Normalize hex codes in unicode escape sequences in f-strings (#4523) - #5526

Closes #4523

Summary

When PR #4401 prevented quotes from normalizing inside f-strings containing backslashes (to avoid syntax errors like f"{'\''}" turning into f"{'''}"), it bypassed visit_STRING, completely suppressing normalize_unicode_escape_sequences() for f-strings.

This PR introduces normalize_fstring_unicode_escape_sequences() to safely normalize hex codes in unicode escape sequences (e.g. \xFF -> \xff, \u00A0 -> \u00a0, \N{...} -> uppercase) within f-strings and t-strings while preserving AST equivalence:

• Raw f-strings: Skips template parts when the f-string prefix contains r/R (e.g. rf"...").
• Raw string literals in replacement fields: Preserves expressions like f"{r'\xFF'}" without mutating escaped characters, preventing runtime value changes and ASTSafetyError.
• Debug f-strings: Leaves expressions preceding = intact (e.g. f"{'\xFF'=}") since Python prints the literal source representation before = at runtime.
• Nested templates: Recursively handles nested f-strings and t-strings.

Test Plan

• Added cases to pep_701.py covering raw f-strings, nested f-strings, debug f-strings (=), raw string literals in fields, \N{...}, \u..., and \U....
• Added test case to format_unicode_escape_seq.py.
• Verified full test suite (pytest) and linters (flake8, black --check).