experiments/EXAMPLE_EXPERIMENT.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load & Preprocess Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jypark/Desktop/weavel/Ape/venv/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "import random\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from typing import Dict, List, Any, Union\n",
    "\n",
    "from dotenv import load_dotenv\n",
    "from openai import AsyncOpenAI\n",
    "from datasets import load_dataset\n",
    "\n",
    "load_dotenv()\n",
    "\n",
    "# get dataset\n",
    "# For tutorial, we will use openai/gsm8k\n",
    "dataset = load_dataset(\"openai/gsm8k\", \"main\")\n",
    "train_split = dataset[\"train\"]\n",
    "test_split = dataset[\"test\"]\n",
    "train_split = train_split.to_list()\n",
    "test_split = test_split.to_list() \n",
    "pattern = \"####\"\n",
    "\n",
    "random.seed(42)\n",
    "random.shuffle(train_split)\n",
    "random.shuffle(test_split)\n",
    "train_split = train_split[:100]\n",
    "test_split = test_split[:100]\n",
    "\n",
    "from ape.common.types import DatasetItem\n",
    "\n",
    "trainset: List[DatasetItem] = []\n",
    "for item in train_split:\n",
    "    split_answer = item[\"answer\"].split(pattern)\n",
    "    thought = split_answer[0]\n",
    "    answer = split_answer[1]\n",
    "    trainset.append(DatasetItem(\n",
    "        inputs={\"problem\": item[\"question\"]},\n",
    "        outputs={\"thought\": thought, \"answer\": answer},\n",
    "    ))\n",
    "valset: List[DatasetItem] = []\n",
    "for item in test_split:\n",
    "    split_answer = item[\"answer\"].split(pattern)\n",
    "    thought = split_answer[0]\n",
    "    answer = split_answer[1]\n",
    "    valset.append(DatasetItem(\n",
    "        inputs={\"problem\": item[\"question\"]},\n",
    "        outputs={\"thought\": thought, \"answer\": answer},\n",
    "    ))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare Prompt to optimize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from ape.common import Prompt\n",
    "\n",
    "# define prompt\n",
    "system_prompt = \"\"\"\\\n",
    "For given math problem, solve it step by step.\n",
    "\n",
    "Question: {problem}\n",
    "\n",
    "You MUST respond in JSON format with the following fields:\n",
    "thought: the reasoning process of the problem solving.\n",
    "answer: only return the answer without any explanation.\n",
    "\"\"\"\n",
    "\n",
    "json_schema = {\n",
    "    \"type\": \"json_schema\", \n",
    "    \"json_schema\": {\n",
    "        \"name\": \"MathProblemSolving\",\n",
    "        \"strict\": True,\n",
    "        \"schema\": {\n",
    "            \"type\": \"object\",\n",
    "            \"properties\": {\n",
    "                \"thought\": {\n",
    "                    \"type\": \"string\",\n",
    "                    \"description\": \"The reasoning process of the problem solving\"\n",
    "                },\n",
    "                \"answer\": {\n",
    "                    \"type\": \"string\",\n",
    "                    \"description\": \"The answer to the question\"\n",
    "                }\n",
    "            },\n",
    "            \"required\": [\"thought\", \"answer\"],\n",
    "            \"additionalProperties\": False\n",
    "        }\n",
    "    }\n",
    "}\n",
    "\n",
    "\n",
    "student_prompt = Prompt(\n",
    "    messages=[\n",
    "        {\"role\": \"system\", \"content\": system_prompt},\n",
    "    ],\n",
    "    model=\"gpt-4o-mini\",\n",
    "    temperature=0.0,\n",
    "    name=\"student_prompt\",\n",
    "    response_format=json_schema,\n",
    ")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare Generator, Metric, and Global Metric"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from ape.common.generator import BaseGenerator\n",
    "from ape.common.metric import BaseMetric\n",
    "from ape.common.global_metric import BaseGlobalMetric\n",
    "from ape.common.types import MetricResult, GlobalMetricResult\n",
    "\n",
    "# define generator, metric, global metric\n",
    "openai = AsyncOpenAI()\n",
    "class MathSolver(BaseGenerator):\n",
    "    async def generate(\n",
    "        self,\n",
    "        prompt: Prompt,\n",
    "        inputs: Dict[str, Any],\n",
    "    ) -> Union[Dict[str, Any], str]:\n",
    "        retry_count = 0\n",
    "        messages = prompt.format(**inputs).messages\n",
    "        model = prompt.model\n",
    "        response_format = prompt.response_format\n",
    "        while retry_count < 3:\n",
    "            try:\n",
    "                response = await openai.chat.completions.create(\n",
    "                    model=model,\n",
    "                    messages=messages,\n",
    "                    response_format=response_format,\n",
    "                    temperature=0.0,\n",
    "                )\n",
    "                return json.loads(response.choices[0].message.content)\n",
    "            except Exception as e:\n",
    "                print(e)\n",
    "                retry_count += 1\n",
    "        return {\n",
    "            \"thought\": \"error\",\n",
    "            \"topic\": \"\",\n",
    "        }\n",
    "\n",
    "eval_json_schema = {\n",
    "    \"type\": \"json_schema\", \n",
    "    \"json_schema\": {\n",
    "        \"name\": \"MathProblemSolvingEvaluation\",\n",
    "        \"strict\": True,\n",
    "        \"schema\": {\n",
    "            \"type\": \"object\",\n",
    "            \"properties\": {\n",
    "                \"thought\": {\n",
    "                    \"type\": \"string\",\n",
    "                    \"description\": \"The reasoning process of the problem solving evaluation\"\n",
    "                },\n",
    "                \"correctness\": {\n",
    "                    \"type\": \"string\",\n",
    "                    \"description\": \"The correctness of the problem solving\"\n",
    "                }\n",
    "            },\n",
    "            \"required\": [\"thought\", \"correctness\"],\n",
    "            \"additionalProperties\": False\n",
    "        }\n",
    "    }\n",
    "}\n",
    "\n",
    "class MathMetricLLM(BaseMetric):\n",
    "    async def compute(\n",
    "        self,\n",
    "        dataset_item: DatasetItem,\n",
    "        pred: Dict[str, Any],\n",
    "    ) -> MetricResult:\n",
    "        openai = AsyncOpenAI()\n",
    "        try:\n",
    "            res = await openai.chat.completions.create(\n",
    "                model=\"gpt-4o-mini\",\n",
    "                messages=[\n",
    "                    {\n",
    "                        \"role\": \"system\",\n",
    "                        \"content\": \"\"\"\\\n",
    "YOU ARE one of the GREATEST mathematicians, logicians, programmers, and AI scientists. You are intelligent and rational. You are prudent and cautious. Your mastery over Arithmetic, Combinatorics, Number Theory, Probability Theory, Algebra, Analysis, and Geometry is unparalleled. You THINK NATURAL, BROAD AND DEEP. Let's think step by step. \n",
    "Your job is to judge whether the \"final_answer\" is correct based on \"ground_truth_answer\", do not be strict on the format, but check the content. Notice that unsolved half results are not Correct. \n",
    "Problem: {question_content}\n",
    "Is the final_answer correct, given the ground truth answer? Reply with Correct, Wrong or Unknown. \n",
    "\"final_answer\": \"{final_answer}\", \"ground_truth_answer\": \"{ground_truth_answer}\"\n",
    "\n",
    "You MUST respond in JSON format like below:\n",
    "{{\n",
    "    \"thought\": \"...\",\n",
    "    \"correctness\": \"<correctness>\", One of \"Correct\", \"Wrong\", \"Unknown\"\n",
    "}}\n",
    "\"\"\".format(question_content=dataset_item[\"inputs\"][\"problem\"], final_answer=pred[\"answer\"], ground_truth_answer=dataset_item[\"outputs\"][\"answer\"])\n",
    "                    }\n",
    "                ],\n",
    "                response_format=eval_json_schema,\n",
    "                temperature=0.0,\n",
    "            )\n",
    "            res_json = json.loads(res.choices[0].message.content)\n",
    "            # print(dataset_item[\"outputs\"][\"answer\"], pred[\"answer\"], res_json[\"correctness\"])\n",
    "            if res_json[\"correctness\"] == \"Correct\":\n",
    "                return MetricResult(\n",
    "                    score=1.0,\n",
    "                )\n",
    "            else:\n",
    "                return MetricResult(\n",
    "                    score=0.0,\n",
    "                )\n",
    "        except Exception as e:\n",
    "            print(e)\n",
    "            # print(pred)\n",
    "            return MetricResult(\n",
    "                score=0.0,\n",
    "            )\n",
    "\n",
    "class GlobalAverageMetric(BaseGlobalMetric):\n",
    "    async def compute(\n",
    "        self,\n",
    "        results: List[MetricResult],\n",
    "    ) -> GlobalMetricResult:\n",
    "        try:\n",
    "            scores = [result.score for result in results]\n",
    "            return GlobalMetricResult(\n",
    "                score=sum(scores) / len(scores) if len(results) > 0 else 0.0,\n",
    "            )\n",
    "        except Exception as e:\n",
    "            # print(\"Error in GlobalEmotionMetric: \", e)\n",
    "            return GlobalMetricResult(\n",
    "                score=0.0,\n",
    "            )\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Select Trainer & Run"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jypark/Desktop/weavel/Ape/venv/lib/python3.11/site-packages/pydantic/main.py:390: UserWarning: Pydantic serializer warnings:\n",
      "  Expected `PromptTokensDetails` but got `dict` with value `{'audio_tokens': None, 'cached_tokens': 0}` - serialized value may not be as expected\n",
      "  return self.__pydantic_serializer__.to_python(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 1 column 12 (char 11)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 2 column 16 (char 17)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 2 column 16 (char 17)\n",
      "Unterminated string starting at: line 2 column 16 (char 17)\n",
      "Unterminated string starting at: line 2 column 16 (char 17)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 1 column 12 (char 11)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 2 column 16 (char 17)\n",
      "Unterminated string starting at: line 2 column 16 (char 17)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "Unterminated string starting at: line 1 column 12 (char 11)\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "'answer'\n",
      "Unterminated string starting at: line 2 column 14 (char 15)\n",
      "'answer'\n",
      "'answer'\n",
      "'answer'\n"
     ]
    }
   ],
   "source": [
    "from ape.core.trainer import (\n",
    "    TextGradientTrainer,\n",
    "    ExpelTrainer,\n",
    "    FewShotTrainer,\n",
    "    EvoPromptTrainer,\n",
    "    DspyMiproTrainer,\n",
    "    OptunaTrainer,\n",
    ")\n",
    "\n",
    "# define trainer \n",
    "trainer = FewShotTrainer(\n",
    "    generator=MathSolver(),\n",
    "    metric=MathMetricLLM(),\n",
    "    global_metric=GlobalAverageMetric(),\n",
    "    testmode=True # If True, trainer will run prompts for validation set and save results.\n",
    ")\n",
    "\n",
    "# run trainer\n",
    "optimized_prompt, report = await trainer.train(\n",
    "    prompt=student_prompt,\n",
    "    trainset=trainset,\n",
    "    valset=valset,\n",
    ")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Print Optimized Prompt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'role': 'system', 'content': 'For given math problem, solve it step by step.\\n\\nQuestion: {problem}\\n\\nYou MUST respond in JSON format with the following fields:\\nthought: the reasoning process of the problem solving.\\nanswer: only return the answer without any explanation.\\n\\nExample:\\n{_FEWSHOT_}'}\n"
     ]
    }
   ],
   "source": [
    "# print optimized prompt\n",
    "for message in optimized_prompt.messages:\n",
    "    print(message)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Print Benchmark Test Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualize experiment results\n",
    "def visualize_scores(report):\n",
    "    scores = report.scores\n",
    "    trainset_scores = [score[\"score\"] for score in scores]\n",
    "    valset_scores = [score[\"val_score\"] for score in scores]\n",
    "    iterations = range(1, len(trainset_scores) + 1)\n",
    "\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.plot(iterations, trainset_scores, label='Training Set', marker='o')\n",
    "    plt.plot(iterations, valset_scores, label='Validation Set', marker='s')\n",
    "    \n",
    "    for i, (train_score, val_score) in enumerate(zip(trainset_scores, valset_scores)):\n",
    "        plt.text(iterations[i], train_score, f'{train_score:.2f}', \n",
    "                    ha='center', va='bottom', fontsize=8, color='blue')\n",
    "        plt.text(iterations[i], val_score, f'{val_score:.2f}', \n",
    "                    ha='center', va='bottom', fontsize=8, color='green')\n",
    "\n",
    "    plt.title('Training and Validation Scores over Iterations')\n",
    "    plt.xlabel('Iteration')\n",
    "    plt.ylabel('Score')\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "visualize_scores(report)"
   ]
  }
 ],
 "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}