app.py

from llama_index.core import SimpleDirectoryReader
from llama_index.core.node_parser import SentenceSplitter
from llama_index.core import Settings
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core import  VectorStoreIndex
from langchain_groq import ChatGroq
from langchain.tools import BaseTool, StructuredTool, tool
from langchain_community.tools.tavily_search import TavilySearchResults
from typing import TypedDict ,Annotated
from langchain_core.runnables import RunnablePassthrough
from langchain_core.output_parsers import StrOutputParser
import os
from typing import TypedDict ,Annotated
from langchain_core.messages import AnyMessage,SystemMessage,HumanMessage,ToolMessage,AIMessage
import operator
from langgraph.checkpoint.memory import InMemorySaver
from langgraph.graph import StateGraph, END
from fastapi import FastAPI
import json
import shutil
import os
from fastapi import FastAPI, File, UploadFile

app = FastAPI()

@app.get("/")
def read_root():
    return {"message": "Connected"}

@tool
def retrieve(query_text):
    """
    Retrieves relevant information from a vector index based on a query.

    Parameters:
    - query_text (str): Query to search for relevant information.

    Returns:
    - str: Retrieved text from the document.
    """
    if vector_index is None:
        return "Vector index not found. Please upload a file first."
    else:
        retriever = vector_index.as_retriever(similarity_top_k=3)
        result = retriever.retrieve(query_text)
        if result:
            return "\n\n".join([node.node.text for node in result])
        return "No relevant information found."

tavily_search = TavilySearchResults(max_results=4)

vector_index = None

@app.post("/uploadpdfs")
async def upload_file(file: UploadFile = File(...)):
    global vector_index

    # Save uploaded file to a temp directory
    temp_dir = "temp_uploads"
    os.makedirs(temp_dir, exist_ok=True)
    file_id = str(uuid.uuid4())
    file_path = os.path.join(temp_dir, f"{file_id}_{file.filename}")

    with open(file_path, "wb") as f:
        shutil.copyfileobj(file.file, f)

    # Load and parse document
    documents = SimpleDirectoryReader(input_files=[file_path]).load_data()
    parser = SentenceSplitter(chunk_size=300, chunk_overlap=50)
    nodes = parser.get_nodes_from_documents(documents)

    # Create or update vector index
    embed_model = HuggingFaceEmbedding(model_name="WhereIsAI/UAE-Large-V1")
    if vector_index is None:
        vector_index = VectorStoreIndex(nodes, embed_model=embed_model)
        message = "New vector index created and file stored."
    else:
        vector_index.insert_nodes(nodes)
        message = "File stored and vector index updated."

    return {"message": message, "filename": file.filename}




class AgentState(TypedDict):
  messages: Annotated[list[AnyMessage], operator.add]

memory = InMemorySaver()

class Agent:
  def __init__(self, model, tools, checkpointer=None, system=""):
    self.system = system
    graph = StateGraph(AgentState)
    graph.add_node('llm',self.call_llm)
    graph.add_node('action',self.take_action)
    graph.add_conditional_edges("llm",self.exists_action,{True :"action",False:END})
    graph.add_edge("action","llm")
    graph.set_entry_point("llm")
    self.graph = graph.compile(checkpointer=checkpointer)
    self.tools = {t.name:t for t in tools}
    self.model = model.bind_tools(tools)

  def call_llm(self, state:AgentState):
    messages = state['messages']
    if self.system :
      messages = [SystemMessage(content=self.system)] + messages
    message = self.model.invoke(messages)
    return {"messages":[message]}

  def exists_action(self, state:AgentState):
    result = state['messages'][-1]
    return len(result.tool_calls) > 0

  def take_action(self, state:AgentState):
    tool_calls = state['messages'][-1].tool_calls
    results = []
    for t in tool_calls:
      result= self.tools[t['name']].invoke(t['args'])
      results.append(ToolMessage(tool_call_id=t['id'],name=t['name'],content=str(result)))
    return {"messages":results}


system_Prompt="""
You are an AI assistant designed to assist users with health benefits, diet, nutrition information, and recipes.
You analyze patient reports to offer guidance on self-care with AI support.
Provide answers directly related to the question, without additional explanation or unrelated information.
"""

tools=[retrieve,tavily_search]

model = ChatGroq(
    model="qwen-qwq-32b")

agent = Agent(model, tools, memory, system=system_Prompt)

thread = {"configurable": {'thread_id': '1'}}


@app.post("/askbot")
async def ask_question(query: QueryRequest):
    messages = [HumanMessage(content=query.message)]
    final_res = ""

    for event in abot.graph.stream({'messages': messages}, thread):
        for v in event.values():
            if isinstance(v, dict) and 'messages' in v:
                for msg in v['messages']:
                    if hasattr(msg, 'content') and isinstance(msg, AIMessage):
                        final_res += msg.content

    return {"answer": final_res}