fix load_bulk_data

[alondmnt]

annoying little bugs. see commits.

test:

pl = PhenoLoader('rna_seq')
pl.load_bulk_data('batch_metadata_parquet')

[alondmnt]

we may also face another potential issue (haven't dived into it, so it's just a hunch).

I think that get_function_for_field_type, when calling load_bulk_data on bulk dictionary fields, does not use the load_func of the parent_dataframe.

for instance, if I ask for a microbe it uses the default read_parquet, but not what is specified for group bulk data.
in the future, if I will ask for a field in a DICOM file, it will not use the load_func for DICOM files, but rather the default read_parquet.

I think...

[MariaGorodetski]

Super! - on the fix and the example.
Regarding the problem, I understand.
We use the field_type of the specific field in the bulk dictionary. I need to think about an elegant solution...