fix: normalize vector store scores to similarity (higher = better)

The threshold parameter in Memory.search() was silently broken for 11
vector stores because they returned raw distance scores (lower = better)
while the threshold check assumed similarity scores (higher = better).
This caused threshold filtering to be inverted — good matches got
dropped and bad matches passed through.

Convert all affected vector stores to return similarity scores:
- Cosine distance stores: score = max(0.0, 1.0 - distance)
- L2 distance stores: score = 1.0 / (1.0 + distance)
- Stores computing similarity then discarding it: return similarity directly

Fixes #4453
Related: #3283

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

mem0/vector_stores/supabase.py

@@ -135,7 +135,8 @@ class Supabase(VectorStoreBase):
             data=vectors, limit=limit, filters=filters, include_metadata=True, include_value=True
         )
 
-        return [OutputData(id=str(result[0]), score=float(result[1]), payload=result[2]) for result in results]
+        # Convert cosine distance to similarity score (higher = more similar)
+        return [OutputData(id=str(result[0]), score=max(0.0, 1.0 - float(result[1])), payload=result[2]) for result in results]
 
     def delete(self, vector_id: str):
         """