Use chatgpt

app.py

@@ -1,18 +1,20 @@
 import os
 import streamlit as st
-from llama_index import GPTSimpleVectorIndex, SimpleDirectoryReader
+from llama_index import GPTSimpleVectorIndex, SimpleDirectoryReader, ServiceContext
+from llama_index.llm_predictor.chatgpt import ChatGPTLLMPredictor
 
 index_name = "./index.json"
 documents_folder = "./documents"
 
-
 @st.cache_resource
 def initialize_index(index_name, documents_folder):
+    llm_predictor = ChatGPTLLMPredictor()
+    service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor)
     if os.path.exists(index_name):
-        index = GPTSimpleVectorIndex.load_from_disk(index_name)
+        index = GPTSimpleVectorIndex.load_from_disk(index_name, service_context=service_context)
     else:
         documents = SimpleDirectoryReader(documents_folder).load_data()
-        index = GPTSimpleVectorIndex.from_documents(documents)
+        index = GPTSimpleVectorIndex.from_documents(documents, service_context=service_context)
         index.save_to_disk(index_name)
 
     return index
@@ -26,7 +28,7 @@ def query_index(_index, query_text):
 
 st.title("🦙 Llama Index Demo 🦙")
 st.header("Welcome to the Llama Index Streamlit Demo")
-st.write("Enter a query about Paul Graham's essays. You can check out the original essay [here](https://raw.githubusercontent.com/jerryjliu/llama_index/main/examples/paul_graham_essay/data/paul_graham_essay.txt)")
+st.write("Enter a query about Paul Graham's essays. You can check out the original essay [here](https://raw.githubusercontent.com/jerryjliu/llama_index/main/examples/paul_graham_essay/data/paul_graham_essay.txt). Your query will be answered using the essay as context, using embeddings from text-ada-002 and LLM completions from ChatGPT. You can read more about Llama Index and how this works in [our docs!](https://gpt-index.readthedocs.io/en/latest/index.html)")
 
 index = None
 api_key = st.text_input("Enter your OpenAI API key here:", type="password")