zopen: explicit binary/text mode , and explicit encoding as UTF-8

[DanielYang59]

Summary

• Explicit binary/text mode for zopen, I have opened a PR from monty's side to prevent this zopen changes: forbid implicit binary/text mode, signature change, default UTF-8 encoding in text mode, drop .z support after one-year materialsvirtuallab/monty#730
• Explicitly use UTF-8 encoding for zopen

Detailed reasoning and example are given in materialsvirtuallab/monty#730 (comment). In a word, not passing b/t in mode would cause different and unexpected behaviour when opening a text file (file.txt) versus a compressed file (file.gz), for the former we would get TextIO which gives string but BinaryIO for the latter which gives bytes.

---

There're few cases where this wouldn't cause problem, for example json.load accepts both binary and text IO:

json.load(fp, *, cls=None, object_hook=None, parse_float=None, parse_int=None, parse_constant=None, object_pairs_hook=None, **kw)

Deserialize fp (a .read()-supporting text file or binary file containing a JSON document) to a Python object using this conversion table.

For such cases, I tested with the pdag3_complete_dos.json.gz file, "text mode" seems slightly faster:

Binary mode: Loaded data (first 100 chars): {'@module': 'pymatgen.electronic_structure.dos', '@class': 'CompleteDos', 'efermi': -0.28973031, 'st...
Time taken (binary mode): 0.3674 seconds
Text mode: Loaded data (first 100 chars): {'@module': 'pymatgen.electronic_structure.dos', '@class': 'CompleteDos', 'efermi': -0.28973031, 'st...
Time taken (text mode): 0.3450 seconds

[DanielYang59]

This would fail if filename is a name for compressed file (e.g. file.gz), in this case file would be BinaryIO and we cannot write str into it:

pymatgen/src/pymatgen/io/cif.py

Lines 1758 to 1765 in 31f1e1f

	def write_file(
	self,
	filename: str | Path,
	mode: Literal["w", "a", "wt", "at"] = "w",
	) -> None:
	"""Write the CIF file."""
	with zopen(filename, mode=mode) as file:
	file.write(str(self))

w and a have been removed in 4fccd59, with default changed to wt