Get Context

[d3buggerdan]

How can I print the context retrieved from a chat_engine?

@dosu

[dosubot[bot]]

To print the context retrieved from a CondensePlusContextChatEngine, you can use the _retrieve_context method. Here is an example:

# Assuming you have an instance of CondensePlusContextChatEngine named `chat_engine`
message = "Your input message here"
context_str, nodes = chat_engine._retrieve_context(message)
print(context_str)

This code snippet will print the context string that is generated based on the input message. The _retrieve_context method retrieves relevant documents and constructs the context string from them ^[1][2][3].

To continue talking to Dosu, mention @dosu.

[d3buggerdan]

@dosu

data = SimpleDirectoryReader(input_dir="./data").load_data()
index = VectorStoreIndex.from_documents(data)

llm = OpenAI(model="gpt-3.5-turbo-0125", temperature=0)
chat_engine = index.as_chat_engine(
    chat_mode="context",
    llm=llm,
    verbose=True,
)
return chat_engine

chat_engine = create_chat_engine()

[dosubot[bot]]

To create and use a CondensePlusContextChatEngine with SimpleDirectoryReader, VectorStoreIndex, and OpenAI in Python, follow these steps:

1. Install the necessary packages:

   %pip install llama-index-llms-openai

2. Download the data:

   !mkdir -p 'data/paul_graham/'
   !wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/paul_graham/paul_graham_essay.txt' -O 'data/paul_graham/paul_graham_essay.txt'

3. Set up the OpenAI API key:

   import openai
   import os
   
   os.environ["OPENAI_API_KEY"] = "sk-..."
   openai.api_key = os.environ["OPENAI_API_KEY"]

4. Load data and build the index:

   from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
   from llama_index.llms.openai import OpenAI
   
   data = SimpleDirectoryReader(input_dir="./data/paul_graham/").load_data()
   index = VectorStoreIndex.from_documents(data)

5. Configure the chat engine:

   from llama_index.core.chat_engine.condense_plus_context import CondensePlusContextChatEngine
   from llama_index.core.memory import ChatMemoryBuffer
   
   retriever = index.as_retriever()
   llm = OpenAI(model="gpt-3.5-turbo-0125", temperature=0)
   memory = ChatMemoryBuffer.from_defaults()
   
   chat_engine = CondensePlusContextChatEngine(
       retriever=retriever,
       llm=llm,
       memory=memory,
       verbose=True
   )

6. Chat with your data:

   response = chat_engine.chat("What did Paul Graham do growing up?")
   print(response.response)

7. Ask a follow-up question:

   response_2 = chat_engine.chat("Can you tell me more?")
   print(response_2.response)

8. Reset the conversation state:

   chat_engine.reset()
   response = chat_engine.chat("Hello! What do you know?")
   print(response.response)

By following these steps, you can create and use a CondensePlusContextChatEngine with SimpleDirectoryReader, VectorStoreIndex, and OpenAI in Python ^[1][2][3].

To continue talking to Dosu, mention @dosu.