Fix normalizing the cosine distance in Qdrant (#12934)

Qdrant was incorrectly calculating the cosine similarity and returning
`0.0` for the best match, instead of `1.0`. Internally Qdrant returns a
cosine score from `-1.0` (worst match) to `1.0` (best match), and the
current formula reflects it.

libs/langchain/langchain/vectorstores/qdrant.py

@@ -1837,6 +1837,11 @@ class Qdrant(VectorStore):
         )
         return qdrant
 
+    @staticmethod
+    def _cosine_relevance_score_fn(distance: float) -> float:
+        """Normalize the distance to a score on a scale [0, 1]."""
+        return (distance + 1.0) / 2.0
+
     def _select_relevance_score_fn(self) -> Callable[[float], float]:
         """
         The 'correct' relevance function