Metadata values type issues #816
Description
Describe the bug
The filter derive_annotation_from_compound_name will fail since numpy 2 stricter dtype handling.
While calculating mass_differences compound_name_annotation_df["monoisotopic_mass"] uses float64 parent_mass is a str.
To Reproduce
Steps to reproduce the behavior:
- load_from_mgf with a file with missing inchi and smiles.
- process the metadata with:
from matchms.filtering import derive_annotation_from_compound_name
def metadata_processing(spectrum):
spectrum = derive_annotation_from_compound_name(spectrum)
return spectrum
spectra = [metadata_processing(s) for s in spectra]
Will result in:
UFuncTypeError Traceback (most recent call last)
Cell In[5], line 1
----> 1 spectra = [metadata_processing(s) for s in spectra]
Cell In[4], line 4, in metadata_processing(spectrum)
3 def metadata_processing(spectrum):
----> 4 spectrum = derive_annotation_from_compound_name(spectrum)
6 return spectrum
File [~\workspace\matchms\matchms\filtering\metadata_processing\derive_annotation_from_compound_name.py:77](http://localhost:8888/lab/tree/~/workspace/matchms/matchms/filtering/metadata_processing/derive_annotation_from_compound_name.py#line=76), in derive_annotation_from_compound_name(spectrum_in, annotated_compound_names_file, mass_tolerance, clone)
75 print(parent_mass)
76 print(type(parent_mass))
---> 77 mass_differences = np.abs(compound_name_annotation_df["monoisotopic_mass"] - parent_mass)
...
UFuncTypeError: ufunc 'subtract' did not contain a loop with signature matching types (dtype('float64'), dtype('<U8')) -> None