enhancement: normalize Unicode in token system for port name matching

[rororowyourboat]

Problem

packages/gds-framework/gds/types/tokens.py — The tokenizer lowercases port names but does not perform Unicode normalization. Different Unicode representations of the same character (e.g., "é" NFC vs NFD) will tokenize differently and fail to match during auto-wiring.

Impact

Low — edge case for non-ASCII port names. But could cause mysterious auto-wiring failures that are very hard to debug.

Fix

Add unicodedata.normalize("NFC", ...) before lowercasing in tokenize().