Module seq_match 

Approximate sequence matching for license detection.

This module implements sequence-based matching using set similarity for candidate selection, followed by sequence alignment to find matching blocks.

Based on Python ScanCode Toolkit implementation at: reference/scancode-toolkit/src/licensedcode/match_seq.py

Near-Duplicate Detection

This module implements Phase 2 of Python’s 3-phase matching pipeline:

1. Phase 1: Hash & Aho-Corasick (exact matches)
2. Phase 2: Near-duplicate detection - check whole file for high-resemblance candidates
3. Phase 3: Query run matching (if no near-duplicates found)

The near-duplicate detection finds rules with high resemblance (>= 0.8) to the entire query, which helps match combined rules instead of partial rules.

Constants

HIGH_RESEMBLANCE_THRESHOLD_TENTHS

MATCH_SEQ

MAX_NEAR_DUPE_CANDIDATES

Default number of top near-duplicate candidates to consider.