t81_559_class_01_4_langchain.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "whjsJasuhstV"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jeffheaton/app_generative_ai/blob/main/t81_559_class_01_4_langchain.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "euOZxlIMhstX"
      },
      "source": [
        "# T81-559: Applications of Generative Artificial Intelligence\n",
        "**Module 1: Course Overview**\n",
        "* Instructor: [Jeff Heaton](https://sites.wustl.edu/jeffheaton/), McKelvey School of Engineering, [Washington University in St. Louis](https://engineering.wustl.edu/Programs/Pages/default.aspx)\n",
        "* For more information visit the [class website](https://sites.wustl.edu/jeffheaton/t81-558/)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d4Yov72PhstY"
      },
      "source": [
        "# Module 1 Material\n",
        "\n",
        "* Part 1.1: Course Overview [[Video]](https://www.youtube.com/watch?v=OVS-6s20Ms0) [[Notebook]](t81_559_class_01_1_overview.ipynb)\n",
        "* Part 1.2: Generative AI Overview [[Video]](https://www.youtube.com/watch?v=ohmPaSsKhMs) [[Notebook]](t81_559_class_01_2_genai.ipynb)\n",
        "* Part 1.3: Introduction to OpenAI [[Video]](https://www.youtube.com/watch?v=C2xyi2Cq-bU) [[Notebook]](t81_559_class_01_3_openai.ipynb)\n",
        "* **Part 1.4: Introduction to LangChain** [[Video]](https://www.youtube.com/watch?v=qQI5AhaKxuI) [[Notebook]](t81_559_class_01_4_langchain.ipynb)\n",
        "* Part 1.5: Prompt Engineering [[Video]](https://www.youtube.com/watch?v=_Uot1i5sIXo) [[Notebook]](t81_559_class_01_5_prompt_engineering.ipynb)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AcAUP0c3hstY"
      },
      "source": [
        "# Google CoLab Instructions\n",
        "\n",
        "The following code ensures that Google CoLab is running and maps Google Drive if needed."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xsI496h5hstZ",
        "outputId": "bbac7309-06e2-4a00-aef7-18efa8abe717"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Note: using Google CoLab\n",
            "Requirement already satisfied: langchain in /usr/local/lib/python3.10/dist-packages (0.3.12)\n",
            "Collecting langchain_openai\n",
            "  Downloading langchain_openai-0.2.14-py3-none-any.whl.metadata (2.7 kB)\n",
            "Requirement already satisfied: PyYAML>=5.3 in /usr/local/lib/python3.10/dist-packages (from langchain) (6.0.2)\n",
            "Requirement already satisfied: SQLAlchemy<3,>=1.4 in /usr/local/lib/python3.10/dist-packages (from langchain) (2.0.36)\n",
            "Requirement already satisfied: aiohttp<4.0.0,>=3.8.3 in /usr/local/lib/python3.10/dist-packages (from langchain) (3.11.10)\n",
            "Requirement already satisfied: async-timeout<5.0.0,>=4.0.0 in /usr/local/lib/python3.10/dist-packages (from langchain) (4.0.3)\n",
            "Requirement already satisfied: langchain-core<0.4.0,>=0.3.25 in /usr/local/lib/python3.10/dist-packages (from langchain) (0.3.25)\n",
            "Requirement already satisfied: langchain-text-splitters<0.4.0,>=0.3.3 in /usr/local/lib/python3.10/dist-packages (from langchain) (0.3.3)\n",
            "Requirement already satisfied: langsmith<0.3,>=0.1.17 in /usr/local/lib/python3.10/dist-packages (from langchain) (0.2.3)\n",
            "Requirement already satisfied: numpy<2,>=1.22.4 in /usr/local/lib/python3.10/dist-packages (from langchain) (1.26.4)\n",
            "Requirement already satisfied: pydantic<3.0.0,>=2.7.4 in /usr/local/lib/python3.10/dist-packages (from langchain) (2.10.3)\n",
            "Requirement already satisfied: requests<3,>=2 in /usr/local/lib/python3.10/dist-packages (from langchain) (2.32.3)\n",
            "Requirement already satisfied: tenacity!=8.4.0,<10,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from langchain) (9.0.0)\n",
            "Collecting langchain-core<0.4.0,>=0.3.25 (from langchain)\n",
            "  Downloading langchain_core-0.3.28-py3-none-any.whl.metadata (6.3 kB)\n",
            "Collecting openai<2.0.0,>=1.58.1 (from langchain_openai)\n",
            "  Downloading openai-1.58.1-py3-none-any.whl.metadata (27 kB)\n",
            "Collecting tiktoken<1,>=0.7 (from langchain_openai)\n",
            "  Downloading tiktoken-0.8.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.6 kB)\n",
            "Requirement already satisfied: aiohappyeyeballs>=2.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (2.4.4)\n",
            "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.3.2)\n",
            "Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (24.3.0)\n",
            "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.5.0)\n",
            "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (6.1.0)\n",
            "Requirement already satisfied: propcache>=0.2.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (0.2.1)\n",
            "Requirement already satisfied: yarl<2.0,>=1.17.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain) (1.18.3)\n",
            "Requirement already satisfied: jsonpatch<2.0,>=1.33 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.4.0,>=0.3.25->langchain) (1.33)\n",
            "Requirement already satisfied: packaging<25,>=23.2 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.4.0,>=0.3.25->langchain) (24.2)\n",
            "Requirement already satisfied: typing-extensions>=4.7 in /usr/local/lib/python3.10/dist-packages (from langchain-core<0.4.0,>=0.3.25->langchain) (4.12.2)\n",
            "Requirement already satisfied: httpx<1,>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from langsmith<0.3,>=0.1.17->langchain) (0.28.1)\n",
            "Requirement already satisfied: orjson<4.0.0,>=3.9.14 in /usr/local/lib/python3.10/dist-packages (from langsmith<0.3,>=0.1.17->langchain) (3.10.12)\n",
            "Requirement already satisfied: requests-toolbelt<2.0.0,>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from langsmith<0.3,>=0.1.17->langchain) (1.0.0)\n",
            "Requirement already satisfied: anyio<5,>=3.5.0 in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.58.1->langchain_openai) (3.7.1)\n",
            "Requirement already satisfied: distro<2,>=1.7.0 in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.58.1->langchain_openai) (1.9.0)\n",
            "Requirement already satisfied: jiter<1,>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.58.1->langchain_openai) (0.8.2)\n",
            "Requirement already satisfied: sniffio in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.58.1->langchain_openai) (1.3.1)\n",
            "Requirement already satisfied: tqdm>4 in /usr/local/lib/python3.10/dist-packages (from openai<2.0.0,>=1.58.1->langchain_openai) (4.67.1)\n",
            "Requirement already satisfied: annotated-types>=0.6.0 in /usr/local/lib/python3.10/dist-packages (from pydantic<3.0.0,>=2.7.4->langchain) (0.7.0)\n",
            "Requirement already satisfied: pydantic-core==2.27.1 in /usr/local/lib/python3.10/dist-packages (from pydantic<3.0.0,>=2.7.4->langchain) (2.27.1)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (3.4.0)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (3.10)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (2.2.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain) (2024.12.14)\n",
            "Requirement already satisfied: greenlet!=0.4.17 in /usr/local/lib/python3.10/dist-packages (from SQLAlchemy<3,>=1.4->langchain) (3.1.1)\n",
            "Requirement already satisfied: regex>=2022.1.18 in /usr/local/lib/python3.10/dist-packages (from tiktoken<1,>=0.7->langchain_openai) (2024.11.6)\n",
            "Requirement already satisfied: exceptiongroup in /usr/local/lib/python3.10/dist-packages (from anyio<5,>=3.5.0->openai<2.0.0,>=1.58.1->langchain_openai) (1.2.2)\n",
            "Requirement already satisfied: httpcore==1.* in /usr/local/lib/python3.10/dist-packages (from httpx<1,>=0.23.0->langsmith<0.3,>=0.1.17->langchain) (1.0.7)\n",
            "Requirement already satisfied: h11<0.15,>=0.13 in /usr/local/lib/python3.10/dist-packages (from httpcore==1.*->httpx<1,>=0.23.0->langsmith<0.3,>=0.1.17->langchain) (0.14.0)\n",
            "Requirement already satisfied: jsonpointer>=1.9 in /usr/local/lib/python3.10/dist-packages (from jsonpatch<2.0,>=1.33->langchain-core<0.4.0,>=0.3.25->langchain) (3.0.0)\n",
            "Downloading langchain_openai-0.2.14-py3-none-any.whl (50 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m50.9/50.9 kB\u001b[0m \u001b[31m2.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading langchain_core-0.3.28-py3-none-any.whl (411 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m411.6/411.6 kB\u001b[0m \u001b[31m21.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading openai-1.58.1-py3-none-any.whl (454 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m454.3/454.3 kB\u001b[0m \u001b[31m21.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading tiktoken-0.8.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.2 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.2/1.2 MB\u001b[0m \u001b[31m37.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: tiktoken, openai, langchain-core, langchain_openai\n",
            "  Attempting uninstall: openai\n",
            "    Found existing installation: openai 1.57.4\n",
            "    Uninstalling openai-1.57.4:\n",
            "      Successfully uninstalled openai-1.57.4\n",
            "  Attempting uninstall: langchain-core\n",
            "    Found existing installation: langchain-core 0.3.25\n",
            "    Uninstalling langchain-core-0.3.25:\n",
            "      Successfully uninstalled langchain-core-0.3.25\n",
            "Successfully installed langchain-core-0.3.28 langchain_openai-0.2.14 openai-1.58.1 tiktoken-0.8.0\n"
          ]
        }
      ],
      "source": [
        "import os\n",
        "\n",
        "try:\n",
        "    from google.colab import drive, userdata\n",
        "    COLAB = True\n",
        "    print(\"Note: using Google CoLab\")\n",
        "except:\n",
        "    print(\"Note: not using Google CoLab\")\n",
        "    COLAB = False\n",
        "\n",
        "# OpenAI Secrets\n",
        "if COLAB:\n",
        "    os.environ[\"OPENAI_API_KEY\"] = userdata.get('OPENAI_API_KEY')\n",
        "\n",
        "# Install needed libraries in CoLab\n",
        "if COLAB:\n",
        "    !pip install langchain langchain_openai"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pC9A-LaYhsta"
      },
      "source": [
        "# Part 1.4: Introduction to LangChain\n",
        "\n",
        "One of the most intriguing and promising developments in the evolving landscape of language models and artificial intelligence is LangChain. This technology represents a significant leap forward in how we interact with and harness the capabilities of large language models (LLMs). As we delve into the intricacies of LangChain in this chapter, it's important to understand not just the technical underpinnings but also the user experience that makes it so revolutionary.\n",
        "\n",
        "## LangChain Chat Conversation Format\n",
        "\n",
        "To explore LangChain comprehensively, we will adopt a format that has become increasingly familiar and effective in LLMs: the chat conversation interface. This interactive style, reminiscent of how many of us communicate daily, offers a unique and accessible means to illustrate LangChain's capabilities, potential applications, and the nuances of its operation.\n",
        "\n",
        "We begin by importing the components from the LangChain library to support a chat-style interface to OpenAI. We will use the ChatOpenAI interface for the OpenAI family of LLM models."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "K2lC_JK3guaj"
      },
      "outputs": [],
      "source": [
        "# Conversation Style Inteface\n",
        "\n",
        "from langchain_core.messages import HumanMessage, SystemMessage, AIMessage\n",
        "from langchain_core.prompts.chat import (\n",
        "    ChatPromptTemplate,\n",
        "    HumanMessagePromptTemplate,\n",
        "    SystemMessagePromptTemplate,\n",
        ")\n",
        "from langchain_openai import ChatOpenAI"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YmWHTAub0sTm"
      },
      "source": [
        "The conversation format consists of arrays of chat entries of the following three types:\n",
        "\n",
        "* **SystemMessage** - This class designates the system prompt that provides instructions to the AI on the nature of the conversation and hints and guidelines. Generally, there will be only one system message at the beginning of the array.\n",
        "* **HumanMessage** - This class designates the chat messages from outside the LLM, typically the human user.\n",
        "* **AIMessage** - This class designates the chat messages from the LLM as responses to the HumanMessage messages.\n",
        "\n",
        "Here we see the chain to ask a simple question."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "V_sJoVYAtE6b"
      },
      "outputs": [],
      "source": [
        "messages = [\n",
        "    SystemMessage(\n",
        "        content=\"You are a helpful assistant that concisely and accurately answers questions.\"\n",
        "    ),\n",
        "    HumanMessage(\n",
        "        content=\"What is the capital of France?\"\n",
        "    ),\n",
        "]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "To7rZtUO4B8t"
      },
      "source": [
        "We now submit these messages and retrieve the output from the model. We will use gpt-4o-mini, which is good enough for this query. Further, we use a zero temperature; we are simply looking for a factual answer, and creativity is not a goal or concern."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xfb875JhtI6J",
        "outputId": "bac5f921-9b6a-4625-c31d-26ede43af116"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model response:\n",
            "The capital of France is Paris.\n",
            "-----------\n",
            "{'token_usage': {'completion_tokens': 8, 'prompt_tokens': 32, 'total_tokens': 40, 'completion_tokens_details': {'accepted_prediction_tokens': 0, 'audio_tokens': 0, 'reasoning_tokens': 0, 'rejected_prediction_tokens': 0}, 'prompt_tokens_details': {'audio_tokens': 0, 'cached_tokens': 0}}, 'model_name': 'gpt-4o-mini-2024-07-18', 'system_fingerprint': 'fp_d02d531b47', 'finish_reason': 'stop', 'logprobs': None}\n"
          ]
        }
      ],
      "source": [
        "MODEL = 'gpt-4o-mini'\n",
        "\n",
        "# Initialize the OpenAI LLM with your API key\n",
        "llm = ChatOpenAI(\n",
        "  model=MODEL,\n",
        "  temperature= 0.0,\n",
        "  n= 1,\n",
        "  max_tokens= 256)\n",
        "\n",
        "print(\"Model response:\")\n",
        "output = llm.invoke(messages)\n",
        "print(output.content)\n",
        "print(\"-----------\")\n",
        "print(output.response_metadata)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "be6a0hXj4lHj"
      },
      "source": [
        "The model that LangChain returns to you returns additional metadata. This data shows the token usage, which might be useful for estimating the total cost expected from this query.\n",
        "\n",
        "We can continue to grow this conversation if we wish. To do so, we added the model's response and another human question. Here, we will ask the model if it was sure about its last response."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2vJuMVHZ1jJ8",
        "outputId": "173e4b5a-7c4a-4ee7-f5af-3aa618e1dd75"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "SystemMessage : You are a helpful assistant that concisely and accurately answers questions.\n",
            "HumanMessage : What is the capital of France?\n",
            "AIMessage : The capital of France is Paris.\n",
            "HumanMessage : Are you sure, I think it was renamed for some reason?\n"
          ]
        }
      ],
      "source": [
        "messages.append(output)\n",
        "messages.append(HumanMessage(content=\"Are you sure, I think it was renamed for some reason?\"))\n",
        "for message in messages:\n",
        "    print(f\"{type(message).__name__} : {message.content}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A8Z7DZRH5TFr"
      },
      "source": [
        "We can submit the conversation array to the model and see its latest response."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9qFxkEmAz1w1",
        "outputId": "0671573e-0680-482c-8635-44ed4b86dd98"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model response:\n",
            "No, Paris has not been renamed. It remains the capital of France.\n"
          ]
        }
      ],
      "source": [
        "print(\"Model response:\")\n",
        "output = llm.invoke(messages)\n",
        "print(output.content)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y9nvX9oA3a0Z"
      },
      "source": [
        "## Asking a Single Question\n",
        "\n",
        "If you wish to ask the model a single question, not as part of a conversation chain, you can pass a string to the model for a response."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        },
        "id": "34-95k0qg8ss",
        "outputId": "66e03c07-702e-4ba5-be5c-f6058f4d5ebc"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\"As of the latest available data, the five largest cities in the USA by population are:\\n\\n1. **New York City, New York**\\n2. **Los Angeles, California**\\n3. **Chicago, Illinois**\\n4. **Houston, Texas**\\n5. **Phoenix, Arizona**\\n\\nPlease note that population figures can change over time, so it's always a good idea to check the most recent census data or estimates for the latest numbers.\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {}
        }
      ],
      "source": [
        "# complete\n",
        "\n",
        "from langchain_openai import OpenAI, ChatOpenAI\n",
        "\n",
        "MODEL = 'gpt-4o-mini'\n",
        "\n",
        "# Initialize the OpenAI LLM (Language Learning Model) with your API key\n",
        "llm = ChatOpenAI(model=MODEL, temperature=0)\n",
        "\n",
        "# Define the question\n",
        "question = \"What are the five largest cities in the USA by population?\"\n",
        "\n",
        "# Use Langchain to call the OpenAI API\n",
        "# The method and parameters might differ based on the Langchain version\n",
        "response = llm.invoke(question)\n",
        "\n",
        "# Print the response\n",
        "display(response.content)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_o2EkUkN3hEX"
      },
      "source": [
        "## Prompt Templates\n",
        "\n",
        "LangChain allows you to create chains of operations typically performed as part of an LLM-enabled application. One of these operations is a prompt template, which allows you to insert text into a previously created prompt. In this example, we will create a prompt template that asks the model to create a random blog post title.\n",
        "\n",
        "```\n",
        "Return only the title of a blog post article title on the topic of {topic} in {language}\n",
        "```\n",
        "\n",
        "To accomplish this objective, we will use a **PromptTemplate** object."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L-g3zOXBmulc",
        "outputId": "0cc8f77c-a590-4951-f653-8faf0efbc19e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\"Data-Driven Pet Care: Using Analytics to Enhance Your Furry Friend's Life\"\n"
          ]
        }
      ],
      "source": [
        "from langchain.prompts import PromptTemplate\n",
        "\n",
        "topic = \"pets for data scientists\"\n",
        "language = \"english\"\n",
        "\n",
        "# Initialize the OpenAI LLM (Language Learning Model) with your API key\n",
        "# Use higher temperature for greater creativity\n",
        "llm = ChatOpenAI(model=MODEL, temperature=0.7)\n",
        "\n",
        "# Define the prompt template\n",
        "title_template = PromptTemplate(\n",
        "    input_variables=['topic', 'language'],\n",
        "    template='Return only the title of a blog post article title on the topic of {topic} in {language}'\n",
        ")\n",
        "\n",
        "# Use RunnableSequence for chaining\n",
        "title_chain = title_template | llm #RunnableSequence(steps=[title_template, llm])\n",
        "\n",
        "# Invoke the chain with inputs\n",
        "response = title_chain.invoke({'topic': topic, 'language': language})\n",
        "print(response.content)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IOpAcj-byBPg"
      },
      "source": [
        "## Create a Simple Sequential Chain\n",
        "\n",
        "We will now use LangChain to tie multiple LLM calls into a longer chain using the **SimpleSequentialChain** class. We will use two smaller chains to create a title and body text for a blog post. We begin by defining the two prompts we will use to construct this blog post. Also, note that we request that the LLM utilize [markdown](https://en.wikipedia.org/wiki/Markdown) to generate the actual blog post.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "ELS-9JD3Sao6"
      },
      "outputs": [],
      "source": [
        "# Create the two prompt templates\n",
        "title_template = PromptTemplate( input_variables = ['topic'], template = 'Give me a blog post title on {topic} in English' )\n",
        "article_template = PromptTemplate( input_variables = ['title'], template = 'Write a blog post for {title}, format in markdown.' )"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k3vy7jf3-_xl"
      },
      "source": [
        "We will create the first chain to generate the random title. Here, we allow the user to specify the topic. We use a higher temperature to increase the creativity of the title. We also use a simpler model to minimize cost for the relatively simple task of title selection."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "_KV8UJUDyP8M"
      },
      "outputs": [],
      "source": [
        "MODEL = 'gpt-4o-mini'\n",
        "\n",
        "# Create a chain to generate a random\n",
        "llm = ChatOpenAI(model=MODEL, temperature=0.7)\n",
        "title_chain = title_template | llm"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cO0ielMf_xRa"
      },
      "source": [
        "Next, we compose the actual blog post; we will use a lower temperature to decrease creativity and cause the LLM to stick to factual information and avoid hallucinations. We also use a more complex model to provide a better article."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "G7OCQCggyHlB"
      },
      "outputs": [],
      "source": [
        "MODEL2 = 'gpt-4'\n",
        "\n",
        "# Create the article chain\n",
        "llm2 = ChatOpenAI(model=MODEL2, temperature=0.1)\n",
        "article_chain = article_template | llm2"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6rftFvMyBa4-"
      },
      "source": [
        "Now, we combine these two chains into one. The input to the first chain will be the selected topic. The first chain will then output the title to the second chain, which will, in turn, output the actual article."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "2ibHBt4eyKZt"
      },
      "outputs": [],
      "source": [
        "# Create a complete chain to create a new blog post\n",
        "#complete_chain=SimpleSequentialChain(chains=[title_chain, article_chain], verbose=True)\n",
        "\n",
        "complete_chain = title_chain | article_chain"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "V_7P-n_DBvpD"
      },
      "source": [
        "We can now display the final article. In this case, we requested an article on \"photography,\" and displayed the final article's markdown."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "mwro_kfKXvml"
      },
      "outputs": [],
      "source": [
        "from IPython.display import display_markdown\n",
        "\n",
        "article = complete_chain.invoke('photography')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h1pN8CbXCE6F"
      },
      "source": [
        "The actual display of the markdown is handled by this code:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 579
        },
        "id": "Vt5uGJawuru_",
        "outputId": "916d6480-54e5-44a5-baea-d78065288984"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/markdown": "# Capturing Moments: The Art and Science of Photography\n\nPhotography is a fascinating blend of art and science, a medium that allows us to capture moments in time, preserving them for future generations. It's a field that has evolved significantly over the years, from the first grainy black-and-white images to the high-resolution digital photos we take today. But at its core, photography remains a way for us to document our world, to tell stories, and to express our unique perspectives.\n\n## The Art of Photography\n\nThe art of photography lies in the ability to see the world in a different light, to find beauty in the mundane, and to capture moments that might otherwise go unnoticed. It's about composition, lighting, and perspective. It's about choosing the right moment to press the shutter, capturing a fleeting expression or a stunning landscape.\n\nA good photographer has an eye for detail, an understanding of color and light, and the ability to tell a story through their images. They know how to use their camera as a tool to express their creativity, to evoke emotion, and to create a connection with the viewer.\n\n## The Science of Photography\n\nOn the other hand, the science of photography is about understanding how cameras work, how light interacts with the camera sensor, and how different settings can affect the final image. It's about understanding the technical aspects of photography, such as aperture, shutter speed, and ISO, and how to use them to achieve the desired effect.\n\nThe science of photography also involves understanding the principles of digital image processing, such as how to edit and enhance photos to bring out their best qualities. It's about using technology to push the boundaries of what's possible in photography, creating images that were unimaginable just a few decades ago.\n\n## The Intersection of Art and Science\n\nThe beauty of photography lies in the intersection of art and science. It's a field where creativity and technical knowledge go hand in hand, each enhancing the other. A photographer who understands the science of photography can use their technical skills to realize their artistic vision, while an artistically inclined photographer can use their creativity to push the boundaries of what's technically possible.\n\nIn the end, photography is about capturing moments, telling stories, and expressing our view of the world. Whether you're a professional photographer or a hobbyist, the art and science of photography offer endless possibilities for exploration and creativity. So grab your camera, start experimenting, and see where your creativity takes you."
          },
          "metadata": {}
        }
      ],
      "source": [
        "display_markdown(article.content, raw=True)"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "anaconda-cloud": {},
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3.11 (genai)",
      "language": "python",
      "name": "genai"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.11.8"
    },
    "varInspector": {
      "cols": {
        "lenName": 16,
        "lenType": 16,
        "lenVar": 40
      },
      "kernels_config": {
        "python": {
          "delete_cmd_postfix": "",
          "delete_cmd_prefix": "del ",
          "library": "var_list.py",
          "varRefreshCmd": "print(var_dic_list())"
        },
        "r": {
          "delete_cmd_postfix": ") ",
          "delete_cmd_prefix": "rm(",
          "library": "var_list.r",
          "varRefreshCmd": "cat(var_dic_list()) "
        }
      },
      "types_to_exclude": [
        "module",
        "function",
        "builtin_function_or_method",
        "instance",
        "_Feature"
      ],
      "window_display": false
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}