app.py

import gradio as gr
from typing import TypedDict, Annotated
from huggingface_hub import InferenceClient, login
import random
from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFacePipeline
#from langchain.schema import AIMessage, HumanMessage
from langgraph.graph.message import add_messages
from langchain_core.messages import AnyMessage, HumanMessage, AIMessage
from langchain.tools import Tool
import os
import datasets
from langchain.docstore.document import Document
from langgraph.graph import START, StateGraph
from langchain_community.retrievers import BM25Retriever
from retriever import extract_text

HUGGINGFACEHUB_API_TOKEN = os.environ["HUGGINGFACEHUB_API_TOKEN"]
login(token=HUGGINGFACEHUB_API_TOKEN)

# Load the dataset
guest_dataset = datasets.load_dataset("agents-course/unit3-invitees", split="train")

# Convert dataset entries into Document objects
docs = [
    Document(
        page_content="\n".join([
            f"Name: {guest['name']}",
            f"Relation: {guest['relation']}",
            f"Description: {guest['description']}",
            f"Email: {guest['email']}"
        ]),
        metadata={"name": guest["name"]}
    )
    for guest in guest_dataset
]


bm25_retriever = BM25Retriever.from_documents(docs)


llm = HuggingFaceEndpoint(
    repo_id="HuggingFaceH4/zephyr-7b-beta",
    task="text-generation",
    max_new_tokens=512,
    do_sample=False,
    repetition_penalty=1.03,
)

model = ChatHuggingFace(llm=llm, verbose=True)

def predict(message, history):
    history_langchain_format = []
    for msg in history:
        if msg['role'] == "user":
            history_langchain_format.append(HumanMessage(content=msg['content']))
        elif msg['role'] == "assistant":
            history_langchain_format.append(AIMessage(content=msg['content']))
    history_langchain_format.append(HumanMessage(content=message))
    gpt_response = model.invoke(history_langchain_format)
    return gpt_response.content

demo = gr.ChatInterface(
    predict,
    type="messages"
)

# setup agents

guest_info_tool = Tool(
    name="guest_info_retriever",
    func=extract_text,
    description="Retrieves detailed information about gala guests based on their name or relation."
)

tools = [guest_info_tool]
chat_with_tools = model.bind_tools(tools)


# Generate the AgentState and Agent graph
class AgentState(TypedDict):
    messages: Annotated[list[AnyMessage], add_messages]

def assistant(state: AgentState):
    return {
        "messages": [chat_with_tools.invoke(state["messages"])],
    }

## The graph
builder = StateGraph(AgentState)

demo.launch()