[Bug]: Adding list attributes to non-existing field raises TypeError

[rettigl]

Contact Details

[email protected]

What happened?

If I try to add attributes for a non-existing field (e.g. because the field was not found in the data), then the write raises a TypeError if one of the attributes contains a list.
This seems to be happening in this line:

pynxtools/src/pynxtools/dataconverter/writer.py

Line 326 in f30bc35

dataset = self.ensure_and_get_parent_node(

An example can be created by removing the value for one of the transformations in the mpes example, then the @axis attribute raises this error.

Relevant log output

Input:
del sp.binned.attrs["metadata"]["file"]["trARPES:Carving:TRX.RBV"]
sp.save(data_path + "/binned.nxs")

Output:
# save to NXmpes NeXus (including standardized metadata)
del sp.binned.attrs["metadata"]["file"]["trARPES:Carving:TRX.RBV"]
sp.save(data_path + "/binned.nxs")

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/pynxtools/src/pynxtools/dataconverter/writer.py:326, in Writer._put_data_into_hdf5(self)
    324 else:
    325     # consider changing the name here the lvalue can also be group!
--> 326     dataset = self.ensure_and_get_parent_node(
    327         path, self.data.undocumented.keys()
    328     )
    329     dataset.attrs[entry_name[1:]] = data

File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/pynxtools/src/pynxtools/dataconverter/writer.py:248, in Writer.ensure_and_get_parent_node(self, path, undocumented_paths)
    246 grp = parent.create_group(parent_path_hdf5)
--> 248 attrs = self.__nxdl_to_attrs(parent_path)
    250 if attrs is not None:

File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/pynxtools/src/pynxtools/dataconverter/writer.py:236, in Writer.__nxdl_to_attrs(self, path)
    235     if key in self.data:
--> 236         elem.attrib[attr_name.get("name")] = self.data[key]
    238 return elem.attrib

File src/lxml/etree.pyx:2447, in lxml.etree._Attrib.__setitem__()

File src/lxml/apihelpers.pxi:594, in lxml.etree._setAttributeValue()

File src/lxml/apihelpers.pxi:1539, in lxml.etree._utf8()

TypeError: Argument must be bytes or unicode, got 'list'

The above exception was the direct cause of the following exception:

OSError                                   Traceback (most recent call last)
Cell In[20], line 3
      1 # save to NXmpes NeXus (including standardized metadata)
      2 del sp.binned.attrs["metadata"]["file"]["trARPES:Carving:TRX.RBV"]
----> 3 sp.save(data_path + "/binned.nxs")

File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/sed/sed/core/processor.py:2537, in SedProcessor.save(self, faddr, **kwds)
   2534     if "eln_data" in kwds:
   2535         input_files.append(kwds.pop("eln_data"))
-> 2537     to_nexus(
   2538         data=data,
   2539         faddr=faddr,
   2540         reader=reader,
   2541         definition=definition,
   2542         input_files=input_files,
   2543         **kwds,
   2544     )
   2546 else:
   2547     raise NotImplementedError(
   2548         f"Unrecognized file format: {extension}.",
   2549     )

File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/sed/sed/io/nexus.py:40, in to_nexus(data, faddr, reader, definition, input_files, **kwds)
     37 else:
     38     input_files = tuple(input_files)
---> 40 convert(
     41     input_file=input_files,
     42     objects=(data),
     43     reader=reader,
     44     nxdl=definition,
     45     output=faddr,
     46     **kwds,
     47 )

File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/pynxtools/src/pynxtools/dataconverter/convert.py:241, in convert(input_file, reader, nxdl, output, skip_verify, **kwargs)
    231 data = transfer_data_into_template(
    232     input_file=input_file,
    233     reader=reader,
   (...)
    237     **kwargs,
    238 )
    240 helpers.add_default_root_attributes(data=data, filename=os.path.basename(output))
--> 241 Writer(data=data, nxdl_f_path=nxdl_f_path, output_path=output).write()
    243 logger.info(f"The output file generated: {output}.")

File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/pynxtools/src/pynxtools/dataconverter/writer.py:339, in Writer.write(self)
    337 """Writes the NeXus file with previously validated data from the reader with NXDL attrs."""
    338 try:
--> 339     self._put_data_into_hdf5()
    340 finally:
    341     self.output_nexus.close()

File /mnt/pcshare/users/Laurenz/AreaB/transformations_testing/pynxtools/src/pynxtools/dataconverter/writer.py:331, in Writer._put_data_into_hdf5(self)
    329         dataset.attrs[entry_name[1:]] = data
    330 except Exception as exc:
--> 331     raise IOError(
    332         f"Unknown error occured writing the path: {path} "
    333         f"with the following message: {str(exc)}"
    334     ) from exc

OSError: Unknown error occured writing the path: /ENTRY[entry]/SAMPLE[sample]/TRANSFORMATIONS[transformations]/AXISNAME[trans_x]/@depends_on with the following message: Argument must be bytes or unicode, got 'list'

[lukaspie]

I can confirm that this error occurs with the latest pynxtools and pynxtools-mpes branches. I guess the default behavior should be that these attributes are not written at all and there's a warning, right?