langgraph_agents.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyNXaUf2nzpLN5dzAn5YKGpR",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ramnarayan-code/llm_reflection/blob/main/langgraph_agents.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Install the required packages"
      ],
      "metadata": {
        "id": "Zoildq02yzeM"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 109,
      "metadata": {
        "id": "3RyD2uydP8yR"
      },
      "outputs": [],
      "source": [
        "%%capture --no-stderr\n",
        "%pip install -U langchain langgraph langchain_openai langchain_experimental langsmith pandas pygraphviz\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Ref: https://medium.com/@anuragmishra_27746/future-of-coding-multi-agent-llm-framework-using-langgraph-092da9493663"
      ],
      "metadata": {
        "id": "EPifZZXxtUXY"
      },
      "execution_count": 110,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Import packages and define LLM"
      ],
      "metadata": {
        "id": "_bnYafr-y7WF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "import operator\n",
        "from typing import Annotated, Any, Dict, Optional, Sequence, TypedDict, List, Tuple\n",
        "\n",
        "from langchain_core.pydantic_v1 import BaseModel\n",
        "from langchain_core.prompts import ChatPromptTemplate\n",
        "from langchain_core.pydantic_v1 import BaseModel, Field\n",
        "from langchain_core.messages import HumanMessage\n",
        "from langchain.chains.openai_functions import create_structured_output_runnable\n",
        "from langchain_openai import ChatOpenAI\n",
        "from langgraph.graph import END, StateGraph\n",
        "\n",
        "import os\n",
        "os.environ['OPENAI_API_KEY'] = '<OPENAI-API-KEY>'\n",
        "\n",
        "llm=ChatOpenAI()"
      ],
      "metadata": {
        "id": "ua69NCrbZYGV"
      },
      "execution_count": 112,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Definition of Product Owner Agent\n"
      ],
      "metadata": {
        "id": "8-BESiLrxUgz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class ProductOwner(BaseModel):\n",
        "    user_stories: str = Field(\n",
        "        description=\"PO defined user stories\"\n",
        "    )\n",
        "\n",
        "product_owner_prompt = ChatPromptTemplate.from_template(\n",
        "    '''\n",
        "    **Role**: You are a product owner. You need to create user stories\n",
        "**Task**: As a product owner, you are required to create 5 user stories based on the use case\n",
        "\n",
        "**Instructions**:\n",
        "1. **Understand and Clarify**: Make sure you understand the use case\n",
        "2. **Define user story and assign priority**: define user story in the below format following the agile standards\n",
        "*FORMAT*\n",
        "User story:\n",
        "\n",
        "*USE CASE*\n",
        "{use_case}\n",
        "\n",
        "'''\n",
        ")\n",
        "product_owner_agent = create_structured_output_runnable(\n",
        "    ProductOwner, llm, product_owner_prompt\n",
        ")\n"
      ],
      "metadata": {
        "id": "cGI2UOTBbHfV"
      },
      "execution_count": 113,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Definition of Developer Agent\n"
      ],
      "metadata": {
        "id": "zrT2T5aRxhSY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class Developer(BaseModel):\n",
        "    dev_team_proposal: str = Field(\n",
        "        description=\"Developer proposal for the use case implementation\"\n",
        "    )\n",
        "\n",
        "developer_prompt = ChatPromptTemplate.from_template(\n",
        "    '''\n",
        "    **Role**: You are a developer. You need to assess the user stories\n",
        "**Task**: As a developer, you need to assess the user stories and provide the technical proposal\n",
        "\n",
        "**Instructions**:\n",
        "1. **Understand and Clarify**: Make sure you understand the use case and its related user stories\n",
        "2. **Design the system**: Design a system in the below output format for the use case defining the architecture pattern, defining the components, define the potential technology stack and estimate the efforts\n",
        "*FORMAT*\n",
        "Architecture pattern:\n",
        "Components:\n",
        "Tech Stack:\n",
        "ETA:\n",
        "\n",
        "\n",
        "*USE CASE*\n",
        "{use_case}\n",
        "\n",
        "*USE STORIES*\n",
        "{po_defined_user_stories}\n",
        "'''\n",
        ")\n",
        "developer_agent = create_structured_output_runnable(\n",
        "    Developer, llm, developer_prompt\n",
        ")\n"
      ],
      "metadata": {
        "id": "kTWJfag9rWFb"
      },
      "execution_count": 114,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Langgraph based collaboration between Product owner and Developer agents"
      ],
      "metadata": {
        "id": "mV2hiAbzxtHy"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class AgentState(TypedDict):\n",
        "    use_case: str\n",
        "    po_defined_user_stories: str\n",
        "    dev_team_proposal: str"
      ],
      "metadata": {
        "id": "L-5_Fcd1bMMy"
      },
      "execution_count": 115,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def product_owner(state):\n",
        "    print(f'Product Owner agent:')\n",
        "    use_case = state['use_case']\n",
        "    po_response = product_owner_agent.invoke({'use_case':use_case})\n",
        "    return {'po_defined_user_stories':po_response.user_stories}\n",
        "\n",
        "def developer(state):\n",
        "    print(f'Developer agent:')\n",
        "    po_defined_user_stories = state['po_defined_user_stories']\n",
        "    use_case = state['use_case']\n",
        "    developer_response = developer_agent.invoke({'po_defined_user_stories':po_defined_user_stories, 'use_case':use_case})\n",
        "    return {'dev_team_proposal':developer_response.dev_team_proposal}"
      ],
      "metadata": {
        "id": "RDmx1TTquiFb"
      },
      "execution_count": 116,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "workflow = StateGraph(AgentState)\n",
        "\n",
        "# Define the nodes\n",
        "workflow.add_node(\"product_owner\", product_owner)\n",
        "workflow.add_node(\"developer\", developer)\n",
        "\n",
        "# Build graph\n",
        "workflow.set_entry_point(\"product_owner\")\n",
        "workflow.add_edge(\"product_owner\", \"developer\")\n",
        "workflow.add_edge(\"developer\", END)\n",
        "\n",
        "app = workflow.compile()\n"
      ],
      "metadata": {
        "id": "LaiHgjMNbPgN"
      },
      "execution_count": 117,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Langgraph Structure:"
      ],
      "metadata": {
        "id": "duaIxJuwyFh6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "from langchain_core.runnables.graph import CurveStyle, NodeColors, MermaidDrawMethod\n",
        "from IPython.display import display, HTML, Image\n",
        "\n",
        "display(\n",
        "    Image(\n",
        "        app.get_graph().draw_mermaid_png(\n",
        "            draw_method=MermaidDrawMethod.API,\n",
        "        )\n",
        "    )\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "id": "fmvI0uaewk3G",
        "outputId": "b26525d2-e123-4c7d-8c3e-77f1dac88e54"
      },
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/jpeg": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Interaction with the agents"
      ],
      "metadata": {
        "id": "r6MeZRlGyju9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "for stream_msg in app.stream(\n",
        "    {\"use_case\": [HumanMessage(content=\"Design a CRM system using AI\")]},\n",
        "    {\"recursion_limit\": 100},\n",
        "):\n",
        "    if \"__end__\" not in stream_msg:\n",
        "        if \"product_owner\" in stream_msg:\n",
        "          po_defined_user_stories = stream_msg[\"product_owner\"][\"po_defined_user_stories\"]\n",
        "          print(po_defined_user_stories)\n",
        "        elif \"developer\" in stream_msg:\n",
        "          dev_team_proposal = stream_msg[\"developer\"][\"dev_team_proposal\"]\n",
        "          print(dev_team_proposal)\n",
        "        print(\"----\")"
      ],
      "metadata": {
        "id": "nT9OfbZ9wldD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a526ccd1-5afc-4b49-ecc7-1bb5c6c83f41"
      },
      "execution_count": 122,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Product Owner agent:\n",
            "User story 1: As a sales manager, I want the CRM system to automatically analyze customer data to provide insights for targeted marketing campaigns.\n",
            "User story 2: As a customer service representative, I want the CRM system to suggest personalized responses based on historical interactions with customers.\n",
            "User story 3: As a marketing analyst, I want the CRM system to predict customer behavior using AI algorithms to optimize marketing strategies.\n",
            "User story 4: As a system administrator, I want the CRM system to automate repetitive tasks such as data entry and lead scoring to improve efficiency.\n",
            "User story 5: As a business owner, I want the CRM system to generate AI-powered reports that help in making data-driven decisions for business growth.\n",
            "----\n",
            "Developer agent:\n",
            "**Technical Proposal for CRM System using AI**\n",
            "\n",
            "**Architecture pattern:**\n",
            "- The proposed architecture pattern for the CRM system is a combination of microservices and serverless computing. This architecture allows for scalability, flexibility, and cost-effectiveness.\n",
            "\n",
            "**Components:**\n",
            "1. Customer Data Analysis Service\n",
            "2. Personalized Response Recommendation Service\n",
            "3. Customer Behavior Prediction Service\n",
            "4. Task Automation Service\n",
            "5. AI-Powered Reporting Service\n",
            "\n",
            "**Tech Stack:**\n",
            "- Programming Languages: Python, JavaScript\n",
            "- Frameworks: Flask, Node.js\n",
            "- AI/ML Libraries: TensorFlow, Scikit-learn\n",
            "- Database: MongoDB\n",
            "- Cloud Services: AWS Lambda, Amazon S3\n",
            "\n",
            "**ETA:**\n",
            "- The estimated time for development and implementation of the CRM system using AI is approximately 6-8 months depending on the complexity of AI algorithms and integrations required.\n",
            "\n",
            "----\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "mtoorckdwLRN"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}