Why isotopes of hydrogen are considered when enumerating through Element

[DanielYang59]

Noticed the following behaviour:

from pymatgen.core import Element

for idx, ele in enumerate(Element):
    print(idx, ele)

0 H
1 H  # Deuterium
2 H  # Tritium
3 He
4 Li
5 Be
6 B
7 C
8 N
9 O
......

I would expect looping through Element to behave similarly to looping through the periodic table, i.e. isotopes should not be considered?

pymatgen/src/pymatgen/core/periodic_table.py

Lines 875 to 888 in a09e7dd

	class Element(ElementBase):
	"""Enum representing an element in the periodic table."""
	# This name = value convention is redundant and dumb, but unfortunately is
	# necessary to preserve backwards compatibility with a time when Element is
	# a regular object that is constructed with Element(symbol).
	H = "H"
	D = "D"
	T = "T"
	He = "He"
	Li = "Li"
	Be = "Be"
	B = "B"
	C = "C"

---

Say if I want to get all elements (for example when plotting periodic tables):

I have to do the following "hack" to skip D/T:

elements = list(Element)[0:1] + list(Element)[3:]

Or make use the fact that "symbol" of "H/D/T" are all "H" if I just need the symbols (I have to do this in pymatviz):

elements = [elem.symbol for elem in list(Element)[2:]]

@mkhorton @janosh Let's discuss here, sorry for going off-topic in the original PR.

[DanielYang59]

I noticed the isotopes were recently added in #3542 , perhaps we could overwrite the __iter__ method to skip isotopes when enumerating through Element #4180.
But yes this is a breaking change (also might be called a fix? if someone expect not to include the isotopes when looping through Element like me), and it looks like this enumeration behaviour is not covered by tests and we could enhance test here.

+ class CustomEnumMeta(EnumMeta):
+    def __iter__(cls):
+        named_isotopes = {"D", "T"} 
+        return (member for member in cls.__members__.values() if member.name not in named_isotopes)

- class Element(ElementBase):
+ class Element(ElementBase, metaclass=CustomEnumMeta):

[DanielYang59]

Sorry for duplicating the same topic, I would close this and prefer #4180