LangChain-OpenTutorial · teddylee777 · Jan 18, 2025 · Jan 13, 2025 · Jan 13, 2025 · Jan 13, 2025
diff --git a/15-Agent/05-Iteration-HumanInTheLoop.ipynb b/15-Agent/05-Iteration-HumanInTheLoop.ipynb
@@ -0,0 +1,362 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Iteration-human-in-the-loop\n",
+    "\n",
+    "- Author: [Wonyoung Lee](https://github.com/BaBetterB)\n",
+    "- Peer Review: \n",
+    "- This is a part of [LangChain Open Tutorial](https://github.com/LangChain-OpenTutorial/LangChain-OpenTutorial)\n",
+    "\n",
+    "[![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/BaBetterB/LangChain-OpenTutorial/blob/main/15-Agent/05-Iteration-HumanInTheLoop.ipynb)\n",
+    "[![Open in GitHub](https://img.shields.io/badge/Open%20in%20GitHub-181717?style=flat-square&logo=github&logoColor=white)](https://github.com/LangChain-OpenTutorial/LangChain-OpenTutorial/blob/main/07-TextSplitter/04-SemanticChunker.ipynb)\n",
+    "\n",
+    "\n",
+    "## Overview\n",
+    "\n",
+    "This tutorial covers the functionality of repeating the agent's execution process or receiving user input to decide whether to proceed during intermediate steps. \n",
+    "\n",
+    "The feature of asking the user whether to continue during the agent's execution process is called human-in-the-loop . \n",
+    "\n",
+    "The `iter()` method creates an iterator that allows you to step through the agent's execution process step-by-step.\n",
+    "\n",
+    "\n",
+    "### Table of Contents\n",
+    "\n",
+    "- [Overview](#overview)\n",
+    "- [Environement Setup](#environment-setup)\n",
+    "- [AgentExecutor](#agentexecutor)\n",
+    "\n",
+    "\n",
+    "\n",
+    "### References\n",
+    "\n",
+    "\n",
+    "- [LangChain ChatOpenAI API reference](https://python.langchain.com/api_reference/openai/chat_models/langchain_openai.chat_models.base.ChatOpenAI.html)\n",
+    "- [LangChain AgentExecutor API reference](https://python.langchain.com/api_reference/langchain/agents/langchain.agents.agent.AgentExecutor.html)\n",
+    "- [LangSmith API reference](https://docs.smith.langchain.com/)\n",
+    "\n",
+    "----\n",
+    "\n",
+    " \n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Environment Setup\n",
+    "\n",
+    "Set up the environment. You may refer to [Environment Setup](https://wikidocs.net/257836) for more details.\n",
+    "\n",
+    "**[Note]**\n",
+    "- `langchain-opentutorial` is a package that provides a set of easy-to-use environment setup, useful functions and utilities for tutorials. \n",
+    "- You can checkout the [ `langchain-opentutorial` ](https://github.com/LangChain-OpenTutorial/langchain-opentutorial-pypi) for more details."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Load sample text and output the content."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "[notice] A new release of pip is available: 24.2 -> 24.3.1\n",
+      "[notice] To update, run: python.exe -m pip install --upgrade pip\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%capture --no-stderr\n",
+    "%pip install langchain-opentutorial"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Install required packages\n",
+    "from langchain_opentutorial import package\n",
+    "\n",
+    "\n",
+    "package.install(\n",
+    "    [\n",
+    "        \"langsmith\",\n",
+    "        \"langchain\",\n",
+    "        \"langchain_core\",\n",
+    "        \"langchain_community\",\n",
+    "        \"load_dotenv\",\n",
+    "        \"langchain_openai\",\n",
+    "    ],\n",
+    "    verbose=False,\n",
+    "    upgrade=False,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Environment variables have been set successfully.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Set environment variables\n",
+    "from langchain_opentutorial import set_env\n",
+    "\n",
+    "set_env(\n",
+    "    {\n",
+    "        \"OPENAI_API_KEY\": \"\",\n",
+    "        \"LANGCHAIN_API_KEY\": \"\",\n",
+    "        \"LANGCHAIN_TRACING_V2\": \"true\",\n",
+    "        \"LANGCHAIN_ENDPOINT\": \"https://api.smith.langchain.com\",\n",
+    "        \"LANGCHAIN_PROJECT\": \"Iteration-human-in-the-loop\",  # title\n",
+    "    }\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You can alternatively set `OPENAI_API_KEY` in `.env` file and load it.\n",
+    "\n",
+    "[Note] This is not necessary if you've already set `OPENAI_API_KEY` in previous steps."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 21,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Configuration File for Managing API Keys as Environment Variables\n",
+    "from dotenv import load_dotenv\n",
+    "\n",
+    "# Load API Key Information\n",
+    "load_dotenv(override=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "First, define the tool."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.agents import tool\n",
+    "\n",
+    "\n",
+    "@tool\n",
+    "def add_function(a: float, b: float) -> float:\n",
+    "    \"\"\"Adds two numbers together.\"\"\"\n",
+    "\n",
+    "    return a + b"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Next, define an agent that performs addition calculations using `add_function`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder\n",
+    "from langchain_openai import ChatOpenAI\n",
+    "from langchain.agents import create_tool_calling_agent, AgentExecutor\n",
+    "\n",
+    "# Define tools\n",
+    "tools = [add_function]\n",
+    "\n",
+    "# Create LLM\n",
+    "gpt = ChatOpenAI(model=\"gpt-4o-mini\")\n",
+    "\n",
+    "# Create prompt\n",
+    "prompt = ChatPromptTemplate.from_messages(\n",
+    "    [\n",
+    "        (\n",
+    "            \"system\",\n",
+    "            \"You are a helpful assistant.\"\n",
+    "            \"Please avoid LaTeX-style formatting and use plain symbols.\",\n",
+    "        ),\n",
+    "        (\"human\", \"{input}\"),\n",
+    "        MessagesPlaceholder(variable_name=\"agent_scratchpad\"),\n",
+    "    ]\n",
+    ")\n",
+    "\n",
+    "# Create Agent\n",
+    "gpt_agent = create_tool_calling_agent(gpt, tools, prompt)\n",
+    "\n",
+    "# Create AgentExecutor\n",
+    "agent_executor = AgentExecutor(\n",
+    "    agent=gpt_agent,\n",
+    "    tools=tools,\n",
+    "    verbose=False,\n",
+    "    max_iterations=10,\n",
+    "    handle_parsing_errors=True,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## AgentExecutor\n",
+    "\n",
+    "This method creates an iterator (`AgentExecutorIterator` ) that allows you to step through the agent's execution process.\n",
+    "\n",
+    "**Function Description**\n",
+    "The `iter()` method returns an `AgentExecutorIterator` object that provides sequential access to each step the agent takes until reaching the final output.\n",
+    "\n",
+    "**Key Features**\n",
+    "- **Step-by-step execution access** : Enables you to examine the agent's execution process step-by-step.\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "**Flow Overview**\n",
+    "\n",
+    "To perform the addition calculation for `\"114.5 + 121.2 + 34.2 + 110.1\"`, the steps are executed as follows:\n",
+    "\n",
+    "1. 114.5 + 121.2 = 235.7\n",
+    "2. 235.7 + 34.2 = 269.9\n",
+    "3. 269.9 + 110.1 = 380.0\n",
+    "\n",
+    "You can observe each step in this calculation process.\n",
+    "\n",
+    "During this process, the system displays the intermediate calculation results to the user and asks if they want to continue. (**Human-in-the-loop**)\n",
+    "\n",
+    "If the user inputs anything other than 'y', the iteration stops.\n",
+    "\n",
+    "In practice, while calculating 114.5 + 121.2 = 235.7, 34.2 + 110.1 = 144.3 is also calculated simultaneously.\n",
+    "\n",
+    "Then, the result of 235.7 + 144.3 = 380.0 is calculated as the second step.\n",
+    "\n",
+    "This process can be observed when `verbose=True` is set in the `AgentExecutor`.\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Tool Name: add_function, Execution Result: 235.7\n",
+      "Tool Name: add_function, Execution Result: 380.0\n",
+      "Final result: The result of 114.5 + 121.2 + 34.2 + 110.1 is 380.0.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Define the user input question\n",
+    "question = \"What is the result of 114.5 + 121.2 + 34.2 + 110.1?\"\n",
+    "\n",
+    "\n",
+    "# Flag to track if the calculation is stopped\n",
+    "calculation_stopped = False\n",
+    "\n",
+    "# Use AgentExecutor's iter() method to run step-by-step execution\n",
+    "for step in agent_executor.iter({\"input\": question}):\n",
+    "    # Access each calculation step through intermediate_step\n",
+    "    if output := step.get(\"intermediate_step\"):\n",
+    "        action, value = output[0]\n",
+    "\n",
+    "        # Print the result of each calculation step\n",
+    "        if action.tool == \"add_function\":\n",
+    "            print(f\"Tool Name: {action.tool}, Execution Result: {value}\")\n",
+    "\n",
+    "        # Ask the user whether to continue\n",
+    "        while True:\n",
+    "            _continue = input(\"Do you want to continue? (y/n):\").strip().lower()\n",
+    "            if _continue in [\"y\", \"n\"]:\n",
+    "                if _continue == \"n\":\n",
+    "                    print(f\"Calculation stopped. Last computed result: {value}\")\n",
+    "                    calculation_stopped = True  # Set flag to indicate calculation stop\n",
+    "                    break  # Break from the loop to stop calculation\n",
+    "                break  # Break the inner while loop after valid input\n",
+    "            else:\n",
+    "                print(\"Invalid input! Please enter 'y' or 'n'.\")\n",
+    "\n",
+    "    # Exit the iteration if the calculation is stopped\n",
+    "    if calculation_stopped:\n",
+    "        break\n",
+    "\n",
+    "# Print the final result\n",
+    "if \"output\" in step:\n",
+    "    print(f\"Final result: {step['output']}\")\n",
+    "else:\n",
+    "    print(f\"Final result (from last computation): {value}\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "langchain-opentutorial-HDS-w_h3-py3.11",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}