Use chainlit to index our codebases (PMAT, PMA) and provide chatbot-based documentation

[kongzii]

Personally, I think that having proper documentation and tutorial is key, but this could provide a low-hanging-fruit kind of way to let users know how to use our tools + it can be automatically re-indexed.

https://github.com/Chainlit/chainlit

[kongzii]

I gave it a quick try. It's not bad, but would need more time to get it working fully. Also, writing more docstrings across the codebases would help a lot:

Code to reproduce:

import os
import openai
import chainlit as cl

from llama_index.core import (
    Settings,
    StorageContext,
    VectorStoreIndex,
    load_index_from_storage,
)
from llama_index.llms.openai import OpenAI
from llama_index.embeddings.openai import OpenAIEmbedding
from llama_index.core.query_engine.retriever_query_engine import RetrieverQueryEngine
from dotenv import load_dotenv
from llama_index.core import VectorStoreIndex
from llama_index.readers.github import GithubRepositoryReader, GithubClient
import os

load_dotenv()

Settings.llm = OpenAI(
    model="gpt-4o",
    temperature=0.1,
    max_tokens=4096,
    streaming=True,
    system_prompt="""You are assistant helping to build prediction market agents.
Agents are based on the PMAT (prediction-market-agent-tooling) library with implementations in PMA (prediction-market-agent repository).
Assume that any question is in context of prediction market agent building.
""",
)
Settings.embed_model = OpenAIEmbedding(model="text-embedding-3-large")
Settings.context_window = 8196

try:
    storage_context = StorageContext.from_defaults(persist_dir="./storage")
    index = load_index_from_storage(storage_context)
except:
    github_token = os.environ.get("GITHUB_TOKEN")
    owner = "gnosis"

    github_client = GithubClient(github_token=github_token)

    documents_pmat = GithubRepositoryReader(
        github_client=github_client,
        owner=owner,
        repo="prediction-market-agent-tooling",
        filter_directories=(
            [
                "prediction_market_agent_tooling",
                "tests",
                "tests_integration",
                "tests_integration_with_local_chain",
            ],
            GithubRepositoryReader.FilterType.INCLUDE,
        ),
    ).load_data(branch="main")
    documents_pma = GithubRepositoryReader(
        github_client=github_client,
        owner=owner,
        repo="prediction-market-agent",
        filter_directories=(
            [
                "prediction_market_agent",
                "tests",
            ],
            GithubRepositoryReader.FilterType.INCLUDE,
        ),
    ).load_data(branch="main")

    all_documents = documents_pmat + documents_pma

    index = VectorStoreIndex.from_documents(all_documents)
    index.storage_context.persist("./storage")


@cl.on_chat_start
async def start():
    query_engine = index.as_query_engine(
        streaming=True,
        similarity_top_k=5,
    )
    cl.user_session.set("query_engine", query_engine)

    await cl.Message(
        author="Assistant", content="Hello! Im an AI assistant. How may I help you?"
    ).send()


@cl.on_message
async def main(message: cl.Message):
    query_engine = cl.user_session.get("query_engine")  # type: RetrieverQueryEngine

    msg = cl.Message(content="", author="Assistant")

    res = await cl.make_async(query_engine.query)(message.content)

    for token in res.response_gen:
        await msg.stream_token(token)
    await msg.send()