test_agent.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Agent Testing Notebook\n",
    "\n",
    "This notebook is designed to test the `Agent` on a single question fetched from the API. You can select which question to run by changing the `QUESTION_INDEX_TO_TEST` variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Optional: Install necessary packages if you haven't already\n",
    "# !pip install python-dotenv langchain langchain_openai langchain_community langchain-tavily langgraph pandas requests openai"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import requests\n",
    "import json\n",
    "from pathlib import Path\n",
    "from dotenv import load_dotenv\n",
    "\n",
    "# Add the project root to the Python path to allow importing from 'agent.py'\n",
    "project_root = Path.cwd()\n",
    "if str(project_root) not in sys.path:\n",
    "    sys.path.append(str(project_root))\n",
    "\n",
    "from agent import Agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load environment variables from a .env file in the project root\n",
    "load_dotenv()\n",
    "\n",
    "# --- Langsmith Tracing Setup ---\n",
    "# Set the following environment variables to enable Langsmith tracing.\n",
    "# You can set them in your .env file or uncomment the lines below and add your keys.\n",
    "# os.environ[\"LANGSMITH_TRACING\"] = \"true\"\n",
    "# os.environ[\"LANGSMITH_API_KEY\"] = \"YOUR_LANGSMITH_API_KEY\"\n",
    "# os.environ[\"OPENAI_API_KEY\"] = \"YOUR_OPENAI_API_KEY\"\n",
    "# os.environ[\"TAVILY_API_KEY\"] = \"YOUR_TAVILY_API_KEY\"\n",
    "# os.environ[\"NEBIUS_API_KEY\"] = \"YOUR_NEBIUS_API_KEY\"\n",
    "\n",
    "# Verify that the necessary keys are set\n",
    "required_keys = [\"NEBIUS_API_KEY\", \"OPENAI_API_KEY\", \"TAVILY_API_KEY\", \"LANGSMITH_API_KEY\"]\n",
    "print(\"Checking for required environment variables...\")\n",
    "for key in required_keys:\n",
    "    if not os.getenv(key):\n",
    "        print(f\"⚠️  Warning: Environment variable {key} is not set. The agent might not work correctly.\")\n",
    "    else:\n",
    "        print(f\"✅ {key} is set.\")\n",
    "\n",
    "print(\"Environment variables loaded and Langsmith tracing enabled.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# --- 1. Fetch Questions from API ---\n",
    "API_URL = \"https://agents-course-unit4-scoring.hf.space\"\n",
    "QUESTIONS_URL = f\"{API_URL}/questions\"\n",
    "\n",
    "print(f\"Fetching questions from {QUESTIONS_URL}...\")\n",
    "try:\n",
    "    response = requests.get(QUESTIONS_URL, timeout=15)\n",
    "    response.raise_for_status()\n",
    "    questions_data = response.json()\n",
    "    print(f\"Successfully fetched {len(questions_data)} questions.\")\n",
    "    # Save to a local file for easy inspection\n",
    "    with open(\"questions.json\", \"w\") as f:\n",
    "        json.dump(questions_data, f, indent=4)\n",
    "    print(\"Questions also saved to questions.json\")\n",
    "except Exception as e:\n",
    "    print(f\"Failed to fetch questions: {e}\")\n",
    "    questions_data = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# --- 2. Display Questions for Selection ---\n",
    "if questions_data:\n",
    "    for i, q in enumerate(questions_data):\n",
    "        print(f\"--- Question {i} ---\")\n",
    "        print(f\"  ID: {q.get('task_id')}\")\n",
    "        print(f\"  Question: {q.get('question')[:300]}...\")\n",
    "        if q.get('file_name'):\n",
    "            print(f\"  File: {q.get('file_name')}\")\n",
    "        print(\"-\" * (len(str(i)) + 14))\n",
    "else:\n",
    "    print(\"No questions to display.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# --- 3. Select a Question to Test ---\n",
    "# Change the index number to test a different question.\n",
    "# The question about the pie recipe is usually at index 4.\n",
    "QUESTION_INDEX_TO_TEST = 16  # <--- CHANGE THIS VALUE\n",
    "\n",
    "print(f\"Will test question at index: {QUESTION_INDEX_TO_TEST}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# --- 4. Run the Agent on the Selected Question ---\n",
    "if not questions_data:\n",
    "    print(\"Cannot run agent, no questions were fetched.\")\n",
    "elif not 0 <= QUESTION_INDEX_TO_TEST < len(questions_data):\n",
    "    print(f\"Invalid index: {QUESTION_INDEX_TO_TEST}. Please choose an index between 0 and {len(questions_data) - 1}.\")\n",
    "else:\n",
    "    try:\n",
    "        # Select the question item\n",
    "        question_item = questions_data[QUESTION_INDEX_TO_TEST]\n",
    "        print(f\"Testing with question ID: {question_item.get('task_id')}\")\n",
    "        print(\"-\" * 30)\n",
    "        \n",
    "        # Instantiate the agent\n",
    "        print(\"Instantiating agent...\")\n",
    "        agent = Agent()\n",
    "        print(\"Agent instantiated successfully.\")\n",
    "        print(\"-\" * 30)\n",
    "\n",
    "        # Run the agent\n",
    "        print(\"Running agent on the question... (This may take a moment)\")\n",
    "        final_answer = agent(item=question_item, api_url=API_URL)\n",
    "        print(\"-\" * 30)\n",
    "\n",
    "        # Display the result\n",
    "        print(\"✅ Agent's Final Answer:\")\n",
    "        print(final_answer)\n",
    "\n",
    "    except Exception as e:\n",
    "        print(f\"An error occurred while running the agent: {e}\")"
   ]
  }
 ],
 "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": 4
}