fix(utils): assign unique chain IDs when parsing multi-record FASTAs

[omar-A-hassan]

Summary of Changes

Fixes a silent chain-truncation bug in parse_fasta (src/simplefold/utils/fasta_utils.py).

Currently, parse_fasta hardcodes "id": "A" for every record in a multi-record FASTA:

for seq_record in records:
    molecule = {
        "protein": {
            "id": "A",  # Hardcoded default
            ...
        }
    }

When parsing a multi-record FASTA (e.g., Heavy Chain + Light Chain), parse_boltz_schema stores parsed chains in a dictionary keyed by "id". Because every record gets id: "A", subsequent records overwrite preceding records under key "A". This causes parse_fasta to silently drop all chains except the last record without raising an error.

Proposed Fix

Enumerate record indices and assign unique chain identifiers ("A", "B", "C", ...):

for i, seq_record in enumerate(records):
    seq = str(seq_record.seq)
    chain_id = chr(ord("A") + i) if i < 26 else f"A{i}"

Verification

Verified using parse_fasta on a 2-record FASTA (Heavy + Light chains):

• Before Fix: Parsed 1 chain (Chain A, 23 residues). Heavy chain dropped.
• After Fix: Parsed 2 chains (Chain A, 26 residues; Chain B, 23 residues).