Update app.py

app.py

@@ -17,6 +17,7 @@ RETRIEVER_URL = os.getenv("RETRIEVER_URL")
 RANKER_URL = os.getenv("RANKER_URL")
 HF_TOKEN = os.getenv("HF_TOKEN")
 
+
 class Retriever(EmbeddingRetriever):
     def __init__(
         self,
@@ -119,19 +120,30 @@ EXAMPLES = [
     "The Sphinx is in Egypt.",
 ]
 
-if os.path.exists("faiss_document_store.db") and os.path.exists("faiss_index"):
-    document_store = FAISSDocumentStore.load("faiss_index")
+if (
+    os.path.exists("./data/faiss_document_store.db")
+    and os.path.exists("./data/faiss_index.json")
+    and os.path.exists("./data/faiss_index")
+):
+    document_store = FAISSDocumentStore.load("./data/faiss_index")
     retriever = Retriever(
         document_store=document_store, top_k=TOP_K, batch_size=BATCH_SIZE
     )
+    document_store.update_embeddings(retriever=retriever)
+    document_store.save(index_path="./data/faiss_index")
 else:
     try:
-        os.remove("faiss_index")
-        os.remove("faiss_document_store.db")
+        os.remove("./data/faiss_index")
+        os.remove("./data/faiss_index.json")
+        os.remove("./data/faiss_document_store.db")
     except FileNotFoundError:
         pass
 
-    document_store = FAISSDocumentStore(embedding_dim=384, return_embedding=True)
+    document_store = FAISSDocumentStore(
+        sql_url="sqlite:///data/faiss_document_store.db",
+        return_embedding=True,
+        embedding_dim=384,
+    )
     document_store.write_documents(
         [Document(content=d, id=i) for i, d in enumerate(EXAMPLES)]
     )
@@ -139,7 +151,7 @@ else:
         document_store=document_store, top_k=TOP_K, batch_size=BATCH_SIZE
     )
     document_store.update_embeddings(retriever=retriever)
-    document_store.save(index_path="faiss_index")
+    document_store.save(index_path="./data/faiss_index")
 
 ranker = Ranker()
 
@@ -150,10 +162,8 @@ pipe.add_node(component=ranker, name="Ranker", inputs=["Retriever"])
 
 def run(query: str) -> dict:
     output = pipe.run(query=query)
-
-    return (
-        f"Closest ({TOP_K}) document(s): {[output['documents'][i].content for i in range(TOP_K)]}"
-    )
+    closest_documents = [d.content for d in output["documents"]]
+    return f"Closest ({TOP_K}) document(s): {closest_documents}"
 
 
 run("What is the capital of France?")