Update chatpdf.py

chatpdf.py

@@ -421,14 +421,13 @@ class Rag:
         sim_contents = self.sim_model.most_similar(query, topn=self.similarity_top_k)
         # Get reference results from corpus
         hit_chunk_dict = dict()
-        threshold_score = 0.5  # Establece un umbral para filtrar fragmentos irrelevantes
-
-        for query_id, id_score_dict in sim_contents.items():
-            for corpus_id, s in id_score_dict.items():
-                if s > threshold_score:  # Filtrar por puntuación de similitud
-                    hit_chunk = self.sim_model.corpus[corpus_id]
-                    reference_results.append(hit_chunk)
-                    hit_chunk_dict[corpus_id] = hit_chunk
+
+        for c in sim_contents:
+            for id_score_dict in c:
+                corpus_id = id_score_dict['corpus_id']
+                hit_chunk = id_score_dict["corpus_doc"]
+                reference_results.append(hit_chunk)
+                hit_chunk_dict[corpus_id] = hit_chunk
 
         if reference_results:
             if self.rerank_model is not None: