[BUG] Fix encode_rna docstring terminology and add return_type validation

[Ishiezz]

Reference Issues/PRs

Fixes #616

What does this implement/fix?

Two bugs fixed in pyaptamer/utils/_rna.py in the encode_rna function:

Bug 1 — Wrong docstring terminology (copy-paste from protein encoder)

The function handles RNA sequences but its docstring incorrectly referenced
"protein sequences", "amino acid patterns", and had a typo ("trunacted"):

• "tokenizes protein sequences" → "tokenizes RNA sequences"
• "amino acid patterns are preferred" → "RNA patterns are preferred"
• "trunacted or zero-padded" → "truncated or zero-padded"
• word_max_len description: "amino acid patterns" → "RNA patterns"
• Inline comment: "single protein input" → "single RNA sequence input"

Bug 2 — Invalid return_type silently returns tensor instead of raising ValueError

Before this fix:

encode_rna("ACG", {"A": 1}, max_len=3, return_type="invalid")
# returned a tensor silently — no error

After this fix:

encode_rna("ACG", {"A": 1}, max_len=3, return_type="invalid")
# raises: ValueError: `return_type` must be either 'tensor' or 'numpy', got 'invalid'.

This is now consistent with how rna2vec handles invalid sequence_type.

Did you add any tests?

Yes. Added test_encode_rna_invalid_return_type to cover Bug 2.
All 20 tests in pyaptamer/utils/tests/test_rna.py pass.

PR checklist

• The PR title starts with [BUG]
• Added/modified tests
• Used pre-commit hooks when committing