BUG: fix regex numeric group replacement in PyArrow string arrays

[zishan044]

• closes BUG: Series.str.replace stopped working with regex groups #62653
• Tests added and passed
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file

Problem

PyArrow string arrays incorrectly raised NotImplementedError for numeric group references (\1, \2) in str.replace, even though PyArrow supports them. This caused failures when using regex patterns with group references on PyArrow-backed string arrays.

Solution

Modified the condition in pandas/core/arrays/_arrow_string_mixins.py to only block named group references (\g<name>) while allowing numeric group references.

Changes Made

• Removed the blanket prohibition on all backslash+digit patterns in replacement strings
• Only block named group references (detected by r"\g<" in repl)
• All other unsupported features (callable repl, case=False, flags != 0, re.Pattern objects) still properly raise NotImplementedError

Testing

Manually verified the fix works with:

• Basic numeric group replacement: s.str.replace(r'\[(\d+)\]', r'(\1)', regex=True)
• Multiple numeric groups: s.str.replace(r'(\w+)\[(\d+)\](\w+)', r'\1-\2-\3', regex=True)
• Confirmed named groups still properly raise NotImplementedError

Example

# This now works (was broken before):
s = pd.Series(["var.one[0]", "var.two[1]"]).convert_dtypes(dtype_backend="pyarrow")
result = s.str.replace(r'\[(\d+)\]', r'(\1)', regex=True)
# result: ["var.one(0)", "var.two(1)"]

# This still properly fails (named groups not supported):
s.str.replace(r'\[(?P<digit>\d+)\]', r'(\g<digit>)', regex=True)  # Raises NotImplementedError

Note: I haven't added formal tests yet due to local environment issues with the test framework, but I've manually verified the fix works. I'm happy to add tests if maintainers can help me resolve the test environment setup.

[Alvaro-Kothe]

I haven't added formal tests yet due to local environment issues with the test framework

Most of the information needed to setup the development environment is available here: https://pandas.pydata.org/docs/dev/development/contributing_environment.html

What are you having difficulty with?