Replace pybtex or make it optional?

[DanielYang59]

Looks like development on pybtex might have stalled (no release since Jan 2021), and the following would error out on MacOS Python 3.13:

from pybtex.database import BibliographyData, Entry

Gives:

Traceback (most recent call last):
  File "/Users/yang/developer/pymatgen/debug/test_pybtex_install.py", line 1, in <module>
    from pybtex.database import BibliographyData, Entry
  File "/Users/yang/developer/pymatgen/venv313/lib/python3.13/site-packages/pybtex/database/__init__.py", line 44, in <module>
    from pybtex.plugin import find_plugin
  File "/Users/yang/developer/pymatgen/venv313/lib/python3.13/site-packages/pybtex/plugin/__init__.py", line 26, in <module>
    import pkg_resources
ModuleNotFoundError: No module named 'pkg_resources'

This is an already fixed issue from pybtex's side but they haven't got a release with this patch yet.

Should we consider replace it (python-bibtexparser for example) or make it optional as it's only used in two places across the code base:

cif.CifParser.get_bibtex_string

pymatgen/src/pymatgen/io/cif.py

Lines 1349 to 1361 in 5b997f7

	def get_bibtex_string(self) -> str:
	"""Get BibTeX reference from CIF file.
	Args:
	data:
	Returns:
	BibTeX string.
	"""
	try:
	from pybtex.database import BibliographyData, Entry
	except ImportError:
	raise RuntimeError("Bibliographic data extraction requires pybtex.")

Helper function needed by util.provenance.StructureNL

pymatgen/src/pymatgen/util/provenance.py

Lines 38 to 53 in 5b997f7

	def is_valid_bibtex(reference: str) -> bool:
	"""Use pybtex to validate that a reference is in proper BibTeX format.
	Args:
	reference: A String reference in BibTeX format.
	Returns:
	bool: True if reference is valid bibtex.
	"""
	# str is necessary since pybtex seems to have an issue with unicode. The
	# filter expression removes all non-ASCII characters.
	str_io = StringIO(reference.encode("ascii", "ignore").decode("ascii"))
	parser = bibtex.Parser()
	errors.set_strict_mode(enable=False)
	bib_data = parser.parse_stream(str_io)
	return len(bib_data.entries) > 0

---

Related:

• Support python3.13 #4100
• Python 3.13 support #4284

[DanielYang59]

I would give it a try and make it optional for now (0b25745)