feat: Add semantic deduplication and threshold configuration for improved query filtering

2026-01-19 05:03:57 +00:00 · 2025-11-03 21:43:21 +01:00
parent a1163976a6
commit 3eb861ef8a
5 changed files with 93 additions and 19 deletions
--- a/src/config.jsonc
+++ b/src/config.jsonc
@@ -48,6 +48,8 @@
            "scrollRandomResults": true,
            "clickRandomResults": true,
            "retryMobileSearchAmount": 2,
+            "semanticDedup": true, // Filter queries with high word similarity (Jaccard). Reduces repetitive patterns.
+            "semanticDedupThreshold": 0.65, // Similarity threshold (0-1). Lower = more strict filtering.
            "delay": {
                "min": "3min",
                "max": "5min"