{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Geometry Problem Solving Agent\n",
    "\n",
    "> 结合 FormalGeo 与 Agent 的几何问题形式化自动求解器\n",
    "\n",
    "本 Notebook 演示如何使用 RAVS 框架求解几何问题。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 导入必要的模块\n",
    "import sys\n",
    "import os\n",
    "\n",
    "# 添加项目路径\n",
    "sys.path.append(\"../src\")\n",
    "os.chdir(\"../src/ravs\")\n",
    "\n",
    "from agent_loop import main"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. 单题求解示例\n",
    "\n",
    "求解编号为 1 的几何问题"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 求解单个题目\n",
    "result = main(\n",
    "    test_pids=[1],          # 题目编号，范围 1-7000\n",
    "    log_path=\"../../outputs/log/test.json\",  # 日志保存路径\n",
    "    model_names=['Deepseek'],  # 使用的 LLM\n",
    "    max_epoch=50,            # 最大交互轮次\n",
    "    max_context=80000,       # 最大上下文长度\n",
    "    solve_again=False,       # 失败后是否重试\n",
    "    debug_mode=True          # 是否输出详细日志\n",
    ")\n",
    "\n",
    "print(f\"求解状态: {result}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. 批量求解示例\n",
    "\n",
    "批量求解编号为 1-5 的几何问题"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 批量求解\n",
    "results = main(\n",
    "    test_pids=list(range(1, 6)),  # 批量求解 1-5 题\n",
    "    log_path=\"../../outputs/log/batch_test.json\",\n",
    "    model_names=['Deepseek'],\n",
    "    max_epoch=50,\n",
    "    max_context=80000,\n",
    "    solve_again=True,\n",
    "    debug_mode=False\n",
    ")\n",
    "\n",
    "print(\"批量求解完成！\")\n",
    "print(f\"结果: {results}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. 多模型并行求解\n",
    "\n",
    "使用多个 LLM 模型并行处理不同题目"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 多模型并行（每个模型名称对应一个进程）\n",
    "results = main(\n",
    "    test_pids=[1, 2, 3, 4],\n",
    "    log_path=\"../../outputs/log/multi_model_test.json\",\n",
    "    model_names=['Deepseek', 'Deepseek', 'BaiLian'],  # 3 个进程\n",
    "    max_epoch=50,\n",
    "    max_context=80000,\n",
    "    solve_again=True,\n",
    "    debug_mode=False\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. 绘制统计图表\n",
    "\n",
    "绘制论文中的统计图（需要先运行过求解）"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from chart import main as chart_main\n",
    "\n",
    "# 生成统计图表\n",
    "chart_main()\n",
    "print(\"图表已保存至 outputs/ 目录\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.12.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}