[acul71]

Summary

• Fixes ENCODINGS_LOOKUP mixes string and bytes keys in the same dict #44
• Introduces _ENCODINGS_BY_NAME and _ENCODINGS_BY_CODE; keeps ENCODINGS_LOOKUP as a backward-compatible mixed alias
• Name-based APIs (encode, is_encoding_supported, get_encoding_info, Encoder) only accept encoding names
• Prefix resolution stays on get_codec / decode
• Updates API docs and adds regression tests

Demo / evidence

is_encoding_supported('base16'): True
is_encoding_supported(b'f'): False
encode(b'f', ...): UnsupportedEncodingError Encoding b'f' not supported.
get_codec(b'f68656c6c6f').encoding: base16
get_codec(emoji).encoding: base256emoji

Explanation: Before this change, mixed keys made is_encoding_supported(b"f") return True. After the split, name APIs reject prefix bytes while get_codec still resolves single-byte and emoji prefixes.

Test plan

• make lint
• make typecheck
• make test
• make docs-ci
• Manual before/after snippet above

Made with Cursor

[acul71]

Demo logs (name vs prefix-code)

is_encoding_supported('base16'): True
is_encoding_supported(b'f'): False
encode(b'f', ...): UnsupportedEncodingError Encoding b'f' not supported.
get_codec(b'f68656c6c6f').encoding: base16
get_codec(emoji).encoding: base256emoji

Name-based APIs no longer treat prefix bytes as encoding names; get_codec still resolves prefixes including base256emoji.

[acul71]

@sumanjeet0012 Could you please review this PR when you have a chance? Thanks!