Log the similarity scores in verbose mode when the embeddings_only_similarity_threshold is used.

Author: drazvan

nemoguardrails/embeddings/basic.py

@@ -14,6 +14,7 @@
 # limitations under the License.
 
 import asyncio
+import logging
 from typing import Any, Dict, List, Optional, Union
 
 from annoy import AnnoyIndex
@@ -23,6 +24,8 @@
 from nemoguardrails.embeddings.providers import EmbeddingModel, init_embedding_model
 from nemoguardrails.rails.llm.config import EmbeddingsCacheConfig
 
+log = logging.getLogger(__name__)
+
 
 class BasicEmbeddingsIndex(EmbeddingsIndex):
  """Basic implementation of an embeddings index.
@@ -286,6 +289,14 @@ async def search(
  include_distances=True,
  )
 
+ # In verbose mode, we show detailed info about the scores
+ if threshold != float("inf"):
+ log_items = []
+ for i in range(len(results[0])):
+ score = 1 - results[1][i] / 2
+ log_items.append((score, self._items[results[0][i]].text))
+ log.info("Similarity scores :: %s", str(log_items))
+
  filtered_results = self._filter_results(results[0], results[1], threshold)
 
  return [self._items[i] for i in filtered_results]