updated index

.vscode/launch.json

@@ -0,0 +1,16 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Python: Current File",
+            "type": "python",
+            "request": "launch",
+            "program": "${file}",
+            "console": "integratedTerminal",
+            "justMyCode": true
+        }
+    ]
+}

app_pinecode.py

@@ -25,8 +25,8 @@ def search(question, retriever_top_k, reader_top_k):
   result = []
   for i in range(0, len(query_result)):
     item = query_result[i]
-    result.append([[i+1], item.answer, item.context[:200], item.meta['title'], item.meta['source'], int(item.meta['page'])])
-    #result.append([[i+1], item.answer, item.context[:200], item.meta['title']])
+    #result.append([[i+1], item.answer, item.context[:200], item.meta['title'], item.meta['source'], int(item.meta['page'])])
+    result.append([[i+1], item.answer, item.context[:200], item.meta['title']])
   
   return result
 
@@ -50,6 +50,8 @@ if __name__ == "__main__":
         # Every form must have a submit button.
         submitted = st.form_submit_button("Buscar")
     
+    results = search("que es el adres", retriever_top_k=5, reader_top_k=3)
+
     # on submit we execute search
     if(submitted):
         # set start time

config.py

@@ -2,7 +2,7 @@ class Config():
     es_host = "saimon-askwdemocracy.es.us-central1.gcp.cloud.es.io"
     es_user = "elastic"
     es_password = "53f2a7a9-ea9d-4fd2-a8bc-f471b67f0262"
-    proposals_index = "props"
+    proposals_index = "semantic-text-search"
     reader_model_name_or_path = "deepset/xlm-roberta-base-squad2-distilled"
     #reader_model_name_or_path = "deepset/xlm-roberta-base-squad2"
     use_gpu = True

pinecode_quieries.py

@@ -34,20 +34,23 @@ class PinecodeProposalQueries(DocumentQueries):
         self.document_store = PineconeDocumentStore(
             api_key=es_password,
             environment = "us-east1-gcp",
-            index="props",
+            index="semantic-text-search",
             similarity="cosine",
-            embedding_dim=768
+            embedding_dim=384
         )        
         #self.retriever = BM25Retriever(document_store = self.document_store)
         self.retriever = EmbeddingRetriever(
             document_store=self.document_store,
-            embedding_model="multi-qa-distilbert-dot-v1",
+            #embedding_model="multi-qa-distilbert-dot-v1",
+            embedding_model = "sentence-transformers/msmarco-MiniLM-L6-cos-v5",
             model_format="sentence_transformers"
         )
-        #self.document_store.update_embeddings(self.retriever, batch_size=16)
+        #self.document_store.update_embeddings(self.retriever, update_existing_embeddings=False)
         self.pipe = ExtractiveQAPipeline(self.reader, self.retriever)
 
     def search_by_query(self, query : str, retriever_top_k: int, reader_top_k: int, es_index: str = None) :
+        #self.document_store.update_embeddings(self.retriever, update_existing_embeddings=False)
+
         #self.log.write_log(query, "hfspace-informecomision")
         #if es_index is not None:
             #self._initialize_pipeline(self.es_host, es_index, self.es_user, self.es_password)