app.py

import os
import json
from langchain_core.documents import Document
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain_community.vectorstores import FAISS
from langchain_huggingface import HuggingFaceEmbeddings
from langchain_openai import ChatOpenAI
from langchain.chains import RetrievalQA
import chainlit as cl
from typing import AsyncGenerator

from operator import itemgetter
from langchain_core.output_parsers import StrOutputParser
from langchain_core.runnables import RunnablePassthrough, RunnableParallel

# === Load and prepare data ===
with open("combined_data.json", "r") as f:
    raw_data = json.load(f)

all_docs = [
    Document(page_content=entry["content"], metadata=entry["metadata"])
    for entry in raw_data
]


from langchain.text_splitter import RecursiveCharacterTextSplitter

text_splitter = RecursiveCharacterTextSplitter(chunk_size=750, chunk_overlap=100)
split_documents = text_splitter.split_documents(all_docs)



embeddings = HuggingFaceEmbeddings(model_name="bsmith3715/legal-ft-demo_final")

from langchain_qdrant import QdrantVectorStore
from qdrant_client import QdrantClient
from qdrant_client.http.models import Distance, VectorParams

client = QdrantClient(":memory:")

client.create_collection(
    collection_name="reformer_docs",
    vectors_config=VectorParams(size=768, distance=Distance.COSINE),
)

vector_store_ft = QdrantVectorStore(
    client=client,
    collection_name="reformer_docs",
    embedding=embeddings,
)



# === Set up FAISS vector store ===
_ = vector_store_ft.add_documents(documents=split_documents)

retriever_finetune = vector_store_ft.as_retriever(search_kwargs={"k": 5})

# === Load LLM ===
llm = ChatOpenAI(model_name="gpt-4o-mini", temperature=0, stream=True)

from langchain.prompts import ChatPromptTemplate

RAG_PROMPT = """\
You are a helpful assistant who answers questions based on provided context. You must only use the provided context, and cannot use your own knowledge.

### Question
{question}

### Context
{context}
"""

rag_prompt = ChatPromptTemplate.from_template(RAG_PROMPT)

finetune_rag_chain = (
    {"context": itemgetter("question") | retriever_finetune, "question": itemgetter("question")}
    | RunnablePassthrough.assign(context=itemgetter("context"))
    | {"response": rag_prompt | llm | StrOutputParser(), "context": itemgetter("context")}
)

# === Chainlit start event ===
@cl.on_chat_start
async def start():
    await cl.Message(content = 
                     
"""👋 Welcome to your Reformer Pilates AI!
Here's what you can do:
• Ask questions about Reformer Pilates
• Get individualized workouts based on your level, goals, and equipment
• Get instant exercise modifications based on injuries or limitations
Let's get started! 🚀""").send()
    cl.user_session.set("qa_chain", finetune_rag_chain)


@cl.on_message
async def main(message):
    # Get retriever
    chain = cl.user_session.get("qa_chain")

    # Run the chain once to get context
    inputs = {"question": message.content}
    context_and_prompt = await chain.ainvoke(inputs)

    # Send a blank message to stream into
    msg = cl.Message(content="")

    # Call LLM manually with streaming
    llm = ChatOpenAI(model_name="gpt-4.1-mini", temperature=0, streaming=True)
    full_prompt = rag_prompt.format(**inputs, context=context_and_prompt["context"])

    async for chunk in llm.astream(full_prompt):
        await msg.stream_token(chunk.content)  # Only stream the text part

    await msg.send()