BUG: A Molecule object in the cclib TaskDocument can't be printed

[Andrew-S-Rosen]

Minimal example:

from atomate2.common.schemas.cclib import TaskDocument

task = TaskDocument.from_logfile(".", "orca.out")
print(task)

orca.out.txt

Traceback:

ValueError                                Traceback (most recent call last)
File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/IPython/core/formatters.py:708, in PlainTextFormatter.__call__(self, obj)
    701 stream = StringIO()
    702 printer = pretty.RepresentationPrinter(stream, self.verbose,
    703     self.max_width, self.newline,
    704     max_seq_length=self.max_seq_length,
    705     singleton_pprinters=self.singleton_printers,
    706     type_pprinters=self.type_printers,
    707     deferred_pprinters=self.deferred_printers)
--> 708 printer.pretty(obj)
    709 printer.flush()
    710 return stream.getvalue()

File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/IPython/lib/pretty.py:410, in RepresentationPrinter.pretty(self, obj)
    407                         return meth(obj, self, cycle)
    408                 if cls is not object \
    409                         and callable(cls.__dict__.get('__repr__')):
--> 410                     return _repr_pprint(obj, self, cycle)
    412     return _default_pprint(obj, self, cycle)
    413 finally:

File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/IPython/lib/pretty.py:778, in _repr_pprint(obj, p, cycle)
    776 """A pprint that just redirects to the normal repr function."""
    777 # Find newlines and replace them with p.break_()
--> 778 output = repr(obj)
    779 lines = output.splitlines()
    780 with p.group():

File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/pydantic/utils.py:411, in pydantic.utils.Representation.__repr__()

File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/pydantic/utils.py:390, in genexpr()

File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/pydantic/utils.py:390, in genexpr()

File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/pymatgen/core/structure.py:3113, in IMolecule.__repr__(self)
   3111 outs = ["Molecule Summary"]
   3112 for s in self:
-> 3113     outs.append(s.__repr__())
   3114 return "\n".join(outs)

File ~/software/miniconda/envs/quacc/lib/python3.10/site-packages/pymatgen/core/sites.py:211, in Site.__repr__(self)
    210 def __repr__(self):
--> 211     return f"Site: {self.species_string} ({self.coords[0]:.4f}, {self.coords[1]:.4f}, {self.coords[2]:.4f})"

ValueError: Unknown format code 'f' for object of type 'numpy.str_'

The issue is in task["attributes"]["molecule_unoriented"].sites I believe. Something is getting stored as a np.str_ instead of str. Annoying.

Tagging @janosh in case this has something to do with Pymatgen directly, but I don't think it does. I'll revisit this in a few days.

[JaGeo]

I had something similar with the Bandstructure Object. materialsproject/pymatgen#3096

[Andrew-S-Rosen]

Interesting... thanks for sharing!

[janosh]

I'm guessing this is a deserialization bug in cclib.io.ccread which is used to parse the logfile

atomate2/src/atomate2/common/schemas/cclib.py

Line 130 in 896604a

cclib_obj = ccread(logfile, logging.ERROR)

and then asked for the atomcoords

atomate2/src/atomate2/common/schemas/cclib.py

Line 190 in 896604a

coords = cclib_obj.atomcoords

Coordinates clearly shouldn't be parsed as strings but floats. I don't think going from np.str_ to str would help here, so I think materialsproject/pymatgen#3096 is unrelated.

We could accommodate for string types in pymatgen by casting coords to floats here but would definitely be cleaner to not read coordinates as strings in the first place.

[Andrew-S-Rosen]

I'm closing this as not reproducible.

I did

conda create --name atomate2 python=3.10
pip install pymatgen atomate2 cclib

and ran the above code block. It works fine. However, in another environment I have, it does not work. Both have the same versions of Python, pymatgen, cclib, atomate2, and numpy, so I'm at a loss. I'll just pray that it doesn't come up again, but if it does, I'll re-open the issue.

[Andrew-S-Rosen]

I've reopened it because I can now reproduce the issue! lol. It only happens with the main branch of Atomate2, probably because that's when I last made a change to the TaskDocument. Okay, now for me to dig into. Sorry about the noise, friends.

conda create --name atomate2 python=3.10
pip install pymatgen cclib
pip install git+https://github.com/materialsproject/atomate2.git