kolzchut-ragbot 1.1.3__py3-none-any.whl → 1.3.0__py3-none-any.whl

kolzchut_ragbot/engine.py

@@ -178,7 +178,7 @@ class Engine:
         all_docs_by_model = self.elastic_model.search(query_embeddings)
         all_docs = []
         ids_for_fusion = []
-        all_docs_and_scores = []
+        all_docs_and_scores = {}
 
         for key, values in all_docs_by_model.items():
             print(f"\nFound {len(values)} documents for model\n")
@@ -190,7 +190,7 @@ class Engine:
                 all_docs.append(doc)
                 scores_for_model.append({"doc": doc["_source"]["title"], "score": doc["_score"]})
             ids_for_fusion.append(model_ids)
-            all_docs_and_scores.append({f'{key}': scores_for_model})
+            all_docs_and_scores[f'{key}'] = scores_for_model
         print(f"\nFusing {len(ids_for_fusion)} results\n")
         fused_ids = self.reciprocal_rank_fusion(ids_for_fusion, k=top_k)
         top_k_documents = []

kolzchut_ragbot/model.py

@@ -93,23 +93,18 @@ class Model:
         index = index_from_page_id(int(doc_id))
 
         if update:
-            hits = []
             try:
-                existing_doc = self.es_client.search(
-                    index=index,
-                    body={
-                        "query": {
-                            "term": {f"{identifier}": {"value": doc_id}}}
-                    })
-                hits = existing_doc["hits"]["hits"]
+                query = {
+                    "query": {
+                        "match": {
+                            f"{identifier}": doc_id
+                        }
+                    }
+                }
+                self.es_client.delete_by_query(index=f"{EMBEDDING_INDEX}*", body=query)
 
             except Exception as e:
                 logging.error(f"Error while searching for existing document: {e}")
-            if len(hits) > 0:
-                print("delete existing documents")
-                for hit in hits:
-                    print("delete existing document", hit["_id"])
-                    self.es_client.delete(index=index, id=hit["_id"])
 
         for i, doc_dict in enumerate(paragraphs_dicts):
             print(f"saving paragraph {i}")

{kolzchut_ragbot-1.1.3.dist-info → kolzchut_ragbot-1.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: kolzchut-ragbot
-Version: 1.1.3
+Version: 1.3.0
 Summary: A search engine using machine learning models and Elasticsearch for advanced document retrieval.
 Home-page: https://github.com/shmuelrob/ragbot
 Author: Shmuel Robinov

{kolzchut_ragbot-1.1.3.dist-info → kolzchut_ragbot-1.3.0.dist-info}/RECORD

@@ -2,10 +2,10 @@ kolzchut_ragbot/Document.py,sha256=ySawnD06HA0zHjHp4Y_CPjMMZqLp8onaEgd1dGP5sbs,3
 kolzchut_ragbot/IntegrateService.py,sha256=CqB9vW6W5oj6Ig3crEa6hXqwro21z97UaG9ngxFTzYs,100
 kolzchut_ragbot/__init__.py,sha256=KKAc2xjCl5Aui2Cj0FWyvJ51nmnFv7MspLMqOYb-QHA,26
 kolzchut_ragbot/config.py,sha256=pcKVJVJ8P2YximjTrmVlrocHXSmzmNu_DFzNoPLa22E,138
-kolzchut_ragbot/engine.py,sha256=LQSYaDCjWZSVQui28hVWpfArkwRgxCG36BDNv9j4tTA,10162
+kolzchut_ragbot/engine.py,sha256=V8WUWyqvBWbGt-rRRf8G6BEyD-4GjsmtJrxBb6aPon8,10154
 kolzchut_ragbot/llm_client.py,sha256=q_cUZq645P7i1PliYzpJRTWlsoSECVIhE-y9wU5eRtQ,352
-kolzchut_ragbot/model.py,sha256=H-p5KQRlauBWXqFqcikjDwK8pbRmnqpqwy35ZAc8hQI,6505
-kolzchut_ragbot-1.1.3.dist-info/METADATA,sha256=sfZmgYmg12fBjViJa124hUCh1wlK6uWrzLZ8dy3Yfvo,2023
-kolzchut_ragbot-1.1.3.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-kolzchut_ragbot-1.1.3.dist-info/top_level.txt,sha256=NTZoY4GGw3v_7jm0MgcdHw8simoZ78PsR7Meqmkgd_Q,16
-kolzchut_ragbot-1.1.3.dist-info/RECORD,,
+kolzchut_ragbot/model.py,sha256=OfFFNj429IUUObOpjDB08ubZo6io45sOaoxkyEU68d4,6228
+kolzchut_ragbot-1.3.0.dist-info/METADATA,sha256=irR9JjoI8OWFxEbh-0UjEogRHooZJtZ9pxtS9JBXy7M,2023
+kolzchut_ragbot-1.3.0.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+kolzchut_ragbot-1.3.0.dist-info/top_level.txt,sha256=NTZoY4GGw3v_7jm0MgcdHw8simoZ78PsR7Meqmkgd_Q,16
+kolzchut_ragbot-1.3.0.dist-info/RECORD,,