{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "0",
   "metadata": {},
   "source": [
    "# Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import argparse\n",
    "import os\n",
    "from pathlib import Path\n",
    "from typing import Annotated, List\n",
    "\n",
    "from dotenv import load_dotenv\n",
    "from langchain.tools import tool\n",
    "from langchain_core.messages import BaseMessage, HumanMessage, SystemMessage\n",
    "from langchain_openai import ChatOpenAI\n",
    "from langgraph.graph import StateGraph, START\n",
    "from langgraph.graph.message import add_messages\n",
    "from langgraph.prebuilt import ToolNode, tools_condition\n",
    "from typing_extensions import TypedDict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2",
   "metadata": {},
   "outputs": [],
   "source": [
    "from dotenv import load_dotenv\n",
    "load_dotenv()\n",
    "\n",
    "import os\n",
    "os.environ[\"LANGSMITH_TRACING\"] = \"true\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3",
   "metadata": {},
   "source": [
    "# Tools"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4",
   "metadata": {},
   "outputs": [],
   "source": [
    "def transcribe_audio(audio_path: str) -> str:\n",
    "    \"\"\"Transcribe the supplied audio file to text using the OpenAI Whisper API (``whisper-1``).\n",
    "\n",
    "    Args:\n",
    "        audio_path: The path to the audio file to transcribe.\n",
    "\n",
    "    Returns:\n",
    "        The transcribed text.\n",
    "\n",
    "    Raises:\n",
    "        RuntimeError: If the ``OPENAI_API_KEY`` environment variable is not set or the API call fails.\n",
    "    \"\"\"\n",
    "    if not Path(audio_path).exists():\n",
    "        return f\"Error: Audio file not found at {audio_path}\"\n",
    "\n",
    "    api_key = os.getenv(\"OPENAI_API_KEY\")\n",
    "    if not api_key:\n",
    "        raise RuntimeError(\"OPENAI_API_KEY environment variable is not set.\")\n",
    "\n",
    "    try:\n",
    "        from openai import OpenAI  # type: ignore\n",
    "\n",
    "        client = OpenAI(api_key=api_key)\n",
    "        with open(audio_path, \"rb\") as f:\n",
    "            transcription = client.audio.transcriptions.create(\n",
    "                model=\"whisper-1\",\n",
    "                file=f,\n",
    "            )\n",
    "        text: str | None = getattr(transcription, \"text\", None)\n",
    "        if text:\n",
    "            return text.strip()\n",
    "        raise RuntimeError(\"Transcription response did not contain text.\")\n",
    "    except Exception as exc:\n",
    "        raise RuntimeError(f\"OpenAI transcription failed: {exc}\") from exc\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5",
   "metadata": {},
   "source": [
    "# Agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6",
   "metadata": {},
   "outputs": [],
   "source": [
    "class State(TypedDict):\n",
    "    messages: Annotated[List[BaseMessage], add_messages]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7",
   "metadata": {},
   "outputs": [],
   "source": [
    "default_system_prompt = Path(\"system_promt.txt\").read_text(encoding=\"utf-8\") \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8",
   "metadata": {},
   "outputs": [],
   "source": [
    "nebius_api_key = os.environ.get(\"NEBIUS_API_KEY\")\n",
    "llm = ChatOpenAI(\n",
    "            model=\"Qwen/Qwen3-14B\",\n",
    "            api_key=nebius_api_key,\n",
    "            base_url=\"https://api.studio.nebius.com/v1/\",\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9",
   "metadata": {},
   "outputs": [],
   "source": [
    "tools = [transcribe_audio]\n",
    "llm_with_tools = llm.bind_tools(tools)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10",
   "metadata": {},
   "source": [
    "## Nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "11",
   "metadata": {},
   "outputs": [],
   "source": [
    "def assistant_node(state: State):\n",
    "    \"\"\"The assistant node in the graph. It calls the LLM with the current state\n",
    "    to decide the next action (respond or call a tool).\n",
    "    \"\"\"\n",
    "    messages = state[\"messages\"]\n",
    "    system_message = SystemMessage(content=default_system_prompt)\n",
    "\n",
    "    if not messages or not isinstance(messages[0], SystemMessage):\n",
    "        messages.insert(0, system_message)\n",
    "\n",
    "    response = llm_with_tools.invoke(messages)\n",
    "    return {\"messages\": [response]}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12",
   "metadata": {},
   "source": [
    "## Graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "13",
   "metadata": {},
   "outputs": [],
   "source": [
    "graph_builder = StateGraph(State)\n",
    "graph_builder.add_node(\"assistant\", assistant_node)\n",
    "graph_builder.add_node(\"tools\", ToolNode(tools))\n",
    "\n",
    "graph_builder.add_edge(START, \"assistant\")\n",
    "graph_builder.add_conditional_edges(\"assistant\", tools_condition)\n",
    "graph_builder.add_edge(\"tools\", \"assistant\")\n",
    "\n",
    "graph = graph_builder.compile()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "14",
   "metadata": {},
   "outputs": [],
   "source": [
    "# visualise\n",
    "from IPython.display import Image, display\n",
    "\n",
    "try:\n",
    "    display(Image(graph.get_graph().draw_mermaid_png()))\n",
    "except Exception:\n",
    "    # This requires some extra dependencies and is optional\n",
    "    pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "15",
   "metadata": {},
   "outputs": [],
   "source": [
    "question_text = \"What are the filling ingredients in the audio file?\"\n",
    "audio_path    = \"questions/files/99c9cc74-fdc8-46c6-8f8d-3ce2d3bfeea3.mp3\"\n",
    "\n",
    "answer = graph.invoke({\"messages\": question_text, \"audio_path\": audio_path})\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "16",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Show the messages\n",
    "for m in answer['messages']:\n",
    "    m.pretty_print()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}