Edit main device _type

ingest.py

@@ -170,6 +170,7 @@ def load_faiss_index(index_path, metadata_path):
 
 
 def main(device_type):
+    print(f"Running on device: {device_type}")
     # Load documents and split in chunks
     logging.info(f"Loading documents from {SOURCE_DIRECTORY}")
     documents = load_documents(SOURCE_DIRECTORY)
@@ -224,9 +225,13 @@ def main(device_type):
         client_settings=CHROMA_SETTINGS,
         )
 
+import argparse
 
 if __name__ == "__main__":
     logging.basicConfig(
         format="%(asctime)s - %(levelname)s - %(filename)s:%(lineno)s - %(message)s", level=logging.INFO
     )
-    main()
+    parser = argparse.ArgumentParser(description="Ingest script for localGPT")
+    parser.add_argument("--device_type", type=str, required=True, help="Device type (cpu or gpu)")
+    args = parser.parse_args()
+    main(args.device_type)

run_localGPT.py

@@ -143,19 +143,24 @@ def retrieval_qa_pipline(device_type, use_history, promptTemplate_type="llama"):
     #             embedding_function=embeddings,
     #             client_settings=CHROMA_SETTINGS)
 
+    # print(embeddings)
+
     # Initialize the FAISS index
-    # faiss_index = faiss.IndexFlatL2(embeddings)
+    faiss_index = faiss.IndexFlatL2(embeddings.embed())
+
     # # Initialize the docstore
-    # docstore = InMemoryDocstore()
+    docstore = InMemoryDocstore()
     # # Initialize the index_to_docstore_id
-    # index_to_docstore_id = {}
+    index_to_docstore_id = {}
+    # Add the embeddings to the index
+    faiss_index.add(embeddings)
 
-    # db = FAISS(
-    #             embedding_function=embeddings,
-    #             index=faiss_index,
-    #             docstore=docstore,
-    #             index_to_docstore_id=index_to_docstore_id
-    #             )
+    db = FAISS(
+                embedding_function=embeddings,
+                index=faiss_index,
+                docstore=docstore,
+                index_to_docstore_id=index_to_docstore_id
+                )
 
     # # Add documents and their embeddings to the FAISS index and the docstore
     # for i, (text, embedding) in enumerate(zip(df['Text'].tolist(), embeddings)):