agente con chroma direttamente

agent.py

@@ -12,6 +12,8 @@ from langchain.tools.retriever import create_retriever_tool
 from langchain_community.tools import DuckDuckGoSearchResults
 from langchain_community.vectorstores import Chroma
 import json
+import chromadb
+chromadb.config.Settings.telemetry_enabled = False
 
 
 
@@ -114,13 +116,35 @@ sys_msg = SystemMessage(content=system_prompt)
 
 
 
+with open('metadata.jsonl', 'r') as jsonl_file:
+        json_list = list(jsonl_file)
+
+json_QA = []
+for json_str in json_list:
+    json_data = json.loads(json_str)
+    json_QA.append(json_data)
+
 # Usa gli stessi embeddings
 embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
 
-# Carica il vector store salvato precedentemente
-vector_store = Chroma(
-    embedding_function=embeddings,
-    persist_directory="./chroma_db"  # stesso path usato durante il salvataggio
+# Inizializza Chroma
+from langchain.schema import Document
+from langchain_community.vectorstores import Chroma
+
+# Prepara la lista di documenti
+docs = []
+for sample in json_QA:
+    print(len(docs))
+    content = f"Question : {sample['Question']}\n\nFinal answer : {sample['Final answer']}"
+    metadata = {"source": sample['task_id']}
+    doc = Document(page_content=content, metadata=metadata)
+    docs.append(doc)
+print('fatto')
+# Inizializza il vector store Chroma
+vector_store = Chroma.from_documents(
+    documents=docs,
+    embedding=embeddings,
+    persist_directory="./chroma_db" 
 )
 
 # Crea il retriever tool