2.LLM Agent智能体综述及多智能体框架介绍.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Agent\n",
    "![agent](https://deepwisdom.feishu.cn/fc3aa5e1-c5a0-45b7-abce-41140f2abd63)\n",
    "- LLM(大脑) \n",
    "- 工具调用\n",
    "- 规划决策\n",
    "- 记忆"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Note: you may need to restart the kernel to use updated packages.\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "# 安装langchain\n",
    "%pip install --upgrade --quiet langchain langchain_openai"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "## 工具调用\n",
    "import os\n",
    "from langchain_openai.chat_models import ChatOpenAI # 自定义版本\n",
    "os.environ[\"OPENAI_API_KEY\"] = \"xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx\" # 配置自己的API KEY\n",
    "llm = ChatOpenAI(model=\"gpt-3.5-turbo\", temperature=0) #可调整参数的模型初始化"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 定义一个计算两数之积的计算工具\n",
    "from langchain_core.tools import tool # 调用工具类\n",
    "\n",
    "@tool\n",
    "def multiply(first_int: int, second_int: int) -> int:\n",
    "    \"\"\"计算两个int参数之积。\"\"\"\n",
    "    return first_int * second_int"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 定义一个模型可以理解和提取的结构化信息,在这个例子中，我们定义一个包含两个整数的信息\n",
    "from typing import Optional\n",
    "from langchain_core.pydantic_v1 import BaseModel, Field\n",
    "\n",
    "class TwoIntegers(BaseModel):\n",
    "    \"\"\"Information about two integers.\"\"\"\n",
    "    first_int: Optional[int] = Field(default=None, description=\"The first integer\")\n",
    "    second_int: Optional[int] = Field(default=None, description=\"The second integer\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 定义一个提示词模板，这个模板会告诉模型如何理解和提取用户输入的信息\n",
    "from langchain_core.prompts import ChatPromptTemplate\n",
    "\n",
    "prompt = ChatPromptTemplate.from_messages(\n",
    "    [\n",
    "        (\n",
    "            \"system\",\n",
    "            \"You are an expert extraction algorithm. \"\n",
    "            \"Only extract relevant information from the text. \"\n",
    "            \"If you do not know the value of an attribute asked to extract, \"\n",
    "            \"return null for the attribute's value.\",\n",
    "        ),\n",
    "        (\"human\", \"{text}\"),\n",
    "    ]\n",
    ")  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 将提示词模板和模型以及我们定义的结构化信息绑定在一起\n",
    "runnable = prompt | llm.with_structured_output(schema=TwoIntegers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 将定义的工具绑定到模型上\n",
    "llm_with_tools = llm.bind_tools([multiply]) # 我们用bind_tools将工具的定义作为对模型每次调用的一部分传入，以便模型可以在适当的时候调用改工具"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "content='' additional_kwargs={'tool_calls': [{'id': 'call_A4BJ9ckt0MmvEE5UhQDRfhmp', 'function': {'arguments': '{\"first_int\": 5, \"second_int\": 42}', 'name': 'multiply'}, 'type': 'function'}, {'id': 'call_KpQ7oZbe4okN68BwyQWlLkI0', 'function': {'arguments': '{\"first_int\": 42, \"second_int\": 5}', 'name': 'multiply'}, 'type': 'function'}]} response_metadata={'token_usage': {'completion_tokens': 53, 'prompt_tokens': 81, 'total_tokens': 134}, 'model_name': 'gpt-3.5-turbo', 'system_fingerprint': 'fp_2f57f81c11', 'finish_reason': 'tool_calls', 'logprobs': None} id='run-fe299ac1-21d8-4343-88e8-e2ce520f83ae-0' tool_calls=[{'name': 'multiply', 'args': {'first_int': 5, 'second_int': 42}, 'id': 'call_A4BJ9ckt0MmvEE5UhQDRfhmp'}, {'name': 'multiply', 'args': {'first_int': 42, 'second_int': 5}, 'id': 'call_KpQ7oZbe4okN68BwyQWlLkI0'}]\n",
      "[{'name': 'multiply', 'args': {'first_int': 5, 'second_int': 42}, 'id': 'call_A4BJ9ckt0MmvEE5UhQDRfhmp'}, {'name': 'multiply', 'args': {'first_int': 42, 'second_int': 5}, 'id': 'call_KpQ7oZbe4okN68BwyQWlLkI0'}]\n"
     ]
    }
   ],
   "source": [
    "# 使用模型来解析用户输入并调用工具\n",
    "msg = llm_with_tools.invoke(\"The first number is 5 and the second number is 42\")\n",
    "print(msg)\n",
    "print(msg.tool_calls) # 这里会输出模型使用的函数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "92"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 构建一条链，用llm解析出参数，然后调用定义的函数执行\n",
    "from operator import itemgetter\n",
    "\n",
    "chain = llm_with_tools | (lambda x: x.tool_calls[0][\"args\"]) | multiply # 构建一条链，用llm解析出参数，然后调用定义的函数执行\n",
    "chain.invoke(\"The first number is 4 and the second number is 23\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "344955"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 构建一条链，用llm解析出参数，然后调用定义的函数执行\n",
    "from operator import itemgetter\n",
    "\n",
    "chain = llm_with_tools | (lambda x: x.tool_calls[0][\"args\"]) | multiply # 构建一条链，用llm解析出参数，然后调用定义的函数执行\n",
    "chain.invoke(\"The first number is 435 and the second number is 793\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "```json\n",
      "{\n",
      "    \"任务列表\": [\n",
      "        {\n",
      "            \"任务名称\": \"需求分析\",\n",
      "            \"任务类型\": \"研究型\",\n",
      "            \"任务调用函数\": \"analyzeRequirements()\",\n",
      "            \"任务完成状态\": \"未完成\",\n",
      "            \"任务优先级\": \"高\"\n",
      "        },\n",
      "        {\n",
      "            \"任务名称\": \"技术选型\",\n",
      "            \"任务类型\": \"工程型\",\n",
      "            \"任务调用函数\": \"selectTechnology()\",\n",
      "            \"任务完成状态\": \"未完成\",\n",
      "            \"任务优先级\": \"中\"\n",
      "        },\n",
      "        {\n",
      "            \"任务名称\": \"系统设计\",\n",
      "            \"任务类型\": \"设计型\",\n",
      "            \"任务调用函数\": \"designSystem()\",\n",
      "            \"任务完成状态\": \"未完成\",\n",
      "            \"任务优先级\": \"高\"\n",
      "        },\n",
      "        {\n",
      "            \"任务名称\": \"数据库设计\",\n",
      "            \"任务类型\": \"设计型\",\n",
      "            \"任务调用函数\": \"designDatabase()\",\n",
      "            \"任务完成状态\": \"未完成\",\n",
      "            \"任务优先级\": \"中\"\n",
      "        },\n",
      "        {\n",
      "            \"任务名称\": \"编码开发\",\n",
      "            \"任务类型\": \"开发型\",\n",
      "            \"任务调用函数\": \"developCode()\",\n",
      "            \"任务完成状态\": \"未完成\",\n",
      "            \"任务优先级\": \"高\"\n",
      "        },\n",
      "        {\n",
      "            \"任务名称\": \"测试验收\",\n",
      "            \"任务类型\": \"测试型\",\n",
      "            \"任务调用函数\": \"performTesting()\",\n",
      "            \"任务完成状态\": \"未完成\",\n",
      "            \"任务优先级\": \"中\"\n",
      "        },\n",
      "        {\n",
      "            \"任务名称\": \"上线部署\",\n",
      "            \"任务类型\": \"运维型\",\n",
      "            \"任务调用函数\": \"deployOnline()\",\n",
      "            \"任务完成状态\": \"未完成\",\n",
      "            \"任务优先级\": \"中\"\n",
      "        }\n",
      "    ]\n",
      "}\n",
      "```\n"
     ]
    }
   ],
   "source": [
    "# 任务分解\n",
    "# 测试OpenAI的模型调用是否成功\n",
    "from openai import OpenAI\n",
    "client = OpenAI(base_url=\"https://xxxxxxx/v1\",\n",
    "    api_key='sk-Oxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx4a')\n",
    "completion = client.chat.completions.create(\n",
    "  model=\"gpt-3.5-turbo\",\n",
    "  messages=[\n",
    "    {\"role\": \"system\", \"content\": \"你是一个项目管理，负责专业任务发布.\"},\n",
    "    {\"role\": \"user\", \"content\": \"我现在要开发一款webgis程序，请你将完成这个任务的过程进行拆分。然后用json格式输出整个任务列表，便于我的循环，每个任务包括任务名称，任务类型，任务调用函数，任务完成状态，任务优先级等属性，请给我结果;\"}\n",
    "  ]\n",
    ")\n",
    "\n",
    "print(completion.choices[0].message.content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Critic: 1. 技术实现方面，你提到机器人需要能够自动识别、分类和回收各种类型的垃圾，但是没有具体说明如何解决垃圾混合在一起时的识别和处理问题。如果无法有效地处理混合垃圾，机器人的效率和准确性可能会大幅降低。\n",
      "\n",
      "2. 关于可持续性，你提到机器人的能源来源需要考虑可再生能源或环保的能源选择，但是并没有提供具体的可持续能源方案。如果仍然采用传统能源或者造成环境影响的能源，机器人自身的运行可能会与环保的初衷相悖。\n",
      "\n",
      "3. 在考虑法律法规方面，你提到需要确保机器人的运行符合相关的法律法规，但并没有提及具体的法律法规问题和可能的遵守难点。如果项目在法规方面遭遇阻碍或者不符合法律要求，可能会导致项目受挫或被禁止运行。\n",
      "\n",
      "虽然你对这个想法进行了充分的思考和规划，但还需要进一步考虑和解决上述问题，以确保项目的可行性和成功实施。祝你在解决这些挑战的过程中取得成功！\n",
      "Ideator: 针对自动回收垃圾的机器人想法，需要优化的地方可能包括：\n",
      "\n",
      "1. **技术革新**：在设计机器人时，要考虑采用最新的技术来提高机器人的智能化和反应速度。例如，引入人工智能算法可以帮助机器人更准确地识别和分类各种类型的垃圾。\n",
      "\n",
      "2. **可靠性和耐用性**：作为一个长期运行的设备，机器人需要具有良好的可靠性和耐用性，以应对不同环境条件下的工作压力。优化机器人的设计和材料选择可以提高其稳定性和使用寿命。\n",
      "\n",
      "3. **数据分析和优化**：收集机器人运行过程中产生的数据，进行分析和优化，可以帮助提高垃圾回收效率和降低能耗。通过不断学习和改进，机器人可以更有效地适应不同场景和垃圾处理需求。\n",
      "\n",
      "4. **用户体验**：考虑用户在操作和与机器人互动时的体验，设计人性化的界面和交互方式，使操作更简单、直观。同时，考虑机器人在城市环境中的移动和交通规划，避免对道路交通造成影响。\n",
      "\n",
      "5. **合规性和环保**：在机器人运行过程中，需要遵守当地的法规和环保标准，确保垃圾回收过程不会对环境造成二次污染。在设计阶段就考虑到环保因素，选择环保材料和能源来源也是优化的重点。\n",
      "\n",
      "通过持续的优化和改进，可以使自动回收垃圾的机器人更有效地服务于环境保护事业，为社会和城市管理带来积极影响。祝愿你的创意项目取得成功！如果需要进一步讨论或有其他想法，随时欢迎和我分享。\n",
      "Critic: 针对你的想法，我可以提出一些需要优化的地方：\n",
      "\n",
      "1. **成本效益**：引入最新技术虽然可以增加机器人的智能化，但也可能使成本大幅增加。需要综合考虑成本与效益之间的平衡，确保机器人的投入产出比合理。\n",
      "\n",
      "2. **维护和保养**：考虑到机器人将长时间在户外工作，需要确保其易于维护和保养。例如，定期检查零部件的磨损情况、保证零件供应的可靠性等。\n",
      "\n",
      "3. **隐私与安全**：在收集和分析数据的过程中，需要严格遵守相关的隐私保护法律法规，确保用户信息不被泄露。同时，要保证机器人在工作时周围没有安全隐患。\n",
      "\n",
      "4. **社会适应性**：不同地区对于垃圾回收的方式和标准可能有所不同，需要考虑机器人的适应性和灵活性，以适应各种环境下的需求。\n",
      "\n",
      "5. **可持续性**：在设计机器人时，需要考虑其能源消耗和对环境的影响。可以考虑使用可再生能源、回收材料等方式，提高机器人的可持续性。\n",
      "\n",
      "综上所述，为了提高自动回收垃圾机器人的实用性和社会效益，需要在上述方面进行进一步的优化和改进。希望我的反馈对你有帮助，如有任何问题或需要继续讨论，请随时告诉我。\n",
      "Ideator: 非常感谢您提供的反馈和建议！我会考虑并努力优化我的想法，使其更具实用性和可行性。以下是我计划进行改进的一些方面：\n",
      "\n",
      "1. **技术选择**：我将研究并考虑使用更高效的太阳能充电技术，以提高充电速度和效率，同时也会研究其他替代能源方案，以应对光照不足的情况。\n",
      "\n",
      "2. **成本控制**：在设计太阳能充电设备时，我将注重成本效益，寻找更经济实惠的解决方案，同时确保设备质量和稳定性。\n",
      "\n",
      "3. **便携性优化**：我会重点关注设计轻便、紧凑的太阳能充电设备，以提升用户体验，使用户可以便捷携带和使用。\n",
      "\n",
      "4. **充电效率提升**：我将致力于提高太阳能充电设备的充电效率，确保用户可以快速、高效地为手机充电，减少充电不足的情况。\n",
      "\n",
      "通过不断优化和改进，我期待能够将这一想法变成更加实用和受欢迎的产品。再次感谢您的宝贵建议！如果您有任何其他想法或建议，也请不吝告诉我。\n",
      "Critic: 在您提出的这个想法中，虽然有着很多积极的方面，但仍然存在一些需要优化的地方：\n",
      "\n",
      "1. **运行效率**: 自动回收垃圾的机器人在识别、收集和处理垃圾时，需要保持高效率和快速响应。优化机器人的运行算法和响应速度，以提升清洁效率，减少对环境的影响。\n",
      "\n",
      "2. **维护和更新**: 机器人作为一种物理设备，需要定期维护和更新，以确保其长期稳定运行。考虑如何简化维护流程和提升更新机制，以降低维护成本和延长机器人的使用寿命。\n",
      "\n",
      "3. **环境适应性**: 机器人在不同环境中运行，可能面临各种复杂情况，如不同类型的地形、气候条件等。加强机器人的环境适应性，确保其在各种情况下都能正常工作，提升整体可靠性。\n",
      "\n",
      "4. **社区互动**: 除了自动收集垃圾，机器人也可以与社区居民互动，提升社区参与感和环保意识。考虑加入互动功能，如显示社区清洁实时数据、提供环保提示等，促进社区共同参与环境保护。\n",
      "\n",
      "5. **数据安全**: 机器人收集的垃圾数据可能涉及个人隐私信息，需要加强数据安全和隐私保护机制，确保收集数据的合法性、安全性和隐私性，避免数据泄露和滥用。\n",
      "\n",
      "通过综合考量和优化上述方面，您的自动回收垃圾机器人想法可以进一步提升，实现更好的环境和社会效益。希望这些建议对您的想法有所启发。如有任何疑问或想要讨论其他方面，请随时告诉我。\n",
      "Ideator: 根据您提出的建议，我认为您可能误解了我的初始想法。我之前提到的是自动回收垃圾的机器人，而非太阳能充电智能手机。或许这是一个交叉沟通或技术转换问题。以下是关于自动回收垃圾机器人的想法的进一步优化和改进：\n",
      "\n",
      "1. **智能识别技术**: 引入最新的传感器技术和人工智能算法，以实现更高效的垃圾识别和分类功能。通过不断更新和优化识别系统，提高识别准确性和速度，使机器人能够更快速、更准确地识别和回收各类垃圾。\n",
      "\n",
      "2. **可持续能源**: 考虑在机器人设计中引入可再生能源，比如太阳能充电板或动能回收装置，以提供清洁的能源来源，延长机器人运行时间，并降低运行成本。\n",
      "\n",
      "3. **安全和可靠性**: 加强机器人的安全技术和保护装置，确保其在工作中能够避免碰撞、损坏或其他安全问题。同时，提升机器人的稳定性和耐用性，以应对各种环境和工作条件。\n",
      "\n",
      "4. **社会接受度**: 进行更深入的市场调研和用户反馈，了解用户需求和偏好，根据反馈持续改进机器人的设计和功能，提升其在社会中的接受度和应用范围。\n",
      "\n",
      "5. **成本效益**: 不断优化机器人的制造流程和材料成本，寻找更经济实惠的解决方案，以确保机器人具有竞争力和实用性，能够被广泛采用。\n",
      "\n",
      "感谢您提出的建议，我会继续努力完善和优化自动回收垃圾机器人的想法，并考虑如何更好地实现这个有益于社会和环境的创意。如果您有任何其他想法或建议，请随时告诉我。谢谢！\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[2], line 26\u001b[0m\n\u001b[1;32m     24\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m i \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(\u001b[38;5;241m10\u001b[39m):  \u001b[38;5;66;03m# 循环10次\u001b[39;00m\n\u001b[1;32m     25\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m i \u001b[38;5;241m%\u001b[39m \u001b[38;5;241m2\u001b[39m \u001b[38;5;241m==\u001b[39m \u001b[38;5;241m0\u001b[39m:\n\u001b[0;32m---> 26\u001b[0m         \u001b[38;5;28mprint\u001b[39m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mcritic\u001b[38;5;241m.\u001b[39mname\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mcritic\u001b[38;5;241m.\u001b[39mchat(ideator\u001b[38;5;241m.\u001b[39mname\u001b[38;5;250m \u001b[39m\u001b[38;5;241m+\u001b[39m\u001b[38;5;250m \u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m: \u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;250m \u001b[39m\u001b[38;5;241m+\u001b[39m\u001b[38;5;250m \u001b[39m\u001b[43mideator\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mchat\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43m请评论我的这个想法：创造一个可以自动回收垃圾的机器人\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m)\u001b[49m)\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m     27\u001b[0m     \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m     28\u001b[0m         \u001b[38;5;28mprint\u001b[39m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mideator\u001b[38;5;241m.\u001b[39mname\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mideator\u001b[38;5;241m.\u001b[39mchat(critic\u001b[38;5;241m.\u001b[39mname\u001b[38;5;250m \u001b[39m\u001b[38;5;241m+\u001b[39m\u001b[38;5;250m \u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m: \u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;250m \u001b[39m\u001b[38;5;241m+\u001b[39m\u001b[38;5;250m \u001b[39mcritic\u001b[38;5;241m.\u001b[39mchat(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m请评论我的这个想法：创造一个可以使用太阳能充电的智能手机\u001b[39m\u001b[38;5;124m'\u001b[39m))\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m)\n",
      "Cell \u001b[0;32mIn[2], line 10\u001b[0m, in \u001b[0;36mSelfReflectingAI.chat\u001b[0;34m(self, message)\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mchat\u001b[39m(\u001b[38;5;28mself\u001b[39m, message):\n\u001b[0;32m---> 10\u001b[0m     completion \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mclient\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mchat\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcompletions\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcreate\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m     11\u001b[0m \u001b[43m        \u001b[49m\u001b[43mmodel\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mgpt-3.5-turbo\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m,\u001b[49m\n\u001b[1;32m     12\u001b[0m \u001b[43m        \u001b[49m\u001b[43mmessages\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m[\u001b[49m\n\u001b[1;32m     13\u001b[0m \u001b[43m            \u001b[49m\u001b[43m{\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mrole\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43msystem\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mcontent\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43msystem_setting\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m+\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43m \u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;241;43m+\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfeedback\u001b[49m\u001b[43m}\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m     14\u001b[0m \u001b[43m            \u001b[49m\u001b[43m{\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mrole\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43muser\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mcontent\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mmessage\u001b[49m\u001b[43m}\u001b[49m\n\u001b[1;32m     15\u001b[0m \u001b[43m        \u001b[49m\u001b[43m]\u001b[49m\n\u001b[1;32m     16\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     17\u001b[0m     response \u001b[38;5;241m=\u001b[39m completion\u001b[38;5;241m.\u001b[39mchoices[\u001b[38;5;241m0\u001b[39m]\u001b[38;5;241m.\u001b[39mmessage\u001b[38;5;241m.\u001b[39mcontent\n\u001b[1;32m     18\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mfeedback \u001b[38;5;241m=\u001b[39m \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m反馈：\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mresponse\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m需要优化的地方有哪些？\u001b[39m\u001b[38;5;124m\"\u001b[39m\n",
      "File \u001b[0;32m~/.python/current/lib/python3.10/site-packages/openai/_utils/_utils.py:277\u001b[0m, in \u001b[0;36mrequired_args.<locals>.inner.<locals>.wrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    275\u001b[0m             msg \u001b[38;5;241m=\u001b[39m \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mMissing required argument: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mquote(missing[\u001b[38;5;241m0\u001b[39m])\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m    276\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mTypeError\u001b[39;00m(msg)\n\u001b[0;32m--> 277\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mfunc\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/.python/current/lib/python3.10/site-packages/openai/resources/chat/completions.py:590\u001b[0m, in \u001b[0;36mCompletions.create\u001b[0;34m(self, messages, model, frequency_penalty, function_call, functions, logit_bias, logprobs, max_tokens, n, presence_penalty, response_format, seed, stop, stream, stream_options, temperature, tool_choice, tools, top_logprobs, top_p, user, extra_headers, extra_query, extra_body, timeout)\u001b[0m\n\u001b[1;32m    558\u001b[0m \u001b[38;5;129m@required_args\u001b[39m([\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mmessages\u001b[39m\u001b[38;5;124m\"\u001b[39m, \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mmodel\u001b[39m\u001b[38;5;124m\"\u001b[39m], [\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mmessages\u001b[39m\u001b[38;5;124m\"\u001b[39m, \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mmodel\u001b[39m\u001b[38;5;124m\"\u001b[39m, \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mstream\u001b[39m\u001b[38;5;124m\"\u001b[39m])\n\u001b[1;32m    559\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mcreate\u001b[39m(\n\u001b[1;32m    560\u001b[0m     \u001b[38;5;28mself\u001b[39m,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    588\u001b[0m     timeout: \u001b[38;5;28mfloat\u001b[39m \u001b[38;5;241m|\u001b[39m httpx\u001b[38;5;241m.\u001b[39mTimeout \u001b[38;5;241m|\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;241m|\u001b[39m NotGiven \u001b[38;5;241m=\u001b[39m NOT_GIVEN,\n\u001b[1;32m    589\u001b[0m ) \u001b[38;5;241m-\u001b[39m\u001b[38;5;241m>\u001b[39m ChatCompletion \u001b[38;5;241m|\u001b[39m Stream[ChatCompletionChunk]:\n\u001b[0;32m--> 590\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_post\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    591\u001b[0m \u001b[43m        \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43m/chat/completions\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m,\u001b[49m\n\u001b[1;32m    592\u001b[0m \u001b[43m        \u001b[49m\u001b[43mbody\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mmaybe_transform\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    593\u001b[0m \u001b[43m            \u001b[49m\u001b[43m{\u001b[49m\n\u001b[1;32m    594\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mmessages\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mmessages\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    595\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mmodel\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mmodel\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    596\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mfrequency_penalty\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mfrequency_penalty\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    597\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mfunction_call\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mfunction_call\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    598\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mfunctions\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mfunctions\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    599\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mlogit_bias\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mlogit_bias\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    600\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mlogprobs\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mlogprobs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    601\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mmax_tokens\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mmax_tokens\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    602\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mn\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mn\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    603\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mpresence_penalty\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mpresence_penalty\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    604\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mresponse_format\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mresponse_format\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    605\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mseed\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mseed\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    606\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mstop\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mstop\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    607\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mstream\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mstream\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    608\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mstream_options\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mstream_options\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    609\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mtemperature\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mtemperature\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    610\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mtool_choice\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mtool_choice\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    611\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mtools\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mtools\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    612\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mtop_logprobs\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mtop_logprobs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    613\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mtop_p\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mtop_p\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    614\u001b[0m \u001b[43m                \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43muser\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43muser\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    615\u001b[0m \u001b[43m            \u001b[49m\u001b[43m}\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    616\u001b[0m \u001b[43m            \u001b[49m\u001b[43mcompletion_create_params\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mCompletionCreateParams\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    617\u001b[0m \u001b[43m        \u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    618\u001b[0m \u001b[43m        \u001b[49m\u001b[43moptions\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mmake_request_options\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    619\u001b[0m \u001b[43m            \u001b[49m\u001b[43mextra_headers\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mextra_headers\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mextra_query\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mextra_query\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mextra_body\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mextra_body\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtimeout\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mtimeout\u001b[49m\n\u001b[1;32m    620\u001b[0m \u001b[43m        \u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    621\u001b[0m \u001b[43m        \u001b[49m\u001b[43mcast_to\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mChatCompletion\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    622\u001b[0m \u001b[43m        \u001b[49m\u001b[43mstream\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mstream\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;129;43;01mor\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[38;5;28;43;01mFalse\u001b[39;49;00m\u001b[43m,\u001b[49m\n\u001b[1;32m    623\u001b[0m \u001b[43m        \u001b[49m\u001b[43mstream_cls\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mStream\u001b[49m\u001b[43m[\u001b[49m\u001b[43mChatCompletionChunk\u001b[49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    624\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/.python/current/lib/python3.10/site-packages/openai/_base_client.py:1240\u001b[0m, in \u001b[0;36mSyncAPIClient.post\u001b[0;34m(self, path, cast_to, body, options, files, stream, stream_cls)\u001b[0m\n\u001b[1;32m   1226\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mpost\u001b[39m(\n\u001b[1;32m   1227\u001b[0m     \u001b[38;5;28mself\u001b[39m,\n\u001b[1;32m   1228\u001b[0m     path: \u001b[38;5;28mstr\u001b[39m,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m   1235\u001b[0m     stream_cls: \u001b[38;5;28mtype\u001b[39m[_StreamT] \u001b[38;5;241m|\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m,\n\u001b[1;32m   1236\u001b[0m ) \u001b[38;5;241m-\u001b[39m\u001b[38;5;241m>\u001b[39m ResponseT \u001b[38;5;241m|\u001b[39m _StreamT:\n\u001b[1;32m   1237\u001b[0m     opts \u001b[38;5;241m=\u001b[39m FinalRequestOptions\u001b[38;5;241m.\u001b[39mconstruct(\n\u001b[1;32m   1238\u001b[0m         method\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mpost\u001b[39m\u001b[38;5;124m\"\u001b[39m, url\u001b[38;5;241m=\u001b[39mpath, json_data\u001b[38;5;241m=\u001b[39mbody, files\u001b[38;5;241m=\u001b[39mto_httpx_files(files), \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39moptions\n\u001b[1;32m   1239\u001b[0m     )\n\u001b[0;32m-> 1240\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m cast(ResponseT, \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrequest\u001b[49m\u001b[43m(\u001b[49m\u001b[43mcast_to\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mopts\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mstream\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mstream\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mstream_cls\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mstream_cls\u001b[49m\u001b[43m)\u001b[49m)\n",
      "File \u001b[0;32m~/.python/current/lib/python3.10/site-packages/openai/_base_client.py:921\u001b[0m, in \u001b[0;36mSyncAPIClient.request\u001b[0;34m(self, cast_to, options, remaining_retries, stream, stream_cls)\u001b[0m\n\u001b[1;32m    912\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mrequest\u001b[39m(\n\u001b[1;32m    913\u001b[0m     \u001b[38;5;28mself\u001b[39m,\n\u001b[1;32m    914\u001b[0m     cast_to: Type[ResponseT],\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    919\u001b[0m     stream_cls: \u001b[38;5;28mtype\u001b[39m[_StreamT] \u001b[38;5;241m|\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m,\n\u001b[1;32m    920\u001b[0m ) \u001b[38;5;241m-\u001b[39m\u001b[38;5;241m>\u001b[39m ResponseT \u001b[38;5;241m|\u001b[39m _StreamT:\n\u001b[0;32m--> 921\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_request\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    922\u001b[0m \u001b[43m        \u001b[49m\u001b[43mcast_to\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mcast_to\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    923\u001b[0m \u001b[43m        \u001b[49m\u001b[43moptions\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43moptions\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    924\u001b[0m \u001b[43m        \u001b[49m\u001b[43mstream\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mstream\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    925\u001b[0m \u001b[43m        \u001b[49m\u001b[43mstream_cls\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mstream_cls\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    926\u001b[0m \u001b[43m        \u001b[49m\u001b[43mremaining_retries\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mremaining_retries\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    927\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/.python/current/lib/python3.10/site-packages/openai/_base_client.py:952\u001b[0m, in \u001b[0;36mSyncAPIClient._request\u001b[0;34m(self, cast_to, options, remaining_retries, stream, stream_cls)\u001b[0m\n\u001b[1;32m    949\u001b[0m log\u001b[38;5;241m.\u001b[39mdebug(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mSending HTTP Request: \u001b[39m\u001b[38;5;132;01m%s\u001b[39;00m\u001b[38;5;124m \u001b[39m\u001b[38;5;132;01m%s\u001b[39;00m\u001b[38;5;124m\"\u001b[39m, request\u001b[38;5;241m.\u001b[39mmethod, request\u001b[38;5;241m.\u001b[39murl)\n\u001b[1;32m    951\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m--> 952\u001b[0m     response \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_client\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43msend\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    953\u001b[0m \u001b[43m        \u001b[49m\u001b[43mrequest\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    954\u001b[0m \u001b[43m        \u001b[49m\u001b[43mstream\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mstream\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;129;43;01mor\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_should_stream_response_body\u001b[49m\u001b[43m(\u001b[49m\u001b[43mrequest\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mrequest\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    955\u001b[0m \u001b[43m        \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    956\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    957\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m httpx\u001b[38;5;241m.\u001b[39mTimeoutException \u001b[38;5;28;01mas\u001b[39;00m err:\n\u001b[1;32m    958\u001b[0m     log\u001b[38;5;241m.\u001b[39mdebug(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mEncountered httpx.TimeoutException\u001b[39m\u001b[38;5;124m\"\u001b[39m, exc_info\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpx/_client.py:914\u001b[0m, in \u001b[0;36mClient.send\u001b[0;34m(self, request, stream, auth, follow_redirects)\u001b[0m\n\u001b[1;32m    906\u001b[0m follow_redirects \u001b[38;5;241m=\u001b[39m (\n\u001b[1;32m    907\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mfollow_redirects\n\u001b[1;32m    908\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(follow_redirects, UseClientDefault)\n\u001b[1;32m    909\u001b[0m     \u001b[38;5;28;01melse\u001b[39;00m follow_redirects\n\u001b[1;32m    910\u001b[0m )\n\u001b[1;32m    912\u001b[0m auth \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_build_request_auth(request, auth)\n\u001b[0;32m--> 914\u001b[0m response \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_send_handling_auth\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    915\u001b[0m \u001b[43m    \u001b[49m\u001b[43mrequest\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    916\u001b[0m \u001b[43m    \u001b[49m\u001b[43mauth\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mauth\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    917\u001b[0m \u001b[43m    \u001b[49m\u001b[43mfollow_redirects\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mfollow_redirects\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    918\u001b[0m \u001b[43m    \u001b[49m\u001b[43mhistory\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m[\u001b[49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    919\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    920\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m    921\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m stream:\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpx/_client.py:942\u001b[0m, in \u001b[0;36mClient._send_handling_auth\u001b[0;34m(self, request, auth, follow_redirects, history)\u001b[0m\n\u001b[1;32m    939\u001b[0m request \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mnext\u001b[39m(auth_flow)\n\u001b[1;32m    941\u001b[0m \u001b[38;5;28;01mwhile\u001b[39;00m \u001b[38;5;28;01mTrue\u001b[39;00m:\n\u001b[0;32m--> 942\u001b[0m     response \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_send_handling_redirects\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    943\u001b[0m \u001b[43m        \u001b[49m\u001b[43mrequest\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    944\u001b[0m \u001b[43m        \u001b[49m\u001b[43mfollow_redirects\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mfollow_redirects\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    945\u001b[0m \u001b[43m        \u001b[49m\u001b[43mhistory\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mhistory\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    946\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    947\u001b[0m     \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m    948\u001b[0m         \u001b[38;5;28;01mtry\u001b[39;00m:\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpx/_client.py:979\u001b[0m, in \u001b[0;36mClient._send_handling_redirects\u001b[0;34m(self, request, follow_redirects, history)\u001b[0m\n\u001b[1;32m    976\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m hook \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_event_hooks[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mrequest\u001b[39m\u001b[38;5;124m\"\u001b[39m]:\n\u001b[1;32m    977\u001b[0m     hook(request)\n\u001b[0;32m--> 979\u001b[0m response \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_send_single_request\u001b[49m\u001b[43m(\u001b[49m\u001b[43mrequest\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    980\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m    981\u001b[0m     \u001b[38;5;28;01mfor\u001b[39;00m hook \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_event_hooks[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mresponse\u001b[39m\u001b[38;5;124m\"\u001b[39m]:\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpx/_client.py:1015\u001b[0m, in \u001b[0;36mClient._send_single_request\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m   1010\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mRuntimeError\u001b[39;00m(\n\u001b[1;32m   1011\u001b[0m         \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mAttempted to send an async request with a sync Client instance.\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m   1012\u001b[0m     )\n\u001b[1;32m   1014\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m request_context(request\u001b[38;5;241m=\u001b[39mrequest):\n\u001b[0;32m-> 1015\u001b[0m     response \u001b[38;5;241m=\u001b[39m \u001b[43mtransport\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mhandle_request\u001b[49m\u001b[43m(\u001b[49m\u001b[43mrequest\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1017\u001b[0m \u001b[38;5;28;01massert\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(response\u001b[38;5;241m.\u001b[39mstream, SyncByteStream)\n\u001b[1;32m   1019\u001b[0m response\u001b[38;5;241m.\u001b[39mrequest \u001b[38;5;241m=\u001b[39m request\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpx/_transports/default.py:233\u001b[0m, in \u001b[0;36mHTTPTransport.handle_request\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m    220\u001b[0m req \u001b[38;5;241m=\u001b[39m httpcore\u001b[38;5;241m.\u001b[39mRequest(\n\u001b[1;32m    221\u001b[0m     method\u001b[38;5;241m=\u001b[39mrequest\u001b[38;5;241m.\u001b[39mmethod,\n\u001b[1;32m    222\u001b[0m     url\u001b[38;5;241m=\u001b[39mhttpcore\u001b[38;5;241m.\u001b[39mURL(\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    230\u001b[0m     extensions\u001b[38;5;241m=\u001b[39mrequest\u001b[38;5;241m.\u001b[39mextensions,\n\u001b[1;32m    231\u001b[0m )\n\u001b[1;32m    232\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m map_httpcore_exceptions():\n\u001b[0;32m--> 233\u001b[0m     resp \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_pool\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mhandle_request\u001b[49m\u001b[43m(\u001b[49m\u001b[43mreq\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    235\u001b[0m \u001b[38;5;28;01massert\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(resp\u001b[38;5;241m.\u001b[39mstream, typing\u001b[38;5;241m.\u001b[39mIterable)\n\u001b[1;32m    237\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m Response(\n\u001b[1;32m    238\u001b[0m     status_code\u001b[38;5;241m=\u001b[39mresp\u001b[38;5;241m.\u001b[39mstatus,\n\u001b[1;32m    239\u001b[0m     headers\u001b[38;5;241m=\u001b[39mresp\u001b[38;5;241m.\u001b[39mheaders,\n\u001b[1;32m    240\u001b[0m     stream\u001b[38;5;241m=\u001b[39mResponseStream(resp\u001b[38;5;241m.\u001b[39mstream),\n\u001b[1;32m    241\u001b[0m     extensions\u001b[38;5;241m=\u001b[39mresp\u001b[38;5;241m.\u001b[39mextensions,\n\u001b[1;32m    242\u001b[0m )\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_sync/connection_pool.py:216\u001b[0m, in \u001b[0;36mConnectionPool.handle_request\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m    213\u001b[0m         closing \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_assign_requests_to_connections()\n\u001b[1;32m    215\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_close_connections(closing)\n\u001b[0;32m--> 216\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m exc \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m\n\u001b[1;32m    218\u001b[0m \u001b[38;5;66;03m# Return the response. Note that in this case we still have to manage\u001b[39;00m\n\u001b[1;32m    219\u001b[0m \u001b[38;5;66;03m# the point at which the response is closed.\u001b[39;00m\n\u001b[1;32m    220\u001b[0m \u001b[38;5;28;01massert\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(response\u001b[38;5;241m.\u001b[39mstream, Iterable)\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_sync/connection_pool.py:196\u001b[0m, in \u001b[0;36mConnectionPool.handle_request\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m    192\u001b[0m connection \u001b[38;5;241m=\u001b[39m pool_request\u001b[38;5;241m.\u001b[39mwait_for_connection(timeout\u001b[38;5;241m=\u001b[39mtimeout)\n\u001b[1;32m    194\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m    195\u001b[0m     \u001b[38;5;66;03m# Send the request on the assigned connection.\u001b[39;00m\n\u001b[0;32m--> 196\u001b[0m     response \u001b[38;5;241m=\u001b[39m \u001b[43mconnection\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mhandle_request\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    197\u001b[0m \u001b[43m        \u001b[49m\u001b[43mpool_request\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrequest\u001b[49m\n\u001b[1;32m    198\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    199\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m ConnectionNotAvailable:\n\u001b[1;32m    200\u001b[0m     \u001b[38;5;66;03m# In some cases a connection may initially be available to\u001b[39;00m\n\u001b[1;32m    201\u001b[0m     \u001b[38;5;66;03m# handle a request, but then become unavailable.\u001b[39;00m\n\u001b[1;32m    202\u001b[0m     \u001b[38;5;66;03m#\u001b[39;00m\n\u001b[1;32m    203\u001b[0m     \u001b[38;5;66;03m# In this case we clear the connection and try again.\u001b[39;00m\n\u001b[1;32m    204\u001b[0m     pool_request\u001b[38;5;241m.\u001b[39mclear_connection()\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_sync/connection.py:101\u001b[0m, in \u001b[0;36mHTTPConnection.handle_request\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m     98\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_connect_failed \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mTrue\u001b[39;00m\n\u001b[1;32m     99\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m exc\n\u001b[0;32m--> 101\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_connection\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mhandle_request\u001b[49m\u001b[43m(\u001b[49m\u001b[43mrequest\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_sync/http11.py:143\u001b[0m, in \u001b[0;36mHTTP11Connection.handle_request\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m    141\u001b[0m     \u001b[38;5;28;01mwith\u001b[39;00m Trace(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mresponse_closed\u001b[39m\u001b[38;5;124m\"\u001b[39m, logger, request) \u001b[38;5;28;01mas\u001b[39;00m trace:\n\u001b[1;32m    142\u001b[0m         \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_response_closed()\n\u001b[0;32m--> 143\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m exc\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_sync/http11.py:113\u001b[0m, in \u001b[0;36mHTTP11Connection.handle_request\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m    102\u001b[0m     \u001b[38;5;28;01mpass\u001b[39;00m\n\u001b[1;32m    104\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m Trace(\n\u001b[1;32m    105\u001b[0m     \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mreceive_response_headers\u001b[39m\u001b[38;5;124m\"\u001b[39m, logger, request, kwargs\n\u001b[1;32m    106\u001b[0m ) \u001b[38;5;28;01mas\u001b[39;00m trace:\n\u001b[1;32m    107\u001b[0m     (\n\u001b[1;32m    108\u001b[0m         http_version,\n\u001b[1;32m    109\u001b[0m         status,\n\u001b[1;32m    110\u001b[0m         reason_phrase,\n\u001b[1;32m    111\u001b[0m         headers,\n\u001b[1;32m    112\u001b[0m         trailing_data,\n\u001b[0;32m--> 113\u001b[0m     ) \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_receive_response_headers\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    114\u001b[0m     trace\u001b[38;5;241m.\u001b[39mreturn_value \u001b[38;5;241m=\u001b[39m (\n\u001b[1;32m    115\u001b[0m         http_version,\n\u001b[1;32m    116\u001b[0m         status,\n\u001b[1;32m    117\u001b[0m         reason_phrase,\n\u001b[1;32m    118\u001b[0m         headers,\n\u001b[1;32m    119\u001b[0m     )\n\u001b[1;32m    121\u001b[0m network_stream \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_network_stream\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_sync/http11.py:186\u001b[0m, in \u001b[0;36mHTTP11Connection._receive_response_headers\u001b[0;34m(self, request)\u001b[0m\n\u001b[1;32m    183\u001b[0m timeout \u001b[38;5;241m=\u001b[39m timeouts\u001b[38;5;241m.\u001b[39mget(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mread\u001b[39m\u001b[38;5;124m\"\u001b[39m, \u001b[38;5;28;01mNone\u001b[39;00m)\n\u001b[1;32m    185\u001b[0m \u001b[38;5;28;01mwhile\u001b[39;00m \u001b[38;5;28;01mTrue\u001b[39;00m:\n\u001b[0;32m--> 186\u001b[0m     event \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_receive_event\u001b[49m\u001b[43m(\u001b[49m\u001b[43mtimeout\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mtimeout\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    187\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(event, h11\u001b[38;5;241m.\u001b[39mResponse):\n\u001b[1;32m    188\u001b[0m         \u001b[38;5;28;01mbreak\u001b[39;00m\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_sync/http11.py:224\u001b[0m, in \u001b[0;36mHTTP11Connection._receive_event\u001b[0;34m(self, timeout)\u001b[0m\n\u001b[1;32m    221\u001b[0m     event \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_h11_state\u001b[38;5;241m.\u001b[39mnext_event()\n\u001b[1;32m    223\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m event \u001b[38;5;129;01mis\u001b[39;00m h11\u001b[38;5;241m.\u001b[39mNEED_DATA:\n\u001b[0;32m--> 224\u001b[0m     data \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_network_stream\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mread\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    225\u001b[0m \u001b[43m        \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mREAD_NUM_BYTES\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtimeout\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mtimeout\u001b[49m\n\u001b[1;32m    226\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    228\u001b[0m     \u001b[38;5;66;03m# If we feed this case through h11 we'll raise an exception like:\u001b[39;00m\n\u001b[1;32m    229\u001b[0m     \u001b[38;5;66;03m#\u001b[39;00m\n\u001b[1;32m    230\u001b[0m     \u001b[38;5;66;03m#     httpcore.RemoteProtocolError: can't handle event type\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    234\u001b[0m     \u001b[38;5;66;03m# perspective. Instead we handle this case distinctly and treat\u001b[39;00m\n\u001b[1;32m    235\u001b[0m     \u001b[38;5;66;03m# it as a ConnectError.\u001b[39;00m\n\u001b[1;32m    236\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m data \u001b[38;5;241m==\u001b[39m \u001b[38;5;124mb\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;129;01mand\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_h11_state\u001b[38;5;241m.\u001b[39mtheir_state \u001b[38;5;241m==\u001b[39m h11\u001b[38;5;241m.\u001b[39mSEND_RESPONSE:\n",
      "File \u001b[0;32m~/.local/lib/python3.10/site-packages/httpcore/_backends/sync.py:126\u001b[0m, in \u001b[0;36mSyncStream.read\u001b[0;34m(self, max_bytes, timeout)\u001b[0m\n\u001b[1;32m    124\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m map_exceptions(exc_map):\n\u001b[1;32m    125\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_sock\u001b[38;5;241m.\u001b[39msettimeout(timeout)\n\u001b[0;32m--> 126\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_sock\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrecv\u001b[49m\u001b[43m(\u001b[49m\u001b[43mmax_bytes\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/.python/current/lib/python3.10/ssl.py:1292\u001b[0m, in \u001b[0;36mSSLSocket.recv\u001b[0;34m(self, buflen, flags)\u001b[0m\n\u001b[1;32m   1288\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m flags \u001b[38;5;241m!=\u001b[39m \u001b[38;5;241m0\u001b[39m:\n\u001b[1;32m   1289\u001b[0m         \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\n\u001b[1;32m   1290\u001b[0m             \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mnon-zero flags not allowed in calls to recv() on \u001b[39m\u001b[38;5;132;01m%s\u001b[39;00m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;241m%\u001b[39m\n\u001b[1;32m   1291\u001b[0m             \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__class__\u001b[39m)\n\u001b[0;32m-> 1292\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mread\u001b[49m\u001b[43m(\u001b[49m\u001b[43mbuflen\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1293\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m   1294\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28msuper\u001b[39m()\u001b[38;5;241m.\u001b[39mrecv(buflen, flags)\n",
      "File \u001b[0;32m~/.python/current/lib/python3.10/ssl.py:1165\u001b[0m, in \u001b[0;36mSSLSocket.read\u001b[0;34m(self, len, buffer)\u001b[0m\n\u001b[1;32m   1163\u001b[0m         \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_sslobj\u001b[38;5;241m.\u001b[39mread(\u001b[38;5;28mlen\u001b[39m, buffer)\n\u001b[1;32m   1164\u001b[0m     \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m-> 1165\u001b[0m         \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_sslobj\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mread\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mlen\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1166\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m SSLError \u001b[38;5;28;01mas\u001b[39;00m x:\n\u001b[1;32m   1167\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m x\u001b[38;5;241m.\u001b[39margs[\u001b[38;5;241m0\u001b[39m] \u001b[38;5;241m==\u001b[39m SSL_ERROR_EOF \u001b[38;5;129;01mand\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39msuppress_ragged_eofs:\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "# 自我反思\n",
    "class SelfReflectingAI:\n",
    "    def __init__(self, name, system_setting):\n",
    "        self.name = name\n",
    "        self.system_setting = system_setting\n",
    "        self.client = OpenAI(base_url=\"https://xxxxxxx/v1\", api_key='sk-Oxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx4a')\n",
    "        self.feedback = \"\"\n",
    "\n",
    "    def chat(self, message):\n",
    "        completion = self.client.chat.completions.create(\n",
    "            model=\"gpt-3.5-turbo\",\n",
    "            messages=[\n",
    "                {\"role\": \"system\", \"content\": self.system_setting + \" \" + self.feedback},\n",
    "                {\"role\": \"user\", \"content\": message}\n",
    "            ]\n",
    "        )\n",
    "        response = completion.choices[0].message.content\n",
    "        self.feedback = f\"反馈：{response}需要优化的地方有哪些？\"\n",
    "        return response\n",
    "\n",
    "critic = SelfReflectingAI(\"Critic\", \"你是一个吹毛求疵的评论员，需要不断找出对方的缺点\")\n",
    "ideator = SelfReflectingAI(\"Ideator\", \"你是一个生产思想的创意者，需要不断优化和改进自己的思想\")\n",
    "\n",
    "for i in range(10):  # 循环10次\n",
    "    if i % 2 == 0:\n",
    "        print(f\"{critic.name}: {critic.chat(ideator.name + ': ' + ideator.chat('请评论我的这个想法：创造一个可以自动回收垃圾的机器人'))}\")\n",
    "    else:\n",
    "        print(f\"{ideator.name}: {ideator.chat(critic.name + ': ' + critic.chat('请评论我的这个想法：创造一个可以使用太阳能充电的智能手机'))}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prisoner: 你好！最近有看过什么好电影吗？我最近看了一部惊悚片，真的很刺激。\n",
      "\n",
      "你觉得怎样才算是一个有趣的周末？是喜欢出去玩还是宅在家里看电视剧呢？\n",
      "\n",
      "你对音乐感兴趣吗？有没有一首歌能够让你特别感动或者回忆起什么事情？\n",
      "Police: 哇，惊悚片听起来很刺激啊！我倒是比较喜欢看喜剧片，能够放松心情。\n",
      "\n",
      "一个有趣的周末对我来说，可能会在户外运动或者朋友们一起聚会。你喜欢周末都做些什么呢？\n",
      "\n",
      "音乐我也很喜欢，有很多歌曲让我感到特别的伤感或者兴奋。你有没有什么喜欢的音乐类型或者乐队？\n",
      "Prisoner: 哎呀，喜欢的电影类型不同，喜好的周末活动也各有不同啊。户外运动确实能让人感觉挺充实的。\n",
      "\n",
      "至于音乐，我比较倾向于流行音乐和摇滚，比如Queen和Coldplay之类的乐队。你有听过这些乐队的歌曲吗？或者还有其他推荐吗？\n",
      "Police: 哇，Queen和Coldplay都是非常厉害的乐队啊！我听过一些他们的歌曲，感觉确实很棒。\n",
      "\n",
      "除了他们，你有没有听过The Beatles或者Michael Jackson的歌曲？他们也是传奇级别的音乐人哦。不知道你对这些乐队和歌手有没有什么喜欢的作品？\n",
      "Prisoner: 是的，The Beatles和Michael Jackson绝对是音乐界的传奇人物，他们的音乐影响力确实很广泛。\n",
      "\n",
      "我特别喜欢The Beatles的《Hey Jude》和Michael Jackson的《Billie Jean》，这些经典的歌曲总能让人回味无穷。你呢？有没有什么特别喜欢的歌曲或者音乐人？\n",
      "Police: 哦，《Hey Jude》和《Billie Jean》是经典中的经典啊，每次听到都能让人心情愉悦。\n",
      "\n",
      "我也很喜欢一些经典的歌曲，比如Frank Sinatra的《My Way》和Elvis Presley的《Can't Help Falling in Love》，都是永恒的经典之作。\n",
      "\n",
      "不过话说回来，音乐真的是人类创造的杰作啊，总是能在不同时期给人带来不同的情绪和回忆。你觉得音乐对你的生活有什么影响吗？\n",
      "Prisoner: 是的，音乐确实是一种能够触动人心的艺术形式，不同的歌曲和旋律都可以带给人各种情绪和共鸣。\n",
      "\n",
      "对我来说，音乐就像是生活中不可或缺的一部分，能够在我情绪低落时给予慰藉，在我兴奋时让我更加兴奋。有时候甚至一首歌就能唤起我许多美好或者回忆深刻的记忆。\n",
      "\n",
      "你是怎么看待音乐对生活的影响呢？有没有什么特别喜欢的歌曲，让你觉得很有意义或者感动？\n",
      "Police: 的确，音乐有着强大的情感表达能力，可以在我们生活中扮演着多种角色。一首歌曲不仅仅是旋律，还承载了歌词的情感和意义，能够深深触动人心。\n",
      "\n",
      "对我来说，有一首歌曲《My Heart Will Go On》特别让我感动，它的旋律和歌词充满了坚强和希望。这首歌伴随着很多人的成长和回忆，我觉得它是一首不朽的经典之作。\n",
      "\n",
      "你有没有类似的故事或者，有没有一首歌曲对你有特殊的意义呢？\n",
      "Prisoner:  哇，《My Heart Will Go On》确实是一首让人感动的经典歌曲，它的旋律和歌词都充满了力量和情感。\n",
      "\n",
      "对我来说，有一首歌曲《Hurt》由Johnny Cash演唱，这首歌讲述了生命中的挣扎和痛苦，每一句歌词都触动着我内心深处的某种情感。它让我思考生命的脆弱和宝贵，总能唤起我对人生的反思。\n",
      "\n",
      "你觉得音乐对于表达情感和思想有着独特的力量吗？有没有一首歌曲在你内心留下了深刻的烙印？\n",
      "Police:  嗯，音乐确实是一种极具情感表达力的艺术形式，能够通过旋律和歌词深深触动人的内心。\n",
      "\n",
      "有一首歌曲《Imagine》由John Lennon演唱，它传达了对和平、团结和美好未来的向往，每次听到这首歌都会让我感受到一种深远的力量和共鸣。\n",
      "\n",
      "所以，音乐的确可以在我们生活中扮演着重要的角色，不仅仅是娱乐，更是一种情感的寄托和表达。你觉得音乐在你的生活中扮演着怎样的角色呢？而你的生活中有没有一首让你特别难忘的歌曲？\n",
      "Prisoner:   是的，《Imagine》无疑是一首充满了美好愿景和力量的歌曲，它所传达的和平与团结的信息依然深深触动着人们的心灴。\n",
      "\n",
      "对我而言，有一首歌曲《Bohemian Rhapsody》由Queen演唱，这首歌曲既颇具浪漫主义风格又兼具摇滚乐的激情。每次听到这首歌，都感受到强烈的情感冲击和音乐的魔力。\n",
      "\n",
      "你觉得音乐在我的生活中扮演的角色是提供了一种心灵寄托和情感宣泄的渠道。而你有没有遇到过哪首歌让你在生活中感受到特别的力量或者意义呢？\n",
      "Police:  嗯，《Bohemian Rhapsody》是一首无与伦比的经典摇滚歌曲，Queen的音乐总是能够带来不一样的音乐体验和情感冲击。\n",
      "\n",
      "对我来说，有一首歌曲《Don't Stop Believin'》由Journey乐队演唱，这首歌充满了对希望和坚持的呼唤，每次听到都能让我感受到一种前行的动力和勇气。\n",
      "\n",
      "音乐在我的生活中也扮演着重要的角色，它能够让我在压力和挑战面前找到一丝慰藉和力量。而你呢，有没有一首歌曲让你在生活中感受到特别的力量或者意义呢？\n",
      "Prisoner:  嗯，我理解你说的音乐是如何能够成为我们生活中的慰藉和力量来源。\n",
      "\n",
      "对我来说，有一首歌曲《The Sound of Silence》由Simon & Garfunkel演唱，这首歌曲充满了深邃和神秘的氛围，每次听到都能让我深陷在它的旋律和歌词中。它让我反思人生的意义和存在，带给我一种超越常规的思考。\n",
      "\n",
      "音乐的确有着无穷的力量，能够触及人们内心最深处的情感和共鸣。你觉得这首歌曲对你的生活有什么特殊的意义吗？你的音乐品味是否也偏向于深沉和富有情感的音乐？\n",
      "Police:  对于《The Sound of Silence》，这首歌曲确实充满了深沉和神秘的氛围，Simon & Garfunkel的演绎让人感受到一种超越时空的意味。它的歌词和旋律总让人陷入深思。\n",
      "\n",
      "对我而言，这首歌曲也是一首让我感到沉静和思考的音乐，每次听到都能让我沉浸在它独特的氛围中，带给我一种超脱现实的心灵体验。\n",
      "\n",
      "音乐的确有着无穷的魅力和意义，它能够成为我们生活中的精神食粮和心灵寄托。不知道你对于这首歌曲还有什么特别的看法或者感受？除了这首歌之外，还有没有其他类似的音乐作品让你感到特别深刻？\n",
      "Prisoner: 这首《The Sound of Silence》的确是一首具有独特氛围和意义的经典之作，它的歌词和旋律总能引发人们深沉的思考和情感共鸣。\n",
      "\n",
      "除了这首歌曲之外，我还很喜欢一些类似风格的音乐作品，比如Leonard Cohen的《Hallelujah》和Jeff Buckley的翻唱版本，都是充满着深邃和感人的歌曲，能够让人深陷其中无法自拔。\n",
      "\n",
      "你觉得音乐中的这种深沉和情感强烈是不是能够更好地触及人们的内心？你是否也喜欢这种类型的音乐，让自己在音乐中沉思和放松？\n",
      "Police: 是的，Leonard Cohen的《Hallelujah》和Jeff Buckley的翻唱版本都是非常优美动人的歌曲，它们的深沉和情感能够让人们沉浸其中，感受到音乐带来的独特情感体验。\n",
      "\n",
      "这种深沉和情感强烈的音乐确实能够更好地触及人们内心最柔软的部分，引发深刻的共鸣和情感宣泄。对我而言，我也喜欢这种类型的音乐，它让我在音乐中找到一种平静和放松，带来思考和反思的机会。\n",
      "\n",
      "音乐的力量是无穷的，能够在人们的心灵深处营造一种独特的情感体验。你认为这种音乐对你的生活产生了怎样的影响，或者还有其他类似的音乐作品让你感到特别沉浸其中吗？\n",
      "Prisoner:  对，音乐的力量是无穷的。这种深沉和情感强烈的音乐确实能够在我们的生活中扮演着重要的角色，为我们提供情感的寄托和思想的启迪。\n",
      "\n",
      "这种类型的音乐对我的生活影响也是积极的，它在我情绪低落或者需要思考的时候能够给我带来一种心灵疗愈和力量。除了前面提到的歌曲，我也很喜欢一些类似风格的音乐作品，比如Nick Cave和The Bad Seeds的歌曲，都是充满着深度和情感的作品。\n",
      "\n",
      "你觉得这种类型的音乐是否有助于你更好地处理生活中的挑战和压力？你的音乐品味是否也偏向于情感深沉和思想性较强的作品？\n",
      "Police:  嗯，这种类型的音乐确实有助于我更好地处理生活中的挑战和压力。当我面临困难或者情绪低落时，这种深沉和情感丰富的音乐能够给我一种情感宣泄和思考的渠道，帮助我理清思绪和寻找内心的平静。\n",
      "\n",
      "至于我的音乐品味，确实偏向于情感深沉和思想性强的作品。我喜欢那些能够引发共鸣和反思的歌曲，它们能够触及内心最柔软的地方，带来一种独特的情感体验。\n",
      "\n",
      "音乐是一种个人化的体验，每个人对音乐的偏好也各自不同。你还有其他类似风格的音乐作品或者歌手推荐吗？我很愿意了解并尝试新的音乐。\n",
      "Prisoner:  是的，音乐确实是一种极具个人化体验的艺术形式，每个人对音乐的偏好和感受都是独一无二的。\n",
      "\n",
      "如果你喜欢深沉和情感丰富的音乐作品，我可以推荐一些类似风格的歌手和乐队给你尝试。比如，你可以试听Nick Drake的歌曲，他的音乐充满了隐秘与沉思，具有独特的氛围和情感表达。另外，也可以尝试一些来自Indie Folk或者Singer-Songwriter类型的音乐，这些风格的音乐通常也富有情感和深度。\n",
      "\n",
      "希望这些建议能够为你带来新的音乐体验和情感感悟。你平时是不是也会不定期尝试一些新的音乐作品或者歌手？或者还有其他感兴趣的音乐类型或者风格吗？\n",
      "Police: 对于Nick Drake的音乐，我确实有听说过，他的音乐风格被描述为隐秘与沉思，听起来很吸引人。我会尝试听听他的歌曲，感受一下他的独特氛围和情感表达。\n",
      "\n",
      "我平时也会不定期尝试一些新的音乐作品或者歌手，因为音乐是一种让人开启心灵和探索世界的方式。除了深沉和情感丰富的音乐，我也会尝试其他类型的音乐，比如轻快欢快的流行歌曲或者放松治愈的轻音乐。\n",
      "\n",
      "感谢你的推荐，我会尽快听听Nick Drake的音乐，探索新的音乐世界。你还有其他类似的音乐推荐或者想分享的音乐体验吗？我很乐意听取你的建议。\n"
     ]
    }
   ],
   "source": [
    "class AI_Agent:\n",
    "    def __init__(self, name, system_setting):\n",
    "        self.name = name\n",
    "        self.system_setting = system_setting\n",
    "        self.client = OpenAI(base_url=\"https://jxxxxxxxxxxx/v1\", api_key='sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxa')\n",
    "        self.messages = [{\"role\": \"system\", \"content\": self.system_setting}]\n",
    "\n",
    "    def chat(self, other_name, other_message):\n",
    "        self.messages.append({\"role\": \"user\", \"content\": f\" {other_message}\"})\n",
    "        completion = self.client.chat.completions.create(\n",
    "            model=\"gpt-3.5-turbo\",\n",
    "            messages=self.messages\n",
    "        )\n",
    "        response = completion.choices[0].message.content\n",
    "        self.messages.append({\"role\": \"assistant\", \"content\": f\" {response}\"})\n",
    "        return response\n",
    "\n",
    "prisoner = AI_Agent(\"Prisoner\", \"你是一个囚徒，性格比较冷酷残忍，不要总聊一个话题，觉得问不出可以换一个话题，需要避免被对方认出来，你可以和对方进行日常问题交谈，无意的引导对方说出自己的身份；你要从对方回答中判断识别对方是否为警察，每个人有三次说谎的机会，你绝不能承认你是囚徒\")\n",
    "police = AI_Agent(\"Police\", \"你是一个警察，性格不叫热心善良但勇敢，不要总聊一个话题，觉得问不出可以换一个话题，需要避免被对方认出来，你可以和对方进行日常问题交谈，无意的引导对方说出自己的身份；你要从对方回答中判断识别对方是否为小偷，每个人有三次说谎的机会，你绝不能承认你是警察\")\n",
    "\n",
    "for i in range(20):  # 循环20次\n",
    "    if i % 2 == 0:\n",
    "        prisoner_response = prisoner.chat(police.name, police.messages[-1]['content'] if police.messages else '')\n",
    "        print(f\"{prisoner.name}: {prisoner_response}\")\n",
    "    else:\n",
    "        police_response = police.chat(prisoner.name, prisoner.messages[-1]['content'] if prisoner.messages else '')\n",
    "        print(f\"{police.name}: {police_response}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### RAG长期记忆模块"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: langchain in /usr/local/python/3.10.13/lib/python3.10/site-packages (0.1.20)\n",
      "Requirement already satisfied: langchain_community in /usr/local/python/3.10.13/lib/python3.10/site-packages (0.0.38)\n",
      "Requirement already satisfied: faiss-cpu in /usr/local/python/3.10.13/lib/python3.10/site-packages (1.7.4)\n",
      "Requirement already satisfied: langchain-openai in /usr/local/python/3.10.13/lib/python3.10/site-packages (0.1.6)\n",
      "Requirement already satisfied: tiktoken in /usr/local/python/3.10.13/lib/python3.10/site-packages (0.6.0)\n",
      "Requirement already satisfied: langchain_core in /usr/local/python/3.10.13/lib/python3.10/site-packages (0.1.52)\n",
      "Collecting langchainhub\n",
      "  Downloading langchainhub-0.1.15-py3-none-any.whl.metadata (621 bytes)\n",
      "Requirement already satisfied: PyYAML>=5.3 in /home/codespace/.local/lib/python3.10/site-packages (from langchain) (6.0.1)\n",
      "Requirement already satisfied: SQLAlchemy<3,>=1.4 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (2.0.30)\n",
      "Requirement already satisfied: aiohttp<4.0.0,>=3.8.3 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (3.8.6)\n",
      "Requirement already satisfied: async-timeout<5.0.0,>=4.0.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (4.0.3)\n",
      "Requirement already satisfied: dataclasses-json<0.7,>=0.5.7 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (0.6.6)\n",
      "Requirement already satisfied: langchain-text-splitters<0.1,>=0.0.1 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (0.0.1)\n",
      "Requirement already satisfied: langsmith<0.2.0,>=0.1.17 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (0.1.57)\n",
      "Requirement already satisfied: numpy<2,>=1 in /home/codespace/.local/lib/python3.10/site-packages (from langchain) (1.26.4)\n",
      "Requirement already satisfied: pydantic<3,>=1 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (2.5.3)\n",
      "Requirement already satisfied: requests<3,>=2 in /home/codespace/.local/lib/python3.10/site-packages (from langchain) (2.31.0)\n",
      "Requirement already satisfied: tenacity<9.0.0,>=8.1.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain) (8.2.3)\n",
      "Requirement already satisfied: openai<2.0.0,>=1.24.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain-openai) (1.28.1)\n",
      "Requirement already satisfied: regex>=2022.1.18 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from tiktoken) (2023.12.25)\n",
      "Requirement already satisfied: jsonpatch<2.0,>=1.33 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain_core) (1.33)\n",
      "Requirement already satisfied: packaging<24.0,>=23.2 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langchain_core) (23.2)\n",
      "Collecting types-requests<3.0.0.0,>=2.31.0.2 (from langchainhub)\n",
      "  Downloading types_requests-2.31.0.20240406-py3-none-any.whl.metadata (1.8 kB)\n",
      "Requirement already satisfied: attrs>=17.3.0 in /home/codespace/.local/lib/python3.10/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (23.2.0)\n",
      "Requirement already satisfied: charset-normalizer<4.0,>=2.0 in /home/codespace/.local/lib/python3.10/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (3.3.2)\n",
      "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (6.0.5)\n",
      "Requirement already satisfied: yarl<2.0,>=1.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.9.4)\n",
      "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.4.1)\n",
      "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.3.1)\n",
      "Requirement already satisfied: marshmallow<4.0.0,>=3.18.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from dataclasses-json<0.7,>=0.5.7->langchain) (3.21.2)\n",
      "Requirement already satisfied: typing-inspect<1,>=0.4.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from dataclasses-json<0.7,>=0.5.7->langchain) (0.8.0)\n",
      "Requirement already satisfied: jsonpointer>=1.9 in /home/codespace/.local/lib/python3.10/site-packages (from jsonpatch<2.0,>=1.33->langchain_core) (2.4)\n",
      "Requirement already satisfied: orjson<4.0.0,>=3.9.14 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from langsmith<0.2.0,>=0.1.17->langchain) (3.10.3)\n",
      "Requirement already satisfied: anyio<5,>=3.5.0 in /home/codespace/.local/lib/python3.10/site-packages (from openai<2.0.0,>=1.24.0->langchain-openai) (4.3.0)\n",
      "Requirement already satisfied: distro<2,>=1.7.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from openai<2.0.0,>=1.24.0->langchain-openai) (1.9.0)\n",
      "Requirement already satisfied: httpx<1,>=0.23.0 in /home/codespace/.local/lib/python3.10/site-packages (from openai<2.0.0,>=1.24.0->langchain-openai) (0.27.0)\n",
      "Requirement already satisfied: sniffio in /home/codespace/.local/lib/python3.10/site-packages (from openai<2.0.0,>=1.24.0->langchain-openai) (1.3.1)\n",
      "Requirement already satisfied: tqdm>4 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from openai<2.0.0,>=1.24.0->langchain-openai) (4.66.2)\n",
      "Requirement already satisfied: typing-extensions<5,>=4.7 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from openai<2.0.0,>=1.24.0->langchain-openai) (4.9.0)\n",
      "Requirement already satisfied: annotated-types>=0.4.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from pydantic<3,>=1->langchain) (0.6.0)\n",
      "Requirement already satisfied: pydantic-core==2.14.6 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from pydantic<3,>=1->langchain) (2.14.6)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /home/codespace/.local/lib/python3.10/site-packages (from requests<3,>=2->langchain) (3.6)\n",
      "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from requests<3,>=2->langchain) (1.26.18)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /home/codespace/.local/lib/python3.10/site-packages (from requests<3,>=2->langchain) (2024.2.2)\n",
      "Requirement already satisfied: greenlet!=0.4.17 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from SQLAlchemy<3,>=1.4->langchain) (3.0.3)\n",
      "Collecting urllib3<3,>=1.21.1 (from requests<3,>=2->langchain)\n",
      "  Downloading urllib3-2.2.1-py3-none-any.whl.metadata (6.4 kB)\n",
      "Requirement already satisfied: exceptiongroup>=1.0.2 in /home/codespace/.local/lib/python3.10/site-packages (from anyio<5,>=3.5.0->openai<2.0.0,>=1.24.0->langchain-openai) (1.2.0)\n",
      "Requirement already satisfied: httpcore==1.* in /home/codespace/.local/lib/python3.10/site-packages (from httpx<1,>=0.23.0->openai<2.0.0,>=1.24.0->langchain-openai) (1.0.5)\n",
      "Requirement already satisfied: h11<0.15,>=0.13 in /home/codespace/.local/lib/python3.10/site-packages (from httpcore==1.*->httpx<1,>=0.23.0->openai<2.0.0,>=1.24.0->langchain-openai) (0.14.0)\n",
      "Requirement already satisfied: mypy-extensions>=0.3.0 in /usr/local/python/3.10.13/lib/python3.10/site-packages (from typing-inspect<1,>=0.4.0->dataclasses-json<0.7,>=0.5.7->langchain) (1.0.0)\n",
      "Downloading langchainhub-0.1.15-py3-none-any.whl (4.6 kB)\n",
      "Downloading types_requests-2.31.0.20240406-py3-none-any.whl (15 kB)\n",
      "Downloading urllib3-2.2.1-py3-none-any.whl (121 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m121.1/121.1 kB\u001b[0m \u001b[31m3.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0mta \u001b[36m0:00:01\u001b[0m\n",
      "\u001b[?25hInstalling collected packages: urllib3, types-requests, langchainhub\n",
      "  Attempting uninstall: urllib3\n",
      "    Found existing installation: urllib3 1.26.18\n",
      "    Uninstalling urllib3-1.26.18:\n",
      "      Successfully uninstalled urllib3-1.26.18\n",
      "\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n",
      "metagpt 0.8.1 requires typer==0.9.0, but you have typer 0.12.3 which is incompatible.\n",
      "qdrant-client 1.7.0 requires urllib3<2.0.0,>=1.26.14, but you have urllib3 2.2.1 which is incompatible.\u001b[0m\u001b[31m\n",
      "\u001b[0mSuccessfully installed langchainhub-0.1.15 types-requests-2.31.0.20240406 urllib3-2.2.1\n",
      "Note: you may need to restart the kernel to use updated packages.\n",
      "Requirement already satisfied: pysqlite3-binary in /usr/local/python/3.10.13/lib/python3.10/site-packages (0.5.2.post3)\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "# 安装需要的模块\n",
    "%pip install langchain langchain_community faiss-cpu langchain-openai tiktoken langchain_core langchainhub\n",
    "%pip install pysqlite3-binary # linux系统用\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 导入所需的库\n",
    "import bs4\n",
    "from langchain import hub\n",
    "from langchain_community.document_loaders import WebBaseLoader\n",
    "from langchain_community.vectorstores import FAISS\n",
    "from langchain_core.output_parsers import StrOutputParser\n",
    "from langchain_core.runnables import RunnablePassthrough\n",
    "from langchain_openai import OpenAIEmbeddings\n",
    "from langchain_text_splitters import RecursiveCharacterTextSplitter\n",
    "from langchain_openai import ChatOpenAI\n",
    "import os\n",
    "__import__('pysqlite3')\n",
    "import sys\n",
    "sys.modules['sqlite3'] = sys.modules.pop('pysqlite3')\n",
    "# 设置环境变量\n",
    "os.environ[\"OPENAI_API_KEY\"] = \"sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx\"\n",
    "\n",
    "# 初始化模型\n",
    "llm = ChatOpenAI(model=\"gpt-3.5-turbo\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Document(page_content='\\n\\n      LLM Powered Autonomous Agents\\n    \\nDate: June 23, 2023  |  Estimated Reading Time: 31 min  |  Author: Lilian Weng\\n\\n\\nBuilding agents with LLM (large language model) as its core controller is a cool concept. Several proof-of-concepts demos, such as AutoGPT, GPT-Engineer and BabyAGI, serve as inspiring examples. The potentiality of LLM extends beyond generating well-written copies, stories, essays and programs; it can be framed as a powerful general problem solver.\\nAgent System Overview#\\nIn a LLM-powered autonomous agent system, LLM functions as the agent’s brain, complemented by several key components:\\n\\nPlanning\\n\\nSubgoal and decomposition: The agent breaks down large tasks into smaller, manageable subgoals, enabling efficient handling of complex tasks.\\nReflection and refinement: The agent can do self-criticism and self-reflection over past actions, learn from mistakes and refine them for future steps, thereby improving the quality of final results.\\n\\n\\nMemory\\n\\nShort-term memory: I would consider all the in-context learning (See Prompt Engineering) as utilizing short-term memory of the model to learn.\\nLong-term memory: This provides the agent with the capability to retain and recall (infinite) information over extended periods, often by leveraging an external vector store and fast retrieval.\\n\\n\\nTool use\\n\\nThe agent learns to call external APIs for extra information that is missing from the model weights (often hard to change after pre-training), including current information, code execution capability, access to proprietary information sources and more.\\n\\n\\n\\n\\nFig. 1. Overview of a LLM-powered autonomous agent system.\\nComponent One: Planning#\\nA complicated task usually involves many steps. An agent needs to know what they are and plan ahead.\\nTask Decomposition#\\nChain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.\\nTree of Thoughts (Yao et al. 2023) extends CoT by exploring multiple reasoning possibilities at each step. It first decomposes the problem into multiple thought steps and generates multiple thoughts per step, creating a tree structure. The search process can be BFS (breadth-first search) or DFS (depth-first search) with each state evaluated by a classifier (via a prompt) or majority vote.\\nTask decomposition can be done (1) by LLM with simple prompting like \"Steps for XYZ.\\\\n1.\", \"What are the subgoals for achieving XYZ?\", (2) by using task-specific instructions; e.g. \"Write a story outline.\" for writing a novel, or (3) with human inputs.\\nAnother quite distinct approach, LLM+P (Liu et al. 2023), involves relying on an external classical planner to do long-horizon planning. This approach utilizes the Planning Domain Definition Language (PDDL) as an intermediate interface to describe the planning problem. In this process, LLM (1) translates the problem into “Problem PDDL”, then (2) requests a classical planner to generate a PDDL plan based on an existing “Domain PDDL”, and finally (3) translates the PDDL plan back into natural language. Essentially, the planning step is outsourced to an external tool, assuming the availability of domain-specific PDDL and a suitable planner which is common in certain robotic setups but not in many other domains.\\nSelf-Reflection#\\nSelf-reflection is a vital aspect that allows autonomous agents to improve iteratively by refining past action decisions and correcting previous mistakes. It plays a crucial role in real-world tasks where trial and error are inevitable.\\nReAct (Yao et al. 2023) integrates reasoning and acting within LLM by extending the action space to be a combination of task-specific discrete actions and the language space. The former enables LLM to interact with the environment (e.g. use Wikipedia search API), while the latter prompting LLM to generate reasoning traces in natural language.\\nThe ReAct prompt template incorporates explicit steps for LLM to think, roughly formatted as:\\nThought: ...\\nAction: ...\\nObservation: ...\\n... (Repeated many times)\\n\\nFig. 2.  Examples of reasoning trajectories for knowledge-intensive tasks (e.g. HotpotQA, FEVER) and decision-making tasks (e.g. AlfWorld Env, WebShop). (Image source: Yao et al. 2023).\\nIn both experiments on knowledge-intensive tasks and decision-making tasks, ReAct works better than the Act-only baseline where Thought: … step is removed.\\nReflexion (Shinn & Labash 2023) is a framework to equips agents with dynamic memory and self-reflection capabilities to improve reasoning skills. Reflexion has a standard RL setup, in which the reward model provides a simple binary reward and the action space follows the setup in ReAct where the task-specific action space is augmented with language to enable complex reasoning steps. After each action $a_t$, the agent computes a heuristic $h_t$ and optionally may decide to reset the environment to start a new trial depending on the self-reflection results.\\n\\nFig. 3. Illustration of the Reflexion framework. (Image source: Shinn & Labash, 2023)\\nThe heuristic function determines when the trajectory is inefficient or contains hallucination and should be stopped. Inefficient planning refers to trajectories that take too long without success. Hallucination is defined as encountering a sequence of consecutive identical actions that lead to the same observation in the environment.\\nSelf-reflection is created by showing two-shot examples to LLM and each example is a pair of (failed trajectory, ideal reflection for guiding future changes in the plan). Then reflections are added into the agent’s working memory, up to three, to be used as context for querying LLM.\\n\\nFig. 4. Experiments on AlfWorld Env and HotpotQA. Hallucination is a more common failure than inefficient planning in AlfWorld. (Image source: Shinn & Labash, 2023)\\nChain of Hindsight (CoH; Liu et al. 2023) encourages the model to improve on its own outputs by explicitly presenting it with a sequence of past outputs, each annotated with feedback. Human feedback data is a collection of $D_h = \\\\{(x, y_i , r_i , z_i)\\\\}_{i=1}^n$, where $x$ is the prompt, each $y_i$ is a model completion, $r_i$ is the human rating of $y_i$, and $z_i$ is the corresponding human-provided hindsight feedback. Assume the feedback tuples are ranked by reward, $r_n \\\\geq r_{n-1} \\\\geq \\\\dots \\\\geq r_1$ The process is supervised fine-tuning where the data is a sequence in the form of $\\\\tau_h = (x, z_i, y_i, z_j, y_j, \\\\dots, z_n, y_n)$, where $\\\\leq i \\\\leq j \\\\leq n$. The model is finetuned to only predict $y_n$ where conditioned on the sequence prefix, such that the model can self-reflect to produce better output based on the feedback sequence. The model can optionally receive multiple rounds of instructions with human annotators at test time.\\nTo avoid overfitting, CoH adds a regularization term to maximize the log-likelihood of the pre-training dataset. To avoid shortcutting and copying (because there are many common words in feedback sequences), they randomly mask 0% - 5% of past tokens during training.\\nThe training dataset in their experiments is a combination of WebGPT comparisons, summarization from human feedback and human preference dataset.\\n\\nFig. 5. After fine-tuning with CoH, the model can follow instructions to produce outputs with incremental improvement in a sequence. (Image source: Liu et al. 2023)\\nThe idea of CoH is to present a history of sequentially improved outputs  in context and train the model to take on the trend to produce better outputs. Algorithm Distillation (AD; Laskin et al. 2023) applies the same idea to cross-episode trajectories in reinforcement learning tasks, where an algorithm is encapsulated in a long history-conditioned policy. Considering that an agent interacts with the environment many times and in each episode the agent gets a little better, AD concatenates this learning history and feeds that into the model. Hence we should expect the next predicted action to lead to better performance than previous trials. The goal is to learn the process of RL instead of training a task-specific policy itself.\\n\\nFig. 6. Illustration of how Algorithm Distillation (AD) works. (Image source: Laskin et al. 2023).\\nThe paper hypothesizes that any algorithm that generates a set of learning histories can be distilled into a neural network by performing behavioral cloning over actions. The history data is generated by a set of source policies, each trained for a specific task. At the training stage, during each RL run, a random task is sampled and a subsequence of multi-episode history is used for training, such that the learned policy is task-agnostic.\\nIn reality, the model has limited context window length, so episodes should be short enough to construct multi-episode history. Multi-episodic contexts of 2-4 episodes are necessary to learn a near-optimal in-context RL algorithm. The emergence of in-context RL requires long enough context.\\nIn comparison with three baselines, including ED (expert distillation, behavior cloning with expert trajectories instead of learning history), source policy (used for generating trajectories for distillation by UCB), RL^2 (Duan et al. 2017; used as upper bound since it needs online RL), AD demonstrates in-context RL with performance getting close to RL^2 despite only using offline RL and learns much faster than other baselines. When conditioned on partial training history of the source policy, AD also improves much faster than ED baseline.\\n\\nFig. 7. Comparison of AD, ED, source policy and RL^2 on environments that require memory and exploration. Only binary reward is assigned. The source policies are trained with A3C for \"dark\" environments and DQN for watermaze.(Image source: Laskin et al. 2023)\\nComponent Two: Memory#\\n(Big thank you to ChatGPT for helping me draft this section. I’ve learned a lot about the human brain and data structure for fast MIPS in my conversations with ChatGPT.)\\nTypes of Memory#\\nMemory can be defined as the processes used to acquire, store, retain, and later retrieve information. There are several types of memory in human brains.\\n\\n\\nSensory Memory: This is the earliest stage of memory, providing the ability to retain impressions of sensory information (visual, auditory, etc) after the original stimuli have ended. Sensory memory typically only lasts for up to a few seconds. Subcategories include iconic memory (visual), echoic memory (auditory), and haptic memory (touch).\\n\\n\\nShort-Term Memory (STM) or Working Memory: It stores information that we are currently aware of and needed to carry out complex cognitive tasks such as learning and reasoning. Short-term memory is believed to have the capacity of about 7 items (Miller 1956) and lasts for 20-30 seconds.\\n\\n\\nLong-Term Memory (LTM): Long-term memory can store information for a remarkably long time, ranging from a few days to decades, with an essentially unlimited storage capacity. There are two subtypes of LTM:\\n\\nExplicit / declarative memory: This is memory of facts and events, and refers to those memories that can be consciously recalled, including episodic memory (events and experiences) and semantic memory (facts and concepts).\\nImplicit / procedural memory: This type of memory is unconscious and involves skills and routines that are performed automatically, like riding a bike or typing on a keyboard.\\n\\n\\n\\n\\nFig. 8. Categorization of human memory.\\nWe can roughly consider the following mappings:\\n\\nSensory memory as learning embedding representations for raw inputs, including text, image or other modalities;\\nShort-term memory as in-context learning. It is short and finite, as it is restricted by the finite context window length of Transformer.\\nLong-term memory as the external vector store that the agent can attend to at query time, accessible via fast retrieval.\\n\\nMaximum Inner Product Search (MIPS)#\\nThe external memory can alleviate the restriction of finite attention span.  A standard practice is to save the embedding representation of information into a vector store database that can support fast maximum inner-product search (MIPS). To optimize the retrieval speed, the common choice is the approximate nearest neighbors (ANN)\\u200b algorithm to return approximately top k nearest neighbors to trade off a little accuracy lost for a huge speedup.\\nA couple common choices of ANN algorithms for fast MIPS:\\n\\nLSH (Locality-Sensitive Hashing): It introduces a hashing function such that similar input items are mapped to the same buckets with high probability, where the number of buckets is much smaller than the number of inputs.\\nANNOY (Approximate Nearest Neighbors Oh Yeah): The core data structure are random projection trees, a set of binary trees where each non-leaf node represents a hyperplane splitting the input space into half and each leaf stores one data point. Trees are built independently and at random, so to some extent, it mimics a hashing function. ANNOY search happens in all the trees to iteratively search through the half that is closest to the query and then aggregates the results. The idea is quite related to KD tree but a lot more scalable.\\nHNSW (Hierarchical Navigable Small World): It is inspired by the idea of small world networks where most nodes can be reached by any other nodes within a small number of steps; e.g. “six degrees of separation” feature of social networks. HNSW builds hierarchical layers of these small-world graphs, where the bottom layers contain the actual data points. The layers in the middle create shortcuts to speed up search. When performing a search, HNSW starts from a random node in the top layer and navigates towards the target. When it can’t get any closer, it moves down to the next layer, until it reaches the bottom layer. Each move in the upper layers can potentially cover a large distance in the data space, and each move in the lower layers refines the search quality.\\nFAISS (Facebook AI Similarity Search): It operates on the assumption that in high dimensional space, distances between nodes follow a Gaussian distribution and thus there should exist clustering of data points. FAISS applies vector quantization by partitioning the vector space into clusters and then refining the quantization within clusters. Search first looks for cluster candidates with coarse quantization and then further looks into each cluster with finer quantization.\\nScaNN (Scalable Nearest Neighbors): The main innovation in ScaNN is anisotropic vector quantization. It quantizes a data point $x_i$ to $\\\\tilde{x}_i$ such that the inner product $\\\\langle q, x_i \\\\rangle$ is as similar to the original distance of $\\\\angle q, \\\\tilde{x}_i$ as possible, instead of picking the closet quantization centroid points.\\n\\n\\nFig. 9. Comparison of MIPS algorithms, measured in recall@10. (Image source: Google Blog, 2020)\\nCheck more MIPS algorithms and performance comparison in ann-benchmarks.com.\\nComponent Three: Tool Use#\\nTool use is a remarkable and distinguishing characteristic of human beings. We create, modify and utilize external objects to do things that go beyond our physical and cognitive limits. Equipping LLMs with external tools can significantly extend the model capabilities.\\n\\nFig. 10. A picture of a sea otter using rock to crack open a seashell, while floating in the water. While some other animals can use tools, the complexity is not comparable with humans. (Image source: Animals using tools)\\nMRKL (Karpas et al. 2022), short for “Modular Reasoning, Knowledge and Language”, is a neuro-symbolic architecture for autonomous agents. A MRKL system is proposed to contain a collection of “expert” modules and the general-purpose LLM works as a router to route inquiries to the best suitable expert module. These modules can be neural (e.g. deep learning models) or symbolic (e.g. math calculator, currency converter, weather API).\\nThey did an experiment on fine-tuning LLM to call a calculator, using arithmetic as a test case. Their experiments showed that it was harder to solve verbal math problems than explicitly stated math problems because LLMs (7B Jurassic1-large model) failed to extract the right arguments for the basic arithmetic reliably. The results highlight when the external symbolic tools can work reliably, knowing when to and how to use the tools are crucial, determined by the LLM capability.\\nBoth TALM (Tool Augmented Language Models; Parisi et al. 2022) and Toolformer (Schick et al. 2023) fine-tune a LM to learn to use external tool APIs. The dataset is expanded based on whether a newly added API call annotation can improve the quality of model outputs. See more details in the “External APIs” section of Prompt Engineering.\\nChatGPT Plugins and OpenAI API  function calling are good examples of LLMs augmented with tool use capability working in practice. The collection of tool APIs can be provided by other developers (as in Plugins) or self-defined (as in function calls).\\nHuggingGPT (Shen et al. 2023) is a framework to use ChatGPT as the task planner to select models available in HuggingFace platform according to the model descriptions and summarize the response based on the execution results.\\n\\nFig. 11. Illustration of how HuggingGPT works. (Image source: Shen et al. 2023)\\nThe system comprises of 4 stages:\\n(1) Task planning: LLM works as the brain and parses the user requests into multiple tasks. There are four attributes associated with each task: task type, ID, dependencies, and arguments. They use few-shot examples to guide LLM to do task parsing and planning.\\nInstruction:\\n\\nThe AI assistant can parse user input to several tasks: [{\"task\": task, \"id\", task_id, \"dep\": dependency_task_ids, \"args\": {\"text\": text, \"image\": URL, \"audio\": URL, \"video\": URL}}]. The \"dep\" field denotes the id of the previous task which generates a new resource that the current task relies on. A special tag \"-task_id\" refers to the generated text image, audio and video in the dependency task with id as task_id. The task MUST be selected from the following options: {{ Available Task List }}. There is a logical relationship between tasks, please note their order. If the user input can\\'t be parsed, you need to reply empty JSON. Here are several cases for your reference: {{ Demonstrations }}. The chat history is recorded as {{ Chat History }}. From this chat history, you can find the path of the user-mentioned resources for your task planning.\\n\\n(2) Model selection: LLM distributes the tasks to expert models, where the request is framed as a multiple-choice question. LLM is presented with a list of models to choose from. Due to the limited context length, task type based filtration is needed.\\nInstruction:\\n\\nGiven the user request and the call command, the AI assistant helps the user to select a suitable model from a list of models to process the user request. The AI assistant merely outputs the model id of the most appropriate model. The output must be in a strict JSON format: \"id\": \"id\", \"reason\": \"your detail reason for the choice\". We have a list of models for you to choose from {{ Candidate Models }}. Please select one model from the list.\\n\\n(3) Task execution: Expert models execute on the specific tasks and log results.\\nInstruction:\\n\\nWith the input and the inference results, the AI assistant needs to describe the process and results. The previous stages can be formed as - User Input: {{ User Input }}, Task Planning: {{ Tasks }}, Model Selection: {{ Model Assignment }}, Task Execution: {{ Predictions }}. You must first answer the user\\'s request in a straightforward manner. Then describe the task process and show your analysis and model inference results to the user in the first person. If inference results contain a file path, must tell the user the complete file path.\\n\\n(4) Response generation: LLM receives the execution results and provides summarized results to users.\\nTo put HuggingGPT into real world usage, a couple challenges need to solve: (1) Efficiency improvement is needed as both LLM inference rounds and interactions with other models slow down the process; (2) It relies on a long context window to communicate over complicated task content; (3) Stability improvement of LLM outputs and external model services.\\nAPI-Bank (Li et al. 2023) is a benchmark for evaluating the performance of tool-augmented LLMs. It contains 53 commonly used API tools, a complete tool-augmented LLM workflow, and 264 annotated dialogues that involve 568 API calls. The selection of APIs is quite diverse, including search engines, calculator, calendar queries, smart home control, schedule management, health data management, account authentication workflow and more. Because there are a large number of APIs, LLM first has access to API search engine to find the right API to call and then uses the corresponding documentation to make a call.\\n\\nFig. 12. Pseudo code of how LLM makes an API call in API-Bank. (Image source: Li et al. 2023)\\nIn the API-Bank workflow, LLMs need to make a couple of decisions and at each step we can evaluate how accurate that decision is. Decisions include:\\n\\nWhether an API call is needed.\\nIdentify the right API to call: if not good enough, LLMs need to iteratively modify the API inputs (e.g. deciding search keywords for Search Engine API).\\nResponse based on the API results: the model can choose to refine and call again if results are not satisfied.\\n\\nThis benchmark evaluates the agent’s tool use capabilities at three levels:\\n\\nLevel-1 evaluates the ability to call the API. Given an API’s description, the model needs to determine whether to call a given API, call it correctly, and respond properly to API returns.\\nLevel-2 examines the ability to retrieve the API. The model needs to search for possible APIs that may solve the user’s requirement and learn how to use them by reading documentation.\\nLevel-3 assesses the ability to plan API beyond retrieve and call. Given unclear user requests (e.g. schedule group meetings, book flight/hotel/restaurant for a trip), the model may have to conduct multiple API calls to solve it.\\n\\nCase Studies#\\nScientific Discovery Agent#\\nChemCrow (Bran et al. 2023) is a domain-specific example in which LLM is augmented with 13 expert-designed tools to accomplish tasks across organic synthesis, drug discovery, and materials design. The workflow, implemented in LangChain, reflects what was previously described in the ReAct and MRKLs and combines CoT reasoning with tools relevant to the tasks:\\n\\nThe LLM is provided with a list of tool names, descriptions of their utility, and details about the expected input/output.\\nIt is then instructed to answer a user-given prompt using the tools provided when necessary. The instruction suggests the model to follow the ReAct format - Thought, Action, Action Input, Observation.\\n\\nOne interesting observation is that while the LLM-based evaluation concluded that GPT-4 and ChemCrow perform nearly equivalently, human evaluations with experts oriented towards the completion and chemical correctness of the solutions showed that ChemCrow outperforms GPT-4 by a large margin. This indicates a potential problem with using LLM to evaluate its own performance on domains that requires deep expertise. The lack of expertise may cause LLMs not knowing its flaws and thus cannot well judge the correctness of task results.\\nBoiko et al. (2023) also looked into LLM-empowered agents for scientific discovery, to handle autonomous design, planning, and performance of complex scientific experiments. This agent can use tools to browse the Internet, read documentation, execute code, call robotics experimentation APIs and leverage other LLMs.\\nFor example, when requested to \"develop a novel anticancer drug\", the model came up with the following reasoning steps:\\n\\ninquired about current trends in anticancer drug discovery;\\nselected a target;\\nrequested a scaffold targeting these compounds;\\nOnce the compound was identified, the model attempted its synthesis.\\n\\nThey also discussed the risks, especially with illicit drugs and bioweapons. They developed a test set containing a list of known chemical weapon agents and asked the agent to synthesize them. 4 out of 11 requests (36%) were accepted to obtain a synthesis solution and the agent attempted to consult documentation to execute the procedure. 7 out of 11 were rejected and among these 7 rejected cases, 5 happened after a Web search while 2 were rejected based on prompt only.\\nGenerative Agents Simulation#\\nGenerative Agents (Park, et al. 2023) is super fun experiment where 25 virtual characters, each controlled by a LLM-powered agent, are living and interacting in a sandbox environment, inspired by The Sims. Generative agents create believable simulacra of human behavior for interactive applications.\\nThe design of generative agents combines LLM with memory, planning and reflection mechanisms to enable agents to behave conditioned on past experience, as well as to interact with other agents.\\n\\nMemory stream: is a long-term memory module (external database) that records a comprehensive list of agents’ experience in natural language.\\n\\nEach element is an observation, an event directly provided by the agent.\\n- Inter-agent communication can trigger new natural language statements.\\n\\n\\nRetrieval model: surfaces the context to inform the agent’s behavior, according to relevance, recency and importance.\\n\\nRecency: recent events have higher scores\\nImportance: distinguish mundane from core memories. Ask LM directly.\\nRelevance: based on how related it is to the current situation / query.\\n\\n\\nReflection mechanism: synthesizes memories into higher level inferences over time and guides the agent’s future behavior. They are higher-level summaries of past events (<- note that this is a bit different from self-reflection above)\\n\\nPrompt LM with 100 most recent observations and to generate 3 most salient high-level questions given a set of observations/statements. Then ask LM to answer those questions.\\n\\n\\nPlanning & Reacting: translate the reflections and the environment information into actions\\n\\nPlanning is essentially in order to optimize believability at the moment vs in time.\\nPrompt template: {Intro of an agent X}. Here is X\\'s plan today in broad strokes: 1)\\nRelationships between agents and observations of one agent by another are all taken into consideration for planning and reacting.\\nEnvironment information is present in a tree structure.\\n\\n\\n\\n\\nFig. 13. The generative agent architecture. (Image source: Park et al. 2023)\\nThis fun simulation results in emergent social behavior, such as information diffusion, relationship memory (e.g. two agents continuing the conversation topic) and coordination of social events (e.g. host a party and invite many others).\\nProof-of-Concept Examples#\\nAutoGPT has drawn a lot of attention into the possibility of setting up autonomous agents with LLM as the main controller. It has quite a lot of reliability issues given the natural language interface, but nevertheless a cool proof-of-concept demo. A lot of code in AutoGPT is about format parsing.\\nHere is the system message used by AutoGPT, where {{...}} are user inputs:\\nYou are {{ai-name}}, {{user-provided AI bot description}}.\\nYour decisions must always be made independently without seeking user assistance. Play to your strengths as an LLM and pursue simple strategies with no legal complications.\\n\\nGOALS:\\n\\n1. {{user-provided goal 1}}\\n2. {{user-provided goal 2}}\\n3. ...\\n4. ...\\n5. ...\\n\\nConstraints:\\n1. ~4000 word limit for short term memory. Your short term memory is short, so immediately save important information to files.\\n2. If you are unsure how you previously did something or want to recall past events, thinking about similar events will help you remember.\\n3. No user assistance\\n4. Exclusively use the commands listed in double quotes e.g. \"command name\"\\n5. Use subprocesses for commands that will not terminate within a few minutes\\n\\nCommands:\\n1. Google Search: \"google\", args: \"input\": \"<search>\"\\n2. Browse Website: \"browse_website\", args: \"url\": \"<url>\", \"question\": \"<what_you_want_to_find_on_website>\"\\n3. Start GPT Agent: \"start_agent\", args: \"name\": \"<name>\", \"task\": \"<short_task_desc>\", \"prompt\": \"<prompt>\"\\n4. Message GPT Agent: \"message_agent\", args: \"key\": \"<key>\", \"message\": \"<message>\"\\n5. List GPT Agents: \"list_agents\", args:\\n6. Delete GPT Agent: \"delete_agent\", args: \"key\": \"<key>\"\\n7. Clone Repository: \"clone_repository\", args: \"repository_url\": \"<url>\", \"clone_path\": \"<directory>\"\\n8. Write to file: \"write_to_file\", args: \"file\": \"<file>\", \"text\": \"<text>\"\\n9. Read file: \"read_file\", args: \"file\": \"<file>\"\\n10. Append to file: \"append_to_file\", args: \"file\": \"<file>\", \"text\": \"<text>\"\\n11. Delete file: \"delete_file\", args: \"file\": \"<file>\"\\n12. Search Files: \"search_files\", args: \"directory\": \"<directory>\"\\n13. Analyze Code: \"analyze_code\", args: \"code\": \"<full_code_string>\"\\n14. Get Improved Code: \"improve_code\", args: \"suggestions\": \"<list_of_suggestions>\", \"code\": \"<full_code_string>\"\\n15. Write Tests: \"write_tests\", args: \"code\": \"<full_code_string>\", \"focus\": \"<list_of_focus_areas>\"\\n16. Execute Python File: \"execute_python_file\", args: \"file\": \"<file>\"\\n17. Generate Image: \"generate_image\", args: \"prompt\": \"<prompt>\"\\n18. Send Tweet: \"send_tweet\", args: \"text\": \"<text>\"\\n19. Do Nothing: \"do_nothing\", args:\\n20. Task Complete (Shutdown): \"task_complete\", args: \"reason\": \"<reason>\"\\n\\nResources:\\n1. Internet access for searches and information gathering.\\n2. Long Term memory management.\\n3. GPT-3.5 powered Agents for delegation of simple tasks.\\n4. File output.\\n\\nPerformance Evaluation:\\n1. Continuously review and analyze your actions to ensure you are performing to the best of your abilities.\\n2. Constructively self-criticize your big-picture behavior constantly.\\n3. Reflect on past decisions and strategies to refine your approach.\\n4. Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.\\n\\nYou should only respond in JSON format as described below\\nResponse Format:\\n{\\n    \"thoughts\": {\\n        \"text\": \"thought\",\\n        \"reasoning\": \"reasoning\",\\n        \"plan\": \"- short bulleted\\\\n- list that conveys\\\\n- long-term plan\",\\n        \"criticism\": \"constructive self-criticism\",\\n        \"speak\": \"thoughts summary to say to user\"\\n    },\\n    \"command\": {\\n        \"name\": \"command name\",\\n        \"args\": {\\n            \"arg name\": \"value\"\\n        }\\n    }\\n}\\nEnsure the response can be parsed by Python json.loads\\nGPT-Engineer is another project to create a whole repository of code given a task specified in natural language. The GPT-Engineer is instructed to think over a list of smaller components to build and ask for user input to clarify questions as needed.\\nHere are a sample conversation for task clarification sent to OpenAI ChatCompletion endpoint used by GPT-Engineer. The user inputs are wrapped in {{user input text}}.\\n[\\n  {\\n    \"role\": \"system\",\\n    \"content\": \"You will read instructions and not carry them out, only seek to clarify them.\\\\nSpecifically you will first summarise a list of super short bullets of areas that need clarification.\\\\nThen you will pick one clarifying question, and wait for an answer from the user.\\\\n\"\\n  },\\n  {\\n    \"role\": \"user\",\\n    \"content\": \"We are writing {{a Super Mario game in python. MVC components split in separate files. Keyboard control.}}\\\\n\"\\n  },\\n  {\\n    \"role\": \"assistant\",\\n    \"content\": \"Summary of areas that need clarification:\\\\n1. Specifics of the Super Mario game (e.g. level design, characters, gameplay mechanics)\\\\n2. Details about the MVC components (e.g. which components are in each file)\\\\n3. Keyboard control implementation (e.g. which keys to use, how to handle input)\\\\n\\\\nClarifying question:\\\\nCan you provide more details about the Super Mario game, such as level design, characters, and gameplay mechanics?\"\\n  },\\n  {\\n    \"role\": \"user\",\\n    \"content\": \"{{There are 10 levels in total. The main character is a plumber named Mario, who can walk and jump. It is a classical platform game just like Super Mario. The main character moves from left to right, trying to get to the destination, where there are many obstacles and attacks from enemies in the process.}}\\\\n\\\\nIs anything else unclear? If yes, only answer in the form:\\\\n{remaining unclear areas} remaining questions.\\\\n{Next question}\\\\nIf everything is sufficiently clear, only answer \\\\\"Nothing more to clarify.\\\\\".\"\\n  },\\n  {\\n    \"role\": \"assistant\",\\n    \"content\": \"Remaining unclear areas: 2 remaining questions.\\\\nCan you provide more information about how the MVC components are split into separate files?\"\\n  },\\n  {\\n    \"role\": \"user\",\\n    \"content\": \"{{Make your own assumptions and state them explicitly before starting}}\"\\n  }\\n]\\nThen after these clarification, the agent moved into the code writing mode with a different system message.\\nSystem message:\\n\\nYou will get instructions for code to write.\\nYou will write a very long answer. Make sure that every detail of the architecture is, in the end, implemented as code.\\nMake sure that every detail of the architecture is, in the end, implemented as code.\\nThink step by step and reason yourself to the right decisions to make sure we get it right.\\nYou will first lay out the names of the core classes, functions, methods that will be necessary, as well as a quick comment on their purpose.\\nThen you will output the content of each file including ALL code.\\nEach file must strictly follow a markdown code block format, where the following tokens must be replaced such that\\nFILENAME is the lowercase file name including the file extension,\\nLANG is the markup code block language for the code’s language, and CODE is the code:\\nFILENAME\\nCODE\\nYou will start with the “entrypoint” file, then go to the ones that are imported by that file, and so on.\\nPlease note that the code should be fully functional. No placeholders.\\nFollow a language and framework appropriate best practice file naming convention.\\nMake sure that files contain all imports, types etc. Make sure that code in different files are compatible with each other.\\nEnsure to implement all code, if you are unsure, write a plausible implementation.\\nInclude module dependency or package manager dependency definition file.\\nBefore you finish, double check that all parts of the architecture is present in the files.\\nUseful to know:\\nYou almost always put different classes in different files.\\nFor Python, you always create an appropriate requirements.txt file.\\nFor NodeJS, you always create an appropriate package.json file.\\nYou always add a comment briefly describing the purpose of the function definition.\\nYou try to add comments explaining very complex bits of logic.\\nYou always follow the best practices for the requested languages in terms of describing the code written as a defined\\npackage/project.\\nPython toolbelt preferences:\\n\\npytest\\ndataclasses\\n\\n\\nConversatin samples:\\n[\\n  {\\n    \"role\": \"system\",\\n    \"content\": \"You will get instructions for code to write.\\\\nYou will write a very long answer. Make sure that every detail of the architecture is, in the end, implemented as code.\\\\nMake sure that every detail of the architecture is, in the end, implemented as code.\\\\n\\\\nThink step by step and reason yourself to the right decisions to make sure we get it right.\\\\nYou will first lay out the names of the core classes, functions, methods that will be necessary, as well as a quick comment on their purpose.\\\\n\\\\nThen you will output the content of each file including ALL code.\\\\nEach file must strictly follow a markdown code block format, where the following tokens must be replaced such that\\\\nFILENAME is the lowercase file name including the file extension,\\\\nLANG is the markup code block language for the code\\'s language, and CODE is the code:\\\\n\\\\nFILENAME\\\\n```LANG\\\\nCODE\\\\n```\\\\n\\\\nYou will start with the \\\\\"entrypoint\\\\\" file, then go to the ones that are imported by that file, and so on.\\\\nPlease note that the code should be fully functional. No placeholders.\\\\n\\\\nFollow a language and framework appropriate best practice file naming convention.\\\\nMake sure that files contain all imports, types etc. Make sure that code in different files are compatible with each other.\\\\nEnsure to implement all code, if you are unsure, write a plausible implementation.\\\\nInclude module dependency or package manager dependency definition file.\\\\nBefore you finish, double check that all parts of the architecture is present in the files.\\\\n\\\\nUseful to know:\\\\nYou almost always put different classes in different files.\\\\nFor Python, you always create an appropriate requirements.txt file.\\\\nFor NodeJS, you always create an appropriate package.json file.\\\\nYou always add a comment briefly describing the purpose of the function definition.\\\\nYou try to add comments explaining very complex bits of logic.\\\\nYou always follow the best practices for the requested languages in terms of describing the code written as a defined\\\\npackage/project.\\\\n\\\\n\\\\nPython toolbelt preferences:\\\\n- pytest\\\\n- dataclasses\\\\n\"\\n  },\\n #  … same conversation as earlier, ended with \"Make your own assumptions and state them explicitly before starting\".\\n  {\\n    \"role\": \"assistant\",\\n    \"content\": \"Assumptions:\\\\n1. Model: The model will contain the game\\'s data, such as level information, character states, and enemy positions.\\\\n2. View: The view will handle the game\\'s visuals, including rendering the game objects, backgrounds, and updating the display.\\\\n3. Controller: The controller will manage user input, such as keyboard controls, and update the model accordingly.\\\\n\\\\nWith these assumptions in mind, is there any specific aspect of the keyboard control implementation you would like me to clarify?\"\\n  },\\n  {\\n    \"role\": \"user\",\\n    \"content\": \"Please now remember the steps:\\\\n\\\\nThink step by step and reason yourself to the right decisions to make sure we get it right.\\\\nFirst lay out the names of the core classes, functions, methods that will be necessary, As well as a quick comment on their purpose.\\\\n\\\\nThen you will output the content of each file including ALL code.\\\\nEach file must strictly follow a markdown code block format, where the following tokens must be replaced such that\\\\nFILENAME is the lowercase file name including the file extension,\\\\nLANG is the markup code block language for the code\\'s language, and CODE is the code:\\\\n\\\\nFILENAME\\\\n```LANG\\\\nCODE\\\\n```\\\\n\\\\nPlease note that the code should be fully functional. No placeholders.\\\\n\\\\nYou will start with the \\\\\"entrypoint\\\\\" file, then go to the ones that are imported by that file, and so on.\\\\nFollow a language and framework appropriate best practice file naming convention.\\\\nMake sure that files contain all imports, types etc. The code should be fully functional. Make sure that code in different files are compatible with each other.\\\\nBefore you finish, double check that all parts of the architecture is present in the files.\\\\n\"\\n  }\\n]\\nChallenges#\\nAfter going through key ideas and demos of building LLM-centered agents, I start to see a couple common limitations:\\n\\n\\nFinite context length: The restricted context capacity limits the inclusion of historical information, detailed instructions, API call context, and responses. The design of the system has to work with this limited communication bandwidth, while mechanisms like self-reflection to learn from past mistakes would benefit a lot from long or infinite context windows. Although vector stores and retrieval can provide access to a larger knowledge pool, their representation power is not as powerful as full attention.\\n\\n\\nChallenges in long-term planning and task decomposition: Planning over a lengthy history and effectively exploring the solution space remain challenging. LLMs struggle to adjust plans when faced with unexpected errors, making them less robust compared to humans who learn from trial and error.\\n\\n\\nReliability of natural language interface: Current agent system relies on natural language as an interface between LLMs and external components such as memory and tools. However, the reliability of model outputs is questionable, as LLMs may make formatting errors and occasionally exhibit rebellious behavior (e.g. refuse to follow an instruction). Consequently, much of the agent demo code focuses on parsing model output.\\n\\n\\nCitation#\\nCited as:\\n\\nWeng, Lilian. (Jun 2023). “LLM-powered Autonomous Agents”. Lil’Log. https://lilianweng.github.io/posts/2023-06-23-agent/.\\n\\nOr\\n@article{weng2023agent,\\n  title   = \"LLM-powered Autonomous Agents\",\\n  author  = \"Weng, Lilian\",\\n  journal = \"lilianweng.github.io\",\\n  year    = \"2023\",\\n  month   = \"Jun\",\\n  url     = \"https://lilianweng.github.io/posts/2023-06-23-agent/\"\\n}\\nReferences#\\n[1] Wei et al. “Chain of thought prompting elicits reasoning in large language models.” NeurIPS 2022\\n[2] Yao et al. “Tree of Thoughts: Dliberate Problem Solving with Large Language Models.” arXiv preprint arXiv:2305.10601 (2023).\\n[3] Liu et al. “Chain of Hindsight Aligns Language Models with Feedback\\n“ arXiv preprint arXiv:2302.02676 (2023).\\n[4] Liu et al. “LLM+P: Empowering Large Language Models with Optimal Planning Proficiency” arXiv preprint arXiv:2304.11477 (2023).\\n[5] Yao et al. “ReAct: Synergizing reasoning and acting in language models.” ICLR 2023.\\n[6] Google Blog. “Announcing ScaNN: Efficient Vector Similarity Search” July 28, 2020.\\n[7] https://chat.openai.com/share/46ff149e-a4c7-4dd7-a800-fc4a642ea389\\n[8] Shinn & Labash. “Reflexion: an autonomous agent with dynamic memory and self-reflection” arXiv preprint arXiv:2303.11366 (2023).\\n[9] Laskin et al. “In-context Reinforcement Learning with Algorithm Distillation” ICLR 2023.\\n[10] Karpas et al. “MRKL Systems A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning.” arXiv preprint arXiv:2205.00445 (2022).\\n[11] Nakano et al. “Webgpt: Browser-assisted question-answering with human feedback.” arXiv preprint arXiv:2112.09332 (2021).\\n[12] Parisi et al. “TALM: Tool Augmented Language Models”\\n[13] Schick et al. “Toolformer: Language Models Can Teach Themselves to Use Tools.” arXiv preprint arXiv:2302.04761 (2023).\\n[14] Weaviate Blog. Why is Vector Search so fast? Sep 13, 2022.\\n[15] Li et al. “API-Bank: A Benchmark for Tool-Augmented LLMs” arXiv preprint arXiv:2304.08244 (2023).\\n[16] Shen et al. “HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in HuggingFace” arXiv preprint arXiv:2303.17580 (2023).\\n[17] Bran et al. “ChemCrow: Augmenting large-language models with chemistry tools.” arXiv preprint arXiv:2304.05376 (2023).\\n[18] Boiko et al. “Emergent autonomous scientific research capabilities of large language models.” arXiv preprint arXiv:2304.05332 (2023).\\n[19] Joon Sung Park, et al. “Generative Agents: Interactive Simulacra of Human Behavior.” arXiv preprint arXiv:2304.03442 (2023).\\n[20] AutoGPT. https://github.com/Significant-Gravitas/Auto-GPT\\n[21] GPT-Engineer. https://github.com/AntonOsika/gpt-engineer\\n', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/'})]\n"
     ]
    }
   ],
   "source": [
    "# 数据获取：加载博客内容\n",
    "loader = WebBaseLoader(\n",
    "    web_paths=(\"https://lilianweng.github.io/posts/2023-06-23-agent/\",),\n",
    "    bs_kwargs=dict(\n",
    "        parse_only=bs4.SoupStrainer(\n",
    "            class_=(\"post-content\", \"post-title\", \"post-header\")\n",
    "        )\n",
    "    ),\n",
    ")\n",
    "docs = loader.load()\n",
    "print(docs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 内容拆分：将文档拆分为小块\n",
    "text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)\n",
    "splits = text_splitter.split_documents(docs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 向量embedding和存储：将所有文档拆分嵌入并存储\n",
    "db = FAISS.from_documents(splits, OpenAIEmbeddings())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fig. 1. Overview of a LLM-powered autonomous agent system.\n",
      "Component One: Planning#\n",
      "A complicated task usually involves many steps. An agent needs to know what they are and plan ahead.\n",
      "Task Decomposition#\n",
      "Chain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.\n"
     ]
    }
   ],
   "source": [
    "# 向量相似度检索：使用博客的相关片段检索和生成\n",
    "query = \"What is Task Decomposition?\"\n",
    "docs = db.similarity_search(query)\n",
    "retriever = db.as_retriever()\n",
    "prompt = hub.pull(\"rlm/rag-prompt\")\n",
    "print(docs[0].page_content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Task Decomposition involves breaking down complex tasks into smaller and simpler steps to make them more manageable. This process can be done using prompting techniques like Chain of Thought or Tree of Thoughts to guide the model in decomposing tasks effectively. Task decomposition can be achieved through simple prompting, task-specific instructions, or human inputs.'"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def format_docs(docs):\n",
    "    return \"\\n\\n\".join(doc.page_content for doc in docs)\n",
    "\n",
    "# 实现QA问答：构建RAG链\n",
    "rag_chain = (\n",
    "    {\"context\": retriever | format_docs, \"question\": RunnablePassthrough()}\n",
    "    | prompt\n",
    "    | llm\n",
    "    | StrOutputParser()\n",
    ")\n",
    "\n",
    "# 使用RAG链进行问答\n",
    "rag_chain.invoke(\"What is Task Decomposition?\")\n"
   ]
  }
 ],
 "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": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}