removed print statements and added await before the recursive_query_engine

app.py

@@ -6,15 +6,10 @@ load_dotenv()
 
 import os
 import sys
-import getpass
-import nest_asyncio
-# import pandas as pd
 import faiss
 import openai
 
 import chainlit as cl  # importing chainlit for our app
-# https://docs.chainlit.io/api-reference/step-class#update-a-step
-# DEPRICATED: from chainlit.prompt import Prompt, PromptMessage  # importing prompt tools
 
 import llama_index
 from llama_index.core import Settings
@@ -28,10 +23,6 @@ from llama_index.embeddings.openai import OpenAIEmbedding
 from llama_index.postprocessor.flag_embedding_reranker import FlagEmbeddingReranker
 from llama_parse import LlamaParse
 
-from openai import AsyncOpenAI  # importing openai for API usage
-
-# os.environ["CUDA_VISIBLE_DEVICES"] = ""
-# GET KEYS
 LLAMA_CLOUD_API_KEY= os.getenv('LLAMA_CLOUD_API_KEY')
 OPENAI_API_KEY=os.getenv("OPENAI_API_KEY")
 
@@ -41,8 +32,6 @@ os.environ["OPENAI_API_KEY"] = getpass.getpass("OpenAI API Key:")
 # os.environ["WANDB_API_KEY"] = getpass.getpass("WandB API Key: ")
 """
 
-# nest_asyncio.apply()
-
 # PARSING the pdf file
 parser = LlamaParse(
     result_type="markdown",
@@ -53,7 +42,7 @@ parser = LlamaParse(
 
 nvidia_docs = parser.load_data(["./nvidia_2tables.pdf"])
 # Note: nvidia_docs contains only one file (it could contain more). nvidia_docs[0] is the pdf we loaded.
-print(nvidia_docs[0].text[:1000])
+# print(nvidia_docs[0].text[:1000])
 
 # Getting Settings out of llama_index.core which is a major part of their v0.10 update!
 Settings.llm = OpenAI(model="gpt-3.5-turbo")
@@ -65,11 +54,11 @@ node_parser = MarkdownElementNodeParser(llm=OpenAI(model="gpt-3.5-turbo"), num_w
 
 nodes = node_parser.get_nodes_from_documents(documents=[nvidia_docs[0]])
 # Let's see what's in the metadata of the nodes:
-for nd in nodes:
-  print(nd.metadata)
-  for k,v in nd:
-    if k=='table_df':
-      print(nd)
+# for nd in nodes:
+#   print(nd.metadata)
+#   for k,v in nd:
+#     if k=='table_df':
+#       print(nd)
 # Now we extract our `base_nodes` and `objects` to create the `VectorStoreIndex`.
 base_nodes, objects = node_parser.get_nodes_and_objects(nodes)
 
@@ -102,11 +91,6 @@ recursive_query_engine = recursive_index_faiss.as_query_engine(
     verbose=True
 )
 
-"""
-# Create pandas dataframe to store query+generated response+added truth
-columns=["Query", "Response", "Truth"]
-gen_df = pd.DataFrame(columns=columns,dtype='str')
-"""
 
 # ChatOpenAI Templates
 system_template = """Use the following pieces of context to answer the user's question.
@@ -119,26 +103,13 @@ user_template = """ Think through your response step by step."""
 
 #user_query = "Who are the E-VP, Operations - and how old are they?"
 
-#response = recursive_query_engine.query(system_template + user_query + user_template)
-
-#str_resp ="{}".format(response)
-
-
-def retriever_resp(prompt):
-    import time
-    response = "this is my response"
-    time.sleep(5)
-    return response
-
 @cl.on_message  # marks a function that should be run each time the chatbot receives a message from a user
 async def main(message: cl.Message):
     settings = cl.user_session.get("settings")
 
     user_query = message.content
-#    prompt = system_template+user_query+user_template
-    response = recursive_query_engine.query(system_template + user_query + user_template)
-#    response = retriever_resp(prompt)
-#    print("AAA",user_query)
+    
+    response = await recursive_query_engine.query(system_template + user_query + user_template)
     str_resp ="{}".format(response)
     msg = cl.Message(content= str_resp)
     await msg.send()