Create retriever.py

utils/retriever.py

@@ -1,10 +1,21 @@
-# utils/retriever.py
-
 import os
-from langchain.vectorstores import Chroma
-from langchain.embeddings import HuggingFaceEmbeddings
+from langchain_community.vectorstores import Chroma
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_core.vectorstores import VectorStoreRetriever
+
+def load_vectorstore(pdf_path: str) -> VectorStoreRetriever:
+    # Ensure Chroma store directory exists
+    folder_path = "chroma_store"
+    os.makedirs(folder_path, exist_ok=True)
 
-def load_vectorstore(folder_path="data/vectorstore"):
+    # Use a local embedding model (no API key needed)
     embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
-    vectordb = Chroma(persist_directory=folder_path, embedding_function=embeddings)
-    return vectordb
+
+    # Initialize Chroma without deprecated Settings
+    vectordb = Chroma(
+        persist_directory=folder_path,
+        embedding_function=embeddings
+    )
+
+    # Return retriever
+    return vectordb.as_retriever()