app.py

pip install -qU langchain-huggingface langchain-community faiss-cpu huggingface-hub==0.27.0

import os
import getpass

# Load environment variables
load_dotenv()

YOUR_LLM_ENDPOINT_URL = "https://z1nsc3eoo5nxnoos.us-east-1.aws.endpoints.huggingface.cloud"

from langchain_huggingface import HuggingFaceEndpoint

hf_llm = HuggingFaceEndpoint(
    endpoint_url=f"{YOUR_LLM_ENDPOINT_URL}",
    task="text-generation",
    max_new_tokens=512,
    top_k=10,
    top_p=0.95,
    typical_p=0.95,
    temperature=0.01,
    repetition_penalty=1.03,
)

from langchain_core.prompts import PromptTemplate

RAG_PROMPT_TEMPLATE = """\
<|start_header_id|>system<|end_header_id|>
You are a helpful assistant. You answer user questions based on provided context. If you can't answer the question with the provided context, say you don't know.<|eot_id|>

<|start_header_id|>user<|end_header_id|>
User Query:
{query}

Context:
{context}<|eot_id|>

<|start_header_id|>assistant<|end_header_id|>
"""

rag_prompt = PromptTemplate.from_template(RAG_PROMPT_TEMPLATE)

from langchain_huggingface.embeddings import HuggingFaceEndpointEmbeddings

YOUR_EMBED_MODEL_URL = "https://jt4esmqgyp7m3fk8.us-east-1.aws.endpoints.huggingface.cloud"

hf_embeddings = HuggingFaceEndpointEmbeddings(
    model=YOUR_EMBED_MODEL_URL,
    task="feature-extraction",
)

!git clone https://github.com/dbredvick/paul-graham-to-kindle.git

from langchain_community.document_loaders import TextLoader

document_loader = TextLoader("./paul-graham-to-kindle/paul_graham_essays.txt")
documents = document_loader.load()

from langchain_text_splitters import RecursiveCharacterTextSplitter

text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=30)
split_documents = text_splitter.split_documents(documents)
len(split_documents)

from langchain_community.vectorstores import FAISS

for i in range(0, len(split_documents), 32):
  if i == 0:
    vectorstore = FAISS.from_documents(split_documents[i:i+32], hf_embeddings)
    continue
  vectorstore.add_documents(split_documents[i:i+32])

hf_retriever = vectorstore.as_retriever()

from operator import itemgetter
from langchain.schema.output_parser import StrOutputParser
from langchain.schema.runnable import RunnablePassthrough

@cl.on_chat_start
async def start_chat():
    """
    This function will be called at the start of every user session. 

    We will build our LCEL RAG chain here, and store it in the user session. 

    The user session is a dictionary that is unique to each user session, and is stored in the memory of the server.
    """
    ### BUILD LCEL RAG CHAIN THAT ONLY RETURNS TEXT
    lcel_rag_chain = {"context": itemgetter("query") | hf_retriever, "query": itemgetter("query")}| rag_prompt | hf_llm

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

@cl.on_message
async def main(message: cl.Message):
    """
    This function will be called whenever a user sends a message to the bot.
    """
    chainlit_question = message.content
    response = lcel_rag_chain.invoke({"question": chainlit_question})
    chainlit_answer = response["response"].content

    msg = cl.Message(content=chainlit_answer)
    await msg.send()