app.py

from datasets import load_dataset
import pandas as pd
import os
from langchain.document_loaders import CSVLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_core.callbacks import BaseCallbackHandler
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnablePassthrough, RunnableConfig
from langchain_openai import OpenAIEmbeddings, ChatOpenAI
from langchain.embeddings import CacheBackedEmbeddings
from langchain.storage import LocalFileStore
from langchain_community.vectorstores import FAISS
import chainlit as cl

csv_path = "./imdb.csv"
if not os.path.exists(csv_path):
    dataset = load_dataset("ShubhamChoksi/IMDB_Movies")

# Convert to pandas DataFrame (assuming 'train' split)
imdb_train = dataset['train']
df = pd.DataFrame(imdb_train)

# Save to a CSV file
csv_file_path = 'imdb_movies.csv'
df.to_csv(csv_file_path, index=False)

loader = CSVLoader(csv_file_path)
data = loader.load()

# Initialize the text splitter for 1000 char chunk size and 100 char overlap
text_splitter = RecursiveCharacterTextSplitter(
    chunk_size=1000,
    chunk_overlap=100
)

documents = text_splitter.split_documents(data)

# Initialize the OpenAIEmbeddings object
#embedding_model = OpenAIEmbeddings(model="text-embedding-ada-002")
embedding_model = OpenAIEmbeddings()

store = LocalFileStore("./cache/")
embedder = CacheBackedEmbeddings.from_bytes_store(
    embedding_model, store, namespace=embedding_model.model
)

# Attempt to load the FAISS index
vector_store = None
faiss_index_path = "faiss_index"

try:
    vector_store = FAISS.load_local(faiss_index_path, embedding_model, allow_dangerous_deserialization=True)
except Exception as e:
    vector_store = FAISS.from_documents(documents, embedding_model)

vector_store.save_local("faiss_index")

#### APP ###
@cl.on_chat_start
async def on_chat_start():
    # Define the Prompt Template
    prompt_text = """
    You are an AI assistant that loves to suggest movies for people to watch. You have access to the following context:

    {documents}

    Answer the following question. If you cannot answer at all just say "Sorry! I couldn't come up with any movies for this question!"

    Question: {question}
    """

    prompt_template = ChatPromptTemplate.from_template(prompt_text)
    retriever = vector_store.as_retriever()
    chat_model = ChatOpenAI(model="gpt-4o", temperature=0, streaming=True)
    parser = StrOutputParser()

    runnable_chain = (
            {"documents": retriever, "question": RunnablePassthrough()}
            | prompt_template
            | chat_model
            | parser
    )

    cl.user_session.set("runnable_chain", runnable_chain)


# @cl.on_message
# async def on_message(message: cl.Message):
#     runnable_chain = cl.user_session.get("runnable_chain")
#
#     msg = cl.Message(content="")
#
#     async for chunk in runnable_chain.astream(
#             message.content,
#             config=RunnableConfig(callbacks=[cl.LangchainCallbackHandler()]),
#     ):
#         await msg.stream_token(chunk)
#
#     await msg.send()

@cl.on_message
async def on_message(message: cl.Message):
    runnable = cl.user_session.get("runnable_chain")
    msg = cl.Message(content="")

    # class PostMessageHandler(BaseCallbackHandler):
    #     """
    #     Callback handler for handling the retriever and LLM processes.
    #     Used to post the sources of the retrieved documents as a Chainlit element.
    #     """
    #
    #     def __init__(self, msg: cl.Message):
    #         BaseCallbackHandler.__init__(self)
    #         self.msg = msg
    #         self.sources = set()  # To store unique pairs
    #
    #     def on_retriever_end(self, documents, *, run_id, parent_run_id, **kwargs):
    #         for d in documents:
    #             source_page_pair = (d.metadata['source'], d.metadata['page'])
    #             self.sources.add(source_page_pair)  # Add unique pairs to the set
    #
    #     def on_llm_end(self, response, *, run_id, parent_run_id, **kwargs):
    #         if len(self.sources):
    #             sources_text = "\n".join([f"{source}#page={page}" for source, page in self.sources])
    #             self.msg.elements.append(
    #                 cl.Text(name="Sources", content=sources_text, display="inline")
    #             )

    async with cl.Step(type="run", name="Movie Assistant"):
        async for chunk in runnable.astream(
            message.content,
            config=RunnableConfig(callbacks=[
                cl.LangchainCallbackHandler(),
                # PostMessageHandler(msg)
            ]),
        ):
            await msg.stream_token(chunk)

    await msg.send()