File "/usr/local/lib/python3.8/dist-packages/pypet2bids/dcm2niix4pet.py", line 471, in __init__
    helper_functions.single_spreadsheet_reader(
  File "/usr/local/lib/python3.8/dist-packages/pypet2bids/helper_functions.py", line 155, in single_spreadsheet_reader
    metadata[field] = flatten_series(series)
  File "/usr/local/lib/python3.8/dist-packages/pypet2bids/helper_functions.py", line 103, in flatten_series
    raise Exception(f"Invalid Series: {series}")
Exception: Invalid Series: 0   NaN
Name: SpecificRadioactivity, dtype: float64

pet_metadata.csv

Expected behaviour Reading in the attached spreadsheet should yield a specific radioactivity of n/a.

Desktop (please complete the following information):

• OS: linux
• Version: python 3.8, pandas 1.0.1, pet2bids 1.2.4

[dlevitas]

This might partially be because having a value of n/a in Excel (for "SpecificRadioactivity" and "SpecificRadioactivityUnits" metadata fields) get interpreted as NaN by python. Explicitly making n/a a string value (i.e. "n/a") resolves this issue but triggers a new error message:

File "/usr/local/lib/python3.8/dist-packages/pypet2bids/dcm2niix4pet.py", line 991, in check_meta_radio_inputs
    tmp = ((InjectedRadioactivity * (10 ** 6) / SpecificRadioactivity) * (10 ** 6))
TypeError: unsupported operand type(s) for /: 'int' and 'str'

Which seems to be caused by an improper variable type for the "SpecificRadioactivity" and "SpecificRadioactivityUnits" metadata. Simply removing these two fields from the metadata files resolves all issues.

[dlevitas]

I may be mis-reading the BIDS-specification for PET data, but for these metadata fields, it appears that "n/a" is acceptable and even should be entered, assuming it isn't an FDG sequence.

[bendhouseart]

Yes, "n/a" is accepted as often the data simply isn't available. We'd rather people share what they have then give up entirely.

[bendhouseart]

@dlevitas closing this as I believe it was fixed some time ago, just not closed. Let me know if otherwise.