{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LPPvT9ymC3IT"
      },
      "source": [
        "# Problem Statement"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qGihLYlPigGF"
      },
      "source": [
        "## Business Context"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pcYVQzSbllJN"
      },
      "source": [
        "In modern manufacturing, predictive maintenance plays a critical role in ensuring equipment availability and minimizing unplanned downtime. Unexpected machine failures can cause significant delays and financial losses. A manufacturing company is working to enhance the efficiency and reliability of its predictive maintenance system by leveraging operational and sensor data collected from machine tools. This data includes variables such as air temperature, process temperature, rotational speed, torque, tool wear, and failure indicators.\n",
        "\n",
        "While the company has developed a machine learning model to predict failures, the current process for data handling, model training, and deployment is manual and time-consuming—requiring engineers to rerun notebooks each time new data becomes available. This introduces inefficiencies and delays in responding to potential failures.\n",
        "\n",
        "To address this, there is a clear need for an MLOps pipeline that can seamlessly handle data registration, preprocessing, model training with tuning, and deployment. Such a pipeline would enable real-time integration of new data, reduce manual effort, and support timely, data-driven maintenance decisions—ultimately improving production reliability and reducing operational costs."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LN0KWTL4i2B-"
      },
      "source": [
        "## Objective"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XupSlU6QlmDQ"
      },
      "source": [
        "As an MLOps engineer, you have to design and implement a comprehensive CI/CD pipeline integrated with MLflow for continuous machine learning experimentation tracking. This pipeline will automate critical tasks including data registration, preprocessing, model training with hyperparameter tuning, and deployment to production. The objective is to streamline the entire machine learning workflow, allowing for automated execution of each component whenever new data is ingested, thereby minimizing manual intervention, enhancing operational efficiency, and ensuring robust tracking of experiments and model artifacts through MLflow. This will facilitate better collaboration among team members, faster iterations, and more reliable deployment of machine learning models into production environments."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cuC5FGgnEmSF"
      },
      "source": [
        "## Pre-requisites"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XEuuy1fIEmSF"
      },
      "source": [
        "This project deploys everything on **GitHub Actions** (the ML pipeline) and **Streamlit Community Cloud** (the live app). No Hugging Face and no Codespaces are used.\n",
        "\n",
        "**1. Create a Personal Access Token (PAT)**\n",
        "- GitHub → **Settings** → **Developer settings** → **Personal access tokens** → **Tokens (classic)**\n",
        "- **Generate new token (classic)** and enable these scopes:\n",
        "  - `repo`  (push code + let the pipeline commit the trained model back)\n",
        "  - `workflow`  (push the Actions workflow file)\n",
        "- Copy the token.\n",
        "\n",
        "**2. Store the token in Colab secrets**\n",
        "- In Colab, click the **key icon** (left sidebar) → **Add new secret**\n",
        "- Name: **GH_TOKEN**  (must match `COLAB_SECRET_NAME` in the config cell below)\n",
        "- Value: paste your token, and toggle **Notebook access** on.\n",
        "\n",
        "**3. Create a Streamlit Community Cloud account**\n",
        "- Go to **https://share.streamlit.io** and sign in with the **same GitHub account** so it can access your repo. You'll deploy the app from there at the very end (see the last section).\n",
        "\n",
        "**4. Get an ngrok authtoken (for local MLflow experimentation)**\n",
        "- Go to **https://dashboard.ngrok.com/authtokens**, sign up/sign in, and copy your authtoken.\n",
        "- You'll paste this into the MLflow/ngrok cell later, in the **Experimentation and Tracking (Development Environment)** section."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Installing and Importing Necessary Libraries"
      ],
      "metadata": {
        "id": "hoTSxB0eV3jf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install -q PyGithub==2.9.1\n",
        "!pip install mlflow==3.0.1 pyngrok==7.2.12 -q"
      ],
      "metadata": {
        "id": "sGegY4MmVi_A"
      },
      "execution_count": 112,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Note:**\n",
        "\n",
        "- After running the above cell, kindly restart the notebook kernel (for Jupyter Notebook) or runtime (for Google Colab) and run all cells sequentially from the next cell.\n",
        "\n",
        "- On executing the above line of code, you might see a warning regarding package dependencies. This error message can be ignored as the above code ensures that all necessary libraries and their dependencies are maintained to successfully execute the code in this notebook."
      ],
      "metadata": {
        "id": "ai3Z1xkFV80E"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "from github import Github, GithubException\n",
        "\n",
        "from pyngrok import ngrok\n",
        "import subprocess\n",
        "import mlflow\n",
        "\n",
        "import pandas as pd\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n",
        "from sklearn.compose import make_column_transformer\n",
        "from sklearn.pipeline import make_pipeline\n",
        "import xgboost as xgb\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.metrics import classification_report\n",
        "import joblib"
      ],
      "metadata": {
        "id": "U54adAhfVsWz"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NoWqvBRDEmSG"
      },
      "source": [
        "## Configuration"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pOQH9MPAEmSG"
      },
      "execution_count": null,
      "outputs": [],
      "source": [
        "# Edit these three values, then run every cell top to bottom.\n",
        "GITHUB_USERNAME   = \"Swetha1929\"                 # Your GitHub username\n",
        "REPO_NAME         = \"https://github.com/Swetha1929/predictive-maintenance-mlops\"  # Repository name\n",
        "COLAB_SECRET_NAME = \"GITHUB_TOKEN\"                # Name of the secret in Colab\n",
        "\n",
        "REPO   = f\"{GITHUB_USERNAME}/{REPO_NAME}\"\n",
        "BRANCH = \"main\""
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Secrets in Colab"
      ],
      "metadata": {
        "id": "Xv4xyPs-YzOh"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Secrets** in Google Colab provide a secure way to store sensitive information such as API keys, access tokens, and passwords. Instead of hardcoding these values in your notebook, you can save them as secrets and access them securely whenever needed."
      ],
      "metadata": {
        "id": "ZTn9QfuPFtMP"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "1. In the left pane of Google Colab, click the fifth icon, which looks like a key."
      ],
      "metadata": {
        "id": "UcXbMr9-Y2b4"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "O4nzux__Yuns"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "2. Then, click **+ Add new secret**. In the **Name** field, enter a name to identify the GitHub Personal Access Token, for example, **Gittoken**. In the **Value** field, paste the token. Once done, close the **Secrets** window. We will access this token programmatically later."
      ],
      "metadata": {
        "id": "_GbLKduqY9-c"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "RkJ6WU2sZJ-S"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "alv74F-cEmSH",
        "outputId": "fd4567ca-47d4-4cd4-90cf-cee3c6e7b1bf"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Token loaded.\n"
          ]
        }
      ],
      "source": [
        "COLAB_SECRET_NAME = \"GITHUB_TOKEN\"\n",
        "\n",
        "import os\n",
        "from google.colab import userdata\n",
        "\n",
        "os.environ[\"GH_TOKEN\"] = userdata.get(COLAB_SECRET_NAME)\n",
        "print(\"Token loaded.\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0LbSu_p2jYfe"
      },
      "source": [
        "# Model Building"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9DtS3gNDjBbR"
      },
      "source": [
        "## Data Registration"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "id": "9R55s4cdrqYW"
      },
      "outputs": [],
      "source": [
        "os.makedirs(\"week_3_mls/data\", exist_ok=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WxXiD9ZXxodF"
      },
      "source": [
        "Once the **data** folder created after executing the above cell, please upload the **machine-failure-prediction.csv** in to the folder"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 114,
      "metadata": {
        "id": "oAyK_beOrqYW"
      },
      "outputs": [],
      "source": [
        "# Create a folder for storing the model building files\n",
        "os.makedirs(\"/content/week_3_mls/model_building\", exist_ok=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OzqvKufQrqYW"
      },
      "source": [
        "1. **Imports Libraries**: Brings in `pandas` to read and validate the dataset.\n",
        "2. **Loads the Raw Dataset**: Reads the CSV that was uploaded into `week_3_mls/data/`.\n",
        "3. **Validates Columns**: Checks that all the expected columns are present before treating the dataset as \"registered\".\n",
        "4. **Reports a Summary**: Prints the row/column counts and the class balance of the target column.\n",
        "\n",
        "This keeps \"registration\" lightweight and file-based: the CSV lives in the GitHub repo itself, so there's no external dataset store to manage."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 149,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BDU0DQGErqYW",
        "outputId": "6be8224d-d088-479f-e5d1-d615cd94dcba"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Dataset registered successfully.\n",
            "Rows: 10000\n",
            "Columns: 14\n",
            "Columns: ['UDI', 'Product ID', 'Type', 'Air temperature [K]', 'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]', 'Tool wear [min]', 'Machine failure', 'TWF', 'HDF', 'PWF', 'OSF', 'RNF']\n",
            "\n",
            "Target distribution:\n",
            "Machine failure\n",
            "0    9661\n",
            "1     339\n",
            "Name: count, dtype: int64\n",
            "\n",
            "Dataset saved at: /content/week_3_mls/data/machine-failure-prediction.csv\n"
          ]
        }
      ],
      "source": [
        "from pathlib import Path\n",
        "import pandas as pd\n",
        "\n",
        "# -----------------------------\n",
        "# Paths\n",
        "# -----------------------------\n",
        "# BASE_DIR = Path(__file__).resolve().parent\n",
        "# In Colab, __file__ is not defined. We use an explicit path based on the\n",
        "# directory where this script (data_register.py) would be written.\n",
        "BASE_DIR = Path(\"/content/week_3_mls/model_building\")\n",
        "DATA_PATH = BASE_DIR.parent / \"data\" / \"machine-failure-prediction.csv\"\n",
        "\n",
        "# -----------------------------\n",
        "# Load dataset\n",
        "# -----------------------------\n",
        "if not DATA_PATH.exists():\n",
        "    raise FileNotFoundError(f\"Dataset not found at: {DATA_PATH}\")\n",
        "\n",
        "df = pd.read_csv(DATA_PATH)\n",
        "df.columns = df.columns.str.strip()\n",
        "\n",
        "# -----------------------------\n",
        "# Validate columns\n",
        "# -----------------------------\n",
        "expected_columns = [\n",
        "    \"UDI\", \"Product ID\", \"Type\", \"Air temperature [K]\",\n",
        "    \"Process temperature [K]\", \"Rotational speed [rpm]\",\n",
        "    \"Torque [Nm]\", \"Tool wear [min]\", \"Machine failure\",\n",
        "    \"TWF\", \"HDF\", \"PWF\", \"OSF\", \"RNF\"\n",
        "]\n",
        "\n",
        "missing = [col for col in expected_columns if col not in df.columns]\n",
        "if missing:\n",
        "    raise ValueError(f\"Dataset is missing expected columns: {missing}\")\n",
        "\n",
        "# -----------------------------\n",
        "# Basic info\n",
        "# -----------------------------\n",
        "print(\"Dataset registered successfully.\")\n",
        "print(f\"Rows: {df.shape[0]}\")\n",
        "print(f\"Columns: {df.shape[1]}\")\n",
        "print(\"Columns:\", df.columns.tolist())\n",
        "print(\"\\nTarget distribution:\")\n",
        "print(df[\"Machine failure\"].value_counts())\n",
        "\n",
        "# -----------------------------\n",
        "# Save cleaned dataset back\n",
        "# -----------------------------\n",
        "DATA_PATH.parent.mkdir(parents=True, exist_ok=True)\n",
        "df.to_csv(DATA_PATH, index=False)\n",
        "\n",
        "print(f\"\\nDataset saved at: {DATA_PATH}\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.columns"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1Gu5l832tlnw",
        "outputId": "72f9c3e4-74a5-444c-d75d-15c02f4684aa"
      },
      "execution_count": 116,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['UDI', 'Product ID', 'Type', 'Air temperature [K]',\n",
              "       'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]',\n",
              "       'Tool wear [min]', 'Machine failure', 'TWF', 'HDF', 'PWF', 'OSF',\n",
              "       'RNF'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 116
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "df = pd.read_csv('/content/week_3_mls/data/machine-failure-prediction.csv')\n",
        "df.columns = df.columns.str.strip() # Strip whitespace from column names\n",
        "print(\"DataFrame Columns after stripping whitespace:\", df.columns.tolist())\n",
        "\n",
        "# Validate that the expected columns are present before registering it\n",
        "expected_columns = [\n",
        "    \"UDI\", \"Type\", \"Air temperature [K]\", \"Process temperature [K]\",\n",
        "    \"Rotational speed [rpm]\", \"Torque [Nm]\", \"Tool wear [min]\", \"Machine failure\"\n",
        "]\n",
        "missing = [c for c in expected_columns if c not in df.columns]\n",
        "if missing:\n",
        "    raise ValueError(f\"Dataset is missing expected columns: {missing}\")\n",
        "\n",
        "print(\"Dataset registered successfully.\")\n",
        "print(f\"Rows: {df.shape[0]}, Columns: {df.shape[1]}\")\n",
        "print(\"Columns:\", list(df.columns))\n",
        "print(\"Failure distribution:\")\n",
        "print(df[\"Machine failure\"].value_counts())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yQabtyHnstre",
        "outputId": "a49ff5e4-159a-4152-f371-b151c9b62606"
      },
      "execution_count": 117,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "DataFrame Columns after stripping whitespace: ['UDI', 'Product ID', 'Type', 'Air temperature [K]', 'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]', 'Tool wear [min]', 'Machine failure', 'TWF', 'HDF', 'PWF', 'OSF', 'RNF']\n",
            "Dataset registered successfully.\n",
            "Rows: 10000, Columns: 14\n",
            "Columns: ['UDI', 'Product ID', 'Type', 'Air temperature [K]', 'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]', 'Tool wear [min]', 'Machine failure', 'TWF', 'HDF', 'PWF', 'OSF', 'RNF']\n",
            "Failure distribution:\n",
            "Machine failure\n",
            "0    9661\n",
            "1     339\n",
            "Name: count, dtype: int64\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hh2TjRG5WJ4Z"
      },
      "source": [
        "## Data Preparation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kHWdLAZCrqYW"
      },
      "source": [
        "1. **Imports Necessary Libraries**: Brings in `pandas` and `train_test_split`.\n",
        "2. **Dataset Loading**: Reads the registered CSV from `week_3_mls/data/`.\n",
        "3. **Data Preparation**: Drops the `UDI` identifier column and splits features/target into train and test sets (stratified on `Failure` to preserve the class imbalance).\n",
        "4. **Saving Prepared Data**: Writes `Xtrain.csv`, `Xtest.csv`, `ytrain.csv`, and `ytest.csv` to disk, which the GitHub Actions workflow passes to the next job as an artifact.\n",
        "\n",
        "**Note:** `Type` is kept as raw H/L/M strings here. It's one-hot-encoded later, inside the model pipeline, not label-encoded, so training and the deployed app stay consistent."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 120,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6mwk_9rsrqYW",
        "outputId": "cfa47120-8574-4bc2-8215-2b34613013bc"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Data preparation completed successfully.\n",
            "Saved files: Xtrain.csv, Xtest.csv, ytrain.csv, ytest.csv\n"
          ]
        }
      ],
      "source": [
        "from pathlib import Path\n",
        "import pandas as pd\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "# Replace Path(__file__).resolve().parent with the absolute path\n",
        "# where this script would conceptually reside if it were a file.\n",
        "# In the context of writing to 'week_3_mls/model_building/prep.py',\n",
        "# its base directory is '/content/week_3_mls/model_building'.\n",
        "BASE_DIR = Path(\"/content/week_3_mls/model_building\")\n",
        "DATA_PATH = BASE_DIR.parent / \"data\" / \"machine-failure-prediction.csv\"\n",
        "\n",
        "if not DATA_PATH.exists():\n",
        "    raise FileNotFoundError(f\"Dataset not found at: {DATA_PATH}\")\n",
        "\n",
        "df = pd.read_csv(DATA_PATH)\n",
        "df.columns = df.columns.str.strip()\n",
        "\n",
        "TARGET_COL = \"Machine failure\"\n",
        "if TARGET_COL not in df.columns:\n",
        "    raise ValueError(\n",
        "        f\"Target column '{TARGET_COL}' not found. Available columns: {df.columns.tolist()}\"\n",
        "    )\n",
        "\n",
        "X = df.drop(columns=[TARGET_COL])\n",
        "y = df[TARGET_COL]\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(\n",
        "    X, y, test_size=0.2, random_state=42, stratify=y\n",
        ")\n",
        "\n",
        "X_train.to_csv(\"Xtrain.csv\", index=False)\n",
        "X_test.to_csv(\"Xtest.csv\", index=False)\n",
        "y_train.to_csv(\"ytrain.csv\", index=False)\n",
        "y_test.to_csv(\"ytest.csv\", index=False)\n",
        "\n",
        "print(\"Data preparation completed successfully.\")\n",
        "print(\"Saved files: Xtrain.csv, Xtest.csv, ytrain.csv, ytest.csv\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "c53UXntcsGqD",
        "outputId": "84c89c6a-a357-4392-ac41-91e04f88b3bf"
      },
      "execution_count": 121,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        UDI Product ID Type  Air temperature [K]  Process temperature [K]  \\\n",
              "0         1     M14860    M                298.1                    308.6   \n",
              "1         2     L47181    L                298.2                    308.7   \n",
              "2         3     L47182    L                298.1                    308.5   \n",
              "3         4     L47183    L                298.2                    308.6   \n",
              "4         5     L47184    L                298.2                    308.7   \n",
              "...     ...        ...  ...                  ...                      ...   \n",
              "9995   9996     M24855    M                298.8                    308.4   \n",
              "9996   9997     H39410    H                298.9                    308.4   \n",
              "9997   9998     M24857    M                299.0                    308.6   \n",
              "9998   9999     H39412    H                299.0                    308.7   \n",
              "9999  10000     M24859    M                299.0                    308.7   \n",
              "\n",
              "      Rotational speed [rpm]  Torque [Nm]  Tool wear [min]  Machine failure  \\\n",
              "0                       1551         42.8                0                0   \n",
              "1                       1408         46.3                3                0   \n",
              "2                       1498         49.4                5                0   \n",
              "3                       1433         39.5                7                0   \n",
              "4                       1408         40.0                9                0   \n",
              "...                      ...          ...              ...              ...   \n",
              "9995                    1604         29.5               14                0   \n",
              "9996                    1632         31.8               17                0   \n",
              "9997                    1645         33.4               22                0   \n",
              "9998                    1408         48.5               25                0   \n",
              "9999                    1500         40.2               30                0   \n",
              "\n",
              "      TWF  HDF  PWF  OSF  RNF  \n",
              "0       0    0    0    0    0  \n",
              "1       0    0    0    0    0  \n",
              "2       0    0    0    0    0  \n",
              "3       0    0    0    0    0  \n",
              "4       0    0    0    0    0  \n",
              "...   ...  ...  ...  ...  ...  \n",
              "9995    0    0    0    0    0  \n",
              "9996    0    0    0    0    0  \n",
              "9997    0    0    0    0    0  \n",
              "9998    0    0    0    0    0  \n",
              "9999    0    0    0    0    0  \n",
              "\n",
              "[10000 rows x 14 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-16ef2210-ff5c-473a-90d5-360889bf2cbd\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>UDI</th>\n",
              "      <th>Product ID</th>\n",
              "      <th>Type</th>\n",
              "      <th>Air temperature [K]</th>\n",
              "      <th>Process temperature [K]</th>\n",
              "      <th>Rotational speed [rpm]</th>\n",
              "      <th>Torque [Nm]</th>\n",
              "      <th>Tool wear [min]</th>\n",
              "      <th>Machine failure</th>\n",
              "      <th>TWF</th>\n",
              "      <th>HDF</th>\n",
              "      <th>PWF</th>\n",
              "      <th>OSF</th>\n",
              "      <th>RNF</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>M14860</td>\n",
              "      <td>M</td>\n",
              "      <td>298.1</td>\n",
              "      <td>308.6</td>\n",
              "      <td>1551</td>\n",
              "      <td>42.8</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>L47181</td>\n",
              "      <td>L</td>\n",
              "      <td>298.2</td>\n",
              "      <td>308.7</td>\n",
              "      <td>1408</td>\n",
              "      <td>46.3</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>L47182</td>\n",
              "      <td>L</td>\n",
              "      <td>298.1</td>\n",
              "      <td>308.5</td>\n",
              "      <td>1498</td>\n",
              "      <td>49.4</td>\n",
              "      <td>5</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>L47183</td>\n",
              "      <td>L</td>\n",
              "      <td>298.2</td>\n",
              "      <td>308.6</td>\n",
              "      <td>1433</td>\n",
              "      <td>39.5</td>\n",
              "      <td>7</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>L47184</td>\n",
              "      <td>L</td>\n",
              "      <td>298.2</td>\n",
              "      <td>308.7</td>\n",
              "      <td>1408</td>\n",
              "      <td>40.0</td>\n",
              "      <td>9</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9995</th>\n",
              "      <td>9996</td>\n",
              "      <td>M24855</td>\n",
              "      <td>M</td>\n",
              "      <td>298.8</td>\n",
              "      <td>308.4</td>\n",
              "      <td>1604</td>\n",
              "      <td>29.5</td>\n",
              "      <td>14</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9996</th>\n",
              "      <td>9997</td>\n",
              "      <td>H39410</td>\n",
              "      <td>H</td>\n",
              "      <td>298.9</td>\n",
              "      <td>308.4</td>\n",
              "      <td>1632</td>\n",
              "      <td>31.8</td>\n",
              "      <td>17</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9997</th>\n",
              "      <td>9998</td>\n",
              "      <td>M24857</td>\n",
              "      <td>M</td>\n",
              "      <td>299.0</td>\n",
              "      <td>308.6</td>\n",
              "      <td>1645</td>\n",
              "      <td>33.4</td>\n",
              "      <td>22</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9998</th>\n",
              "      <td>9999</td>\n",
              "      <td>H39412</td>\n",
              "      <td>H</td>\n",
              "      <td>299.0</td>\n",
              "      <td>308.7</td>\n",
              "      <td>1408</td>\n",
              "      <td>48.5</td>\n",
              "      <td>25</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9999</th>\n",
              "      <td>10000</td>\n",
              "      <td>M24859</td>\n",
              "      <td>M</td>\n",
              "      <td>299.0</td>\n",
              "      <td>308.7</td>\n",
              "      <td>1500</td>\n",
              "      <td>40.2</td>\n",
              "      <td>30</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10000 rows × 14 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-16ef2210-ff5c-473a-90d5-360889bf2cbd')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-16ef2210-ff5c-473a-90d5-360889bf2cbd button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-16ef2210-ff5c-473a-90d5-360889bf2cbd');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "  <div id=\"id_65f928d9-b74e-47f9-bea7-b5ac0a4e41c3\">\n",
              "    <style>\n",
              "      .colab-df-generate {\n",
              "        background-color: #E8F0FE;\n",
              "        border: none;\n",
              "        border-radius: 50%;\n",
              "        cursor: pointer;\n",
              "        display: none;\n",
              "        fill: #1967D2;\n",
              "        height: 32px;\n",
              "        padding: 0 0 0 0;\n",
              "        width: 32px;\n",
              "      }\n",
              "\n",
              "      .colab-df-generate:hover {\n",
              "        background-color: #E2EBFA;\n",
              "        box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "        fill: #174EA6;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate {\n",
              "        background-color: #3B4455;\n",
              "        fill: #D2E3FC;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate:hover {\n",
              "        background-color: #434B5C;\n",
              "        box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "        filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "        fill: #FFFFFF;\n",
              "      }\n",
              "    </style>\n",
              "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('df')\"\n",
              "            title=\"Generate code using this dataframe.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "    <script>\n",
              "      (() => {\n",
              "      const buttonEl =\n",
              "        document.querySelector('#id_65f928d9-b74e-47f9-bea7-b5ac0a4e41c3 button.colab-df-generate');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      buttonEl.onclick = () => {\n",
              "        google.colab.notebook.generateWithVariable('df');\n",
              "      }\n",
              "      })();\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df",
              "summary": "{\n  \"name\": \"df\",\n  \"rows\": 10000,\n  \"fields\": [\n    {\n      \"column\": \"UDI\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 2886,\n        \"min\": 1,\n        \"max\": 10000,\n        \"num_unique_values\": 10000,\n        \"samples\": [\n          6253,\n          4685,\n          1732\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Product ID\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 10000,\n        \"samples\": [\n          \"L53432\",\n          \"M19544\",\n          \"M16591\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Type\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 3,\n        \"samples\": [\n          \"M\",\n          \"L\",\n          \"H\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Air temperature [K]\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 2.0002586829157574,\n        \"min\": 295.3,\n        \"max\": 304.5,\n        \"num_unique_values\": 93,\n        \"samples\": [\n          299.3,\n          296.9,\n          300.8\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Process temperature [K]\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.4837342191657419,\n        \"min\": 305.7,\n        \"max\": 313.8,\n        \"num_unique_values\": 82,\n        \"samples\": [\n          307.2,\n          308.6,\n          310.1\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Rotational speed [rpm]\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 179,\n        \"min\": 1168,\n        \"max\": 2886,\n        \"num_unique_values\": 941,\n        \"samples\": [\n          1274,\n          1576,\n          2010\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Torque [Nm]\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 9.968933725121401,\n        \"min\": 3.8,\n        \"max\": 76.6,\n        \"num_unique_values\": 577,\n        \"samples\": [\n          36.1,\n          65.9,\n          12.6\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Tool wear [min]\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 63,\n        \"min\": 0,\n        \"max\": 253,\n        \"num_unique_values\": 246,\n        \"samples\": [\n          93,\n          14,\n          215\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Machine failure\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"TWF\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"HDF\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"PWF\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"OSF\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"RNF\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 121
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eZZKnLkLjeM4"
      },
      "source": [
        "## Model Training"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G7nMvH4vtnLQ"
      },
      "source": [
        "\n",
        "1. **Set Ngrok Authentication**: Authenticates Ngrok using a personal token to enable secure tunneling from local to public network.\n",
        "\n",
        "2. **Launch MLflow UI**: Starts the MLflow Tracking UI as a background process on local port 5000 for experiment visualization and tracking.\n",
        "\n",
        "3. **Create Public Tunnel**: Uses Ngrok to expose the local MLflow UI to the internet, generating a public URL that can be accessed remotely.\n",
        "\n",
        "4. **Display Public URL**: Prints the Ngrok-generated URL, allowing users to open and interact with the MLflow UI in their browser."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PzE083BHJhx1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "76bd0420-f1da-478a-8abf-ec25620cf56a"
      },
      "execution_count": 129,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "2026/07/31 18:12:05 INFO mlflow.tracking.fluent: Experiment with name 'mlops-training-experiment' does not exist. Creating a new experiment.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MLflow started.\n"
          ]
        }
      ],
      "source": [
        "import subprocess\n",
        "import time\n",
        "import mlflow\n",
        "\n",
        "process = subprocess.Popen(\n",
        "    [\"mlflow\", \"ui\", \"--host\", \"0.0.0.0\", \"--port\", \"5000\"],\n",
        "    stdout=subprocess.DEVNULL,\n",
        "    stderr=subprocess.DEVNULL\n",
        ")\n",
        "\n",
        "time.sleep(5)\n",
        "\n",
        "mlflow.set_tracking_uri(\"http://127.0.0.1:5000\")\n",
        "mlflow.set_experiment(\"mlops-training-experiment\")\n",
        "\n",
        "print(\"MLflow started.\")"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3EnwkJHmK7Ax",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "da70d44d-3a33-4abc-8f3b-79098ff50f08"
      },
      "execution_count": 131,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Experiment: artifact_location='mlflow-artifacts:/876009104830794530', creation_time=1785521525692, experiment_id='876009104830794530', last_update_time=1785521525692, lifecycle_stage='active', name='mlops-training-experiment', tags={}>"
            ]
          },
          "metadata": {},
          "execution_count": 131
        }
      ],
      "source": [
        "# Set the tracking URL for MLflow\n",
        "mlflow.set_experiment(\"mlops-training-experiment\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HnqCGXq6rqYX"
      },
      "source": [
        "1. **Imports Necessary Libraries**: Essential libraries for data manipulation, model training, evaluation, and experiment tracking are brought in.\n",
        "2. **Data Loading**: The registered dataset is read directly from `week_3_mls/data/`.\n",
        "3. **Feature Definition**: Numerical and categorical features are defined, detailing the characteristics of each.\n",
        "4. **Class Weight Calculation**: Class weights are calculated to address the imbalance in the target variable, improving model training effectiveness.\n",
        "5. **Preprocessing Steps**: Preprocessing is set up using a column transformer that scales numerical features and one-hot-encodes the categorical `Type` feature.\n",
        "6. **Model Definition**: An XGBoost classifier is defined with the calculated class weight.\n",
        "7. **Experimentation Tracking**: MLflow logs each parameter combination tested during grid search as a nested run, so every trial can be compared side by side in the MLflow UI.\n",
        "8. **Prediction and Evaluation**: Predictions are made on both training and test datasets (using a custom classification threshold), and classification reports are generated to evaluate performance.\n",
        "9. **Model Storage**: The best model is saved locally with joblib, ready to be committed by the pipeline and served by the Streamlit app."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 156,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9xPJaYf2rqYX",
        "outputId": "633a580c-eff4-4230-a8a3-2353227fa8c1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "🏃 View run trial_1 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/8f7ec3472cb04bf3b786e3e192f15e21\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 1/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_2 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/b29e1762b65c431b88079bf4789505a9\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 2/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.6}\n",
            "🏃 View run trial_3 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/a782359bff7c4a878c81cee6085caf0d\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 3/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.8}\n",
            "🏃 View run trial_4 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/31721225db944c4a9532136ff6b24448\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 4/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 1.0}\n",
            "🏃 View run trial_5 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/05cc699023984ce0b923f62a7ff874b1\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 5/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_6 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/59fc767f8f9e4a4aa31e1e598ec04662\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 6/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.6}\n",
            "🏃 View run trial_7 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/cd4e25d2733e4ff586b5b6014b26418e\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 7/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.8}\n",
            "🏃 View run trial_8 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/67534d7a1b454830bd4c089b1fd7296d\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 8/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 1.0}\n",
            "🏃 View run trial_9 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/50b19e299800413ea5e58b59895ee120\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 9/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_10 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/09d0263ae3484c71889097d3723a49e9\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 10/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 0.6}\n",
            "🏃 View run trial_11 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/0e6c54b1898e414db334f0ea68d29b49\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 11/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 0.8}\n",
            "🏃 View run trial_12 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/3c7e22d6972f4545b87c30ed2cdbdf15\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 12/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 1.0}\n",
            "🏃 View run trial_13 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/17897031174f4a048740311882f96788\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 13/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_14 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/e57bbd7fa1ca494f91210bf13b4a0cf6\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 14/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.6}\n",
            "🏃 View run trial_15 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/6a82970d588140fdbc84fa88322fb7bd\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 15/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.8}\n",
            "🏃 View run trial_16 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/49b266e1fb794a2a8e7f6b9e9bc7f6c3\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 16/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 1.0}\n",
            "🏃 View run trial_17 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/d19a1eec8dd4420f9430199efdd13275\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 17/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_18 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/02eb3f3bb00342b0ad967be461bb52d3\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 18/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.6}\n",
            "🏃 View run trial_19 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/7d38a0477dc0469bb7958a4ce6cfbad7\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 19/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.8}\n",
            "🏃 View run trial_20 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/75fd4344e2514b5497ad8c4119544f6c\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 20/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 1.0}\n",
            "🏃 View run trial_21 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/d040eaf3b3cc45d88ed3f7c50fd0d295\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 21/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_22 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/f107e9e31c5f446eb25ede79341b1525\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 22/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 0.6}\n",
            "🏃 View run trial_23 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/124f530077b94b80bc20edd45c87b55c\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 23/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 0.8}\n",
            "🏃 View run trial_24 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/784191cfbaa4423cb381de1c58f792f8\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 24/30 - Recall: 0.9559 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 3, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 1.0}\n",
            "🏃 View run trial_25 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/446143408db34b3abf8c0f2c8c3c5a04\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 25/30 - Recall: 0.9265 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 4, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_26 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/d2c6bb81e2444f14a5b1537efe97b8c2\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 26/30 - Recall: 0.9265 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 4, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.6}\n",
            "🏃 View run trial_27 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/e8e79ac00159420eae097c1fcbf61f44\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 27/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 4, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 0.8}\n",
            "🏃 View run trial_28 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/9c6fa3aa5a73432dbe3257ac799fc64e\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 28/30 - Recall: 0.9412 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 4, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__reg_lambda': 1.0}\n",
            "🏃 View run trial_29 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/41c77c4f9c064ee09a01597b1bd542e8\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 29/30 - Recall: 0.9118 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 4, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.4}\n",
            "🏃 View run trial_30 at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/1e831da91df14dcd90bbc134b09d1e0f\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n",
            "Trial 30/30 - Recall: 0.9118 - Params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 4, 'xgbclassifier__n_estimators': 75, 'xgbclassifier__reg_lambda': 0.6}\n",
            "\n",
            "Best params: {'xgbclassifier__learning_rate': 0.01, 'xgbclassifier__max_depth': 2, 'xgbclassifier__n_estimators': 100, 'xgbclassifier__reg_lambda': 0.4}\n",
            "Best test recall: 0.9558823529411765\n",
            "\n",
            "Final Test Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      0.91      0.95      1932\n",
            "           1       0.27      0.91      0.41        68\n",
            "\n",
            "    accuracy                           0.91      2000\n",
            "   macro avg       0.63      0.91      0.68      2000\n",
            "weighted avg       0.97      0.91      0.93      2000\n",
            "\n",
            "\n",
            "Model saved to week_3_mls/deployment/best_machine_failure_model_v1.joblib\n",
            "🏃 View run xgb_grid_search at: http://127.0.0.1:5000/#/experiments/876009104830794530/runs/56bfe72036a74017a0d684b29dce2b38\n",
            "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/876009104830794530\n"
          ]
        }
      ],
      "source": [
        "import pandas as pd\n",
        "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n",
        "from sklearn.compose import make_column_transformer\n",
        "from sklearn.pipeline import make_pipeline\n",
        "from sklearn.model_selection import GridSearchCV, ParameterGrid\n",
        "from sklearn.metrics import classification_report\n",
        "import xgboost as xgb\n",
        "import joblib\n",
        "import mlflow\n",
        "\n",
        "# If you are running in GitHub Actions with a local MLflow server,\n",
        "# uncomment the next line:\n",
        "# mlflow.set_tracking_uri(\"http://localhost:5000\")\n",
        "\n",
        "mlflow.set_experiment(\"mlops-training-experiment\")\n",
        "\n",
        "# -----------------------------\n",
        "# Load train/test splits\n",
        "# -----------------------------\n",
        "Xtrain = pd.read_csv(\"Xtrain.csv\")\n",
        "Xtest = pd.read_csv(\"Xtest.csv\")\n",
        "ytrain = pd.read_csv(\"ytrain.csv\").squeeze()\n",
        "ytest = pd.read_csv(\"ytest.csv\").squeeze()\n",
        "\n",
        "Xtrain.columns = Xtrain.columns.str.strip()\n",
        "Xtest.columns = Xtest.columns.str.strip()\n",
        "\n",
        "# -----------------------------\n",
        "# Features\n",
        "# -----------------------------\n",
        "numeric_features = [\n",
        "    \"Air temperature [K]\",\n",
        "    \"Process temperature [K]\",\n",
        "    \"Rotational speed [rpm]\",\n",
        "    \"Torque [Nm]\",\n",
        "    \"Tool wear [min]\"\n",
        "]\n",
        "categorical_features = [\"Type\"]\n",
        "\n",
        "# -----------------------------\n",
        "# Handle class imbalance\n",
        "# -----------------------------\n",
        "class_weight = ytrain.value_counts()[0] / ytrain.value_counts()[1]\n",
        "\n",
        "# -----------------------------\n",
        "# Preprocessing\n",
        "# -----------------------------\n",
        "preprocessor = make_column_transformer(\n",
        "    (StandardScaler(), numeric_features),\n",
        "    (OneHotEncoder(handle_unknown=\"ignore\"), categorical_features)\n",
        ")\n",
        "\n",
        "# -----------------------------\n",
        "# Base model\n",
        "# -----------------------------\n",
        "xgb_model = xgb.XGBClassifier(\n",
        "    scale_pos_weight=class_weight,\n",
        "    random_state=42,\n",
        "    eval_metric=\"logloss\",\n",
        "    tree_method=\"hist\"\n",
        ")\n",
        "\n",
        "# -----------------------------\n",
        "# Hyperparameter grid\n",
        "# Create 30 combinations\n",
        "# -----------------------------\n",
        "full_grid = {\n",
        "    \"xgbclassifier__n_estimators\": [50, 75, 100],\n",
        "    \"xgbclassifier__max_depth\": [2, 3, 4],\n",
        "    \"xgbclassifier__learning_rate\": [0.01, 0.05, 0.1],\n",
        "    \"xgbclassifier__reg_lambda\": [0.4, 0.6, 0.8, 1.0]\n",
        "}\n",
        "\n",
        "param_list = list(ParameterGrid(full_grid))[:30]\n",
        "\n",
        "# -----------------------------\n",
        "# Pipeline\n",
        "# -----------------------------\n",
        "model_pipeline = make_pipeline(preprocessor, xgb_model)\n",
        "\n",
        "# -----------------------------\n",
        "# Main MLflow run\n",
        "# -----------------------------\n",
        "with mlflow.start_run(run_name=\"xgb_grid_search\"):\n",
        "\n",
        "    best_score = -1\n",
        "    best_params = None\n",
        "    best_model = None\n",
        "    best_threshold = 0.45\n",
        "\n",
        "    for i, params in enumerate(param_list):\n",
        "        # Update model params\n",
        "        model_pipeline.set_params(**params)\n",
        "\n",
        "        # Fit model\n",
        "        model_pipeline.fit(Xtrain, ytrain)\n",
        "\n",
        "        # Predict on validation/test set\n",
        "        y_pred_test_proba = model_pipeline.predict_proba(Xtest)[:, 1]\n",
        "        y_pred_test = (y_pred_test_proba >= best_threshold).astype(int)\n",
        "\n",
        "        # Metrics\n",
        "        test_report = classification_report(ytest, y_pred_test, output_dict=True)\n",
        "        recall_score = test_report[\"1\"][\"recall\"]\n",
        "\n",
        "        # Log each param set as a nested MLflow run\n",
        "        with mlflow.start_run(run_name=f\"trial_{i+1}\", nested=True):\n",
        "            mlflow.log_params(params)\n",
        "            mlflow.log_metric(\"test_accuracy\", test_report[\"accuracy\"])\n",
        "            mlflow.log_metric(\"test_precision\", test_report[\"1\"][\"precision\"])\n",
        "            mlflow.log_metric(\"test_recall\", test_report[\"1\"][\"recall\"])\n",
        "            mlflow.log_metric(\"test_f1-score\", test_report[\"1\"][\"f1-score\"])\n",
        "\n",
        "        print(f\"Trial {i+1}/30 - Recall: {recall_score:.4f} - Params: {params}\")\n",
        "\n",
        "        # Keep best model\n",
        "        if recall_score > best_score:\n",
        "            best_score = recall_score\n",
        "            best_params = params\n",
        "            best_model = model_pipeline\n",
        "\n",
        "    # -------------------------\n",
        "    # Log best params\n",
        "    # -------------------------\n",
        "    mlflow.log_params(best_params)\n",
        "    mlflow.log_metric(\"best_test_recall\", best_score)\n",
        "\n",
        "    print(\"\\nBest params:\", best_params)\n",
        "    print(\"Best test recall:\", best_score)\n",
        "\n",
        "    # -------------------------\n",
        "    # Final evaluation\n",
        "    # -------------------------\n",
        "    threshold = 0.45\n",
        "\n",
        "    y_pred_train_proba = best_model.predict_proba(Xtrain)[:, 1]\n",
        "    y_pred_train = (y_pred_train_proba >= threshold).astype(int)\n",
        "\n",
        "    y_pred_test_proba = best_model.predict_proba(Xtest)[:, 1]\n",
        "    y_pred_test = (y_pred_test_proba >= threshold).astype(int)\n",
        "\n",
        "    train_report = classification_report(ytrain, y_pred_train, output_dict=True)\n",
        "    test_report = classification_report(ytest, y_pred_test, output_dict=True)\n",
        "\n",
        "    print(\"\\nFinal Test Classification Report:\")\n",
        "    print(classification_report(ytest, y_pred_test))\n",
        "\n",
        "    mlflow.log_metrics({\n",
        "        \"train_accuracy\": train_report[\"accuracy\"],\n",
        "        \"train_precision\": train_report[\"1\"][\"precision\"],\n",
        "        \"train_recall\": train_report[\"1\"][\"recall\"],\n",
        "        \"train_f1-score\": train_report[\"1\"][\"f1-score\"],\n",
        "        \"test_accuracy\": test_report[\"accuracy\"],\n",
        "        \"test_precision\": test_report[\"1\"][\"precision\"],\n",
        "        \"test_recall\": test_report[\"1\"][\"recall\"],\n",
        "        \"test_f1-score\": test_report[\"1\"][\"f1-score\"]\n",
        "    })\n",
        "\n",
        "    # -------------------------\n",
        "    # Save model\n",
        "    # -------------------------\n",
        "    model_path = \"week_3_mls/deployment/best_machine_failure_model_v1.joblib\"\n",
        "    joblib.dump(best_model, model_path)\n",
        "    mlflow.log_artifact(model_path, artifact_path=\"model\")\n",
        "\n",
        "    print(f\"\\nModel saved to {model_path}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ABqaxuW5TvRW"
      },
      "source": [
        "- As we can see, all the experiments conducted during hyperparameter tuning are being logged by MLflow.\n",
        "- Upon clicking the links in the output of the above code cell, we can check the tracking on MLflow."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LCH6DuM-T_DA"
      },
      "source": [
        "Now that we've tested the experimentation tracking with MLflow in a development environment, let's convert this to the required script for production environment usage."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oV__RG6dOgYW"
      },
      "source": [
        "### Experimentation and Tracking (Production Environment)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 146,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "owM3g9lVrqYZ",
        "outputId": "47324741-c75a-44ac-994c-79b32f2fa523"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overwriting week_3_mls/model_building/train.py\n"
          ]
        }
      ],
      "source": [
        "%%writefile week_3_mls/model_building/train.py\n",
        "import pandas as pd\n",
        "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n",
        "from sklearn.compose import make_column_transformer\n",
        "from sklearn.pipeline import make_pipeline\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.metrics import classification_report\n",
        "import xgboost as xgb\n",
        "import joblib\n",
        "import mlflow\n",
        "\n",
        "mlflow.set_experiment(\"mlops-training-experiment\")\n",
        "\n",
        "Xtrain = pd.read_csv(\"Xtrain.csv\")\n",
        "Xtest = pd.read_csv(\"Xtest.csv\")\n",
        "ytrain = pd.read_csv(\"ytrain.csv\").squeeze()\n",
        "ytest = pd.read_csv(\"ytest.csv\").squeeze()\n",
        "\n",
        "Xtrain.columns = Xtrain.columns.str.strip()\n",
        "Xtest.columns = Xtest.columns.str.strip()\n",
        "\n",
        "numeric_features = [\n",
        "    \"Air temperature [K]\",\n",
        "    \"Process temperature [K]\",\n",
        "    \"Rotational speed [rpm]\",\n",
        "    \"Torque [Nm]\",\n",
        "    \"Tool wear [min]\"\n",
        "]\n",
        "categorical_features = [\"Type\"]\n",
        "\n",
        "class_weight = ytrain.value_counts()[0] / ytrain.value_counts()[1]\n",
        "\n",
        "preprocessor = make_column_transformer(\n",
        "    (StandardScaler(), numeric_features),\n",
        "    (OneHotEncoder(handle_unknown=\"ignore\"), categorical_features)\n",
        ")\n",
        "\n",
        "xgb_model = xgb.XGBClassifier(\n",
        "    scale_pos_weight=class_weight,\n",
        "    random_state=42,\n",
        "    eval_metric=\"logloss\",\n",
        "    tree_method=\"hist\"\n",
        ")\n",
        "\n",
        "param_grid = {\n",
        "    \"xgbclassifier__n_estimators\": [50, 100],\n",
        "    \"xgbclassifier__max_depth\": [2, 3],\n",
        "    \"xgbclassifier__learning_rate\": [0.05, 0.1],\n",
        "}\n",
        "\n",
        "model_pipeline = make_pipeline(preprocessor, xgb_model)\n",
        "\n",
        "with mlflow.start_run():\n",
        "    grid_search = GridSearchCV(\n",
        "        model_pipeline,\n",
        "        param_grid,\n",
        "        cv=3,\n",
        "        scoring=\"recall\",\n",
        "        n_jobs=-1\n",
        "    )\n",
        "    grid_search.fit(Xtrain, ytrain)\n",
        "\n",
        "    mlflow.log_params(grid_search.best_params_)\n",
        "\n",
        "    best_model = grid_search.best_estimator_\n",
        "    print(\"Best params:\", grid_search.best_params_)\n",
        "\n",
        "    threshold = 0.45\n",
        "    y_pred_train = (best_model.predict_proba(Xtrain)[:, 1] >= threshold).astype(int)\n",
        "    y_pred_test = (best_model.predict_proba(Xtest)[:, 1] >= threshold).astype(int)\n",
        "\n",
        "    train_report = classification_report(ytrain, y_pred_train, output_dict=True)\n",
        "    test_report = classification_report(ytest, y_pred_test, output_dict=True)\n",
        "\n",
        "    print(classification_report(ytest, y_pred_test))\n",
        "\n",
        "    mlflow.log_metrics({\n",
        "        \"train_accuracy\": train_report[\"accuracy\"],\n",
        "        \"train_precision\": train_report[\"1\"][\"precision\"],\n",
        "        \"train_recall\": train_report[\"1\"][\"recall\"],\n",
        "        \"train_f1-score\": train_report[\"1\"][\"f1-score\"],\n",
        "        \"test_accuracy\": test_report[\"accuracy\"],\n",
        "        \"test_precision\": test_report[\"1\"][\"precision\"],\n",
        "        \"test_recall\": test_report[\"1\"][\"recall\"],\n",
        "        \"test_f1-score\": test_report[\"1\"][\"f1-score\"]\n",
        "    })\n",
        "\n",
        "    model_path = \"week_3_mls/deployment/best_machine_failure_model_v1.joblib\"\n",
        "    joblib.dump(best_model, model_path)\n",
        "    mlflow.log_artifact(model_path, artifact_path=\"model\")\n",
        "    print(f\"Model saved to {model_path}\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%%writefile week_3_mls/model_building/prep.py\n",
        "\n",
        "from pathlib import Path\n",
        "import pandas as pd\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "# -----------------------------\n",
        "# Paths\n",
        "# -----------------------------\n",
        "BASE_DIR = Path(__file__).resolve().parent\n",
        "DATA_PATH = BASE_DIR.parent / \"data\" / \"machine-failure-prediction.csv\"\n",
        "\n",
        "# -----------------------------\n",
        "# Load dataset\n",
        "# -----------------------------\n",
        "if not DATA_PATH.exists():\n",
        "    raise FileNotFoundError(f\"Dataset not found at: {DATA_PATH}\")\n",
        "\n",
        "df = pd.read_csv(DATA_PATH)\n",
        "df.columns = df.columns.str.strip()\n",
        "\n",
        "print(\"Dataset loaded successfully.\")\n",
        "print(\"Shape:\", df.shape)\n",
        "\n",
        "# -----------------------------\n",
        "# Drop unnecessary columns\n",
        "# -----------------------------\n",
        "drop_cols = [\"UDI\", \"Product ID\", \"TWF\", \"HDF\", \"PWF\", \"OSF\", \"RNF\"]\n",
        "\n",
        "existing_cols = [c for c in drop_cols if c in df.columns]\n",
        "df = df.drop(columns=existing_cols)\n",
        "\n",
        "# -----------------------------\n",
        "# Target\n",
        "# -----------------------------\n",
        "TARGET_COL = \"Machine failure\"\n",
        "\n",
        "if TARGET_COL not in df.columns:\n",
        "    raise ValueError(\n",
        "        f\"Target column '{TARGET_COL}' not found. \"\n",
        "        f\"Available columns: {df.columns.tolist()}\"\n",
        "    )\n",
        "\n",
        "X = df.drop(columns=[TARGET_COL])\n",
        "y = df[TARGET_COL]\n",
        "\n",
        "# -----------------------------\n",
        "# Train/Test Split\n",
        "# -----------------------------\n",
        "Xtrain, Xtest, ytrain, ytest = train_test_split(\n",
        "    X,\n",
        "    y,\n",
        "    test_size=0.20,\n",
        "    random_state=42,\n",
        "    stratify=y\n",
        ")\n",
        "\n",
        "# -----------------------------\n",
        "# Save splits\n",
        "# -----------------------------\n",
        "Xtrain.to_csv(\"Xtrain.csv\", index=False)\n",
        "Xtest.to_csv(\"Xtest.csv\", index=False)\n",
        "ytrain.to_csv(\"ytrain.csv\", index=False)\n",
        "ytest.to_csv(\"ytest.csv\", index=False)\n",
        "\n",
        "print(\"Data preparation completed successfully.\")\n",
        "print(f\"Training samples : {len(Xtrain)}\")\n",
        "print(f\"Testing samples  : {len(Xtest)}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZLLI5Ql7goZ1",
        "outputId": "54ab693f-9371-4f90-a758-e13edc6bc0d6"
      },
      "execution_count": 151,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overwriting week_3_mls/model_building/prep.py\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0McYCZzkji5I"
      },
      "source": [
        "# Deployment"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4roQt1bFrqYa"
      },
      "source": [
        "## Streamlit App"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 134,
      "metadata": {
        "id": "sSXNru68rqYa"
      },
      "outputs": [],
      "source": [
        "os.makedirs(\"week_3_mls/deployment\", exist_ok=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7RR6I3ePrqYb"
      },
      "source": [
        "This application provides an interactive tool for machine failure prediction, predicting whether a machine is likely to fail based on its operational parameters. It loads the model that the GitHub Actions pipeline trained and committed into the repo, collects user inputs, makes predictions, and displays the results in a simple interface. Streamlit Community Cloud runs this file directly from the repo, so there's nothing to configure beyond pointing it at `week_3_mls/deployment/app.py` when you deploy at the end."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 135,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gLOCkgotrqYb",
        "outputId": "9af8f813-8515-4b43-d847-a2f0a1fa1886"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overwriting week_3_mls/deployment/app.py\n"
          ]
        }
      ],
      "source": [
        "%%writefile week_3_mls/deployment/app.py\n",
        "\n",
        "import os\n",
        "import streamlit as st\n",
        "import pandas as pd\n",
        "import joblib\n",
        "\n",
        "# -----------------------------\n",
        "# Load trained model\n",
        "# -----------------------------\n",
        "model_path = os.path.join(\n",
        "    os.path.dirname(__file__),\n",
        "    \"best_machine_failure_model_v1.joblib\"\n",
        ")\n",
        "\n",
        "model = joblib.load(model_path)\n",
        "\n",
        "# -----------------------------\n",
        "# Streamlit page configuration\n",
        "# -----------------------------\n",
        "st.set_page_config(\n",
        "    page_title=\"Machine Failure Prediction\",\n",
        "    page_icon=\"⚙️\",\n",
        "    layout=\"centered\"\n",
        ")\n",
        "\n",
        "st.title(\"⚙️ Machine Failure Prediction\")\n",
        "st.write(\n",
        "    \"\"\"\n",
        "    Enter the machine operating parameters below to predict\n",
        "    whether the machine is likely to fail.\n",
        "    \"\"\"\n",
        ")\n",
        "\n",
        "# -----------------------------\n",
        "# User Inputs\n",
        "# -----------------------------\n",
        "machine_type = st.selectbox(\n",
        "    \"Machine Type\",\n",
        "    [\"H\", \"L\", \"M\"]\n",
        ")\n",
        "\n",
        "air_temp = st.number_input(\n",
        "    \"Air Temperature (K)\",\n",
        "    min_value=250.0,\n",
        "    max_value=400.0,\n",
        "    value=298.0,\n",
        "    step=0.1\n",
        ")\n",
        "\n",
        "process_temp = st.number_input(\n",
        "    \"Process Temperature (K)\",\n",
        "    min_value=250.0,\n",
        "    max_value=500.0,\n",
        "    value=308.0,\n",
        "    step=0.1\n",
        ")\n",
        "\n",
        "rot_speed = st.number_input(\n",
        "    \"Rotational Speed (rpm)\",\n",
        "    min_value=0,\n",
        "    max_value=3000,\n",
        "    value=1500,\n",
        "    step=1\n",
        ")\n",
        "\n",
        "torque = st.number_input(\n",
        "    \"Torque (Nm)\",\n",
        "    min_value=0.0,\n",
        "    max_value=100.0,\n",
        "    value=40.0,\n",
        "    step=0.1\n",
        ")\n",
        "\n",
        "tool_wear = st.number_input(\n",
        "    \"Tool Wear (min)\",\n",
        "    min_value=0,\n",
        "    max_value=300,\n",
        "    value=10,\n",
        "    step=1\n",
        ")\n",
        "\n",
        "# -----------------------------\n",
        "# Prediction\n",
        "# -----------------------------\n",
        "if st.button(\"Predict Machine Failure\"):\n",
        "\n",
        "    input_df = pd.DataFrame({\n",
        "        \"Type\": [machine_type],\n",
        "        \"Air temperature [K]\": [air_temp],\n",
        "        \"Process temperature [K]\": [process_temp],\n",
        "        \"Rotational speed [rpm]\": [rot_speed],\n",
        "        \"Torque [Nm]\": [torque],\n",
        "        \"Tool wear [min]\": [tool_wear]\n",
        "    })\n",
        "\n",
        "    prediction = model.predict(input_df)[0]\n",
        "\n",
        "    # Only calculate probability if the model supports it\n",
        "    probability = None\n",
        "    if hasattr(model, \"predict_proba\"):\n",
        "        probability = model.predict_proba(input_df)[0][1]\n",
        "\n",
        "    st.subheader(\"Prediction Result\")\n",
        "\n",
        "    if prediction == 1:\n",
        "        st.error(\"⚠ Machine Failure Predicted\")\n",
        "    else:\n",
        "        st.success(\"✅ No Machine Failure Predicted\")\n",
        "\n",
        "    if probability is not None:\n",
        "        st.write(f\"**Failure Probability:** {probability:.2%}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gw5Uyox_rqYb"
      },
      "source": [
        "## App Dependencies"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 137,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cUz77AcOrqYb",
        "outputId": "4d34a7bb-3359-4151-d99e-add6dc4bed50"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overwriting week_3_mls/deployment/requirements.txt\n"
          ]
        }
      ],
      "source": [
        "%%writefile week_3_mls/deployment/requirements.txt\n",
        "streamlit==1.43.2\n",
        "pandas==2.2.2\n",
        "numpy==2.2.3\n",
        "scikit-learn==1.6.0\n",
        "xgboost==2.1.4\n",
        "joblib==1.5.1"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PuCgAW2hktli"
      },
      "source": [
        "# Create and Automate MLOps Pipeline with GitHub Action Workflows using CI/CD"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yo_xWLLvkqSC"
      },
      "source": [
        "## Action Workflow YAML File"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ICEDG3zwkLh2"
      },
      "source": [
        "* A YAML file is a simple, human-readable file used to store configuration settings.\n",
        "* YAML stands for Yet Another Markup Language or YAML Ain't Markup Language (a recursive acronym).\n",
        "* It uses indentation (spaces) to show structure, like folders inside folders.\n",
        "* Each line contains a key and a value, making it easy to organize data.\n",
        "* YAML is often used in automation tools, cloud setups, and app settings."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_oH3zSIIrqYb"
      },
      "source": [
        "Three jobs run in sequence — **Register Dataset → Data Preparation → Model Training (with MLflow tracking)** — then the trained model is committed back into the repo so Streamlit Community Cloud can serve it.\n",
        "\n",
        "**Important:** GitHub only picks up workflow files from `.github/workflows/` at the repo **root** — not nested inside `week_3_mls/`. The cell below creates that folder at the top level, alongside `week_3_mls/`."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 138,
      "metadata": {
        "id": "qZiLae0QrqYb"
      },
      "outputs": [],
      "source": [
        "os.makedirs(\".github/workflows\", exist_ok=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 140,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JDX-bDMmrqYc",
        "outputId": "f576bb35-18e4-4757-c53d-c4a80350d685"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overwriting .github/workflows/pipeline.yml\n"
          ]
        }
      ],
      "source": [
        "%%writefile .github/workflows/pipeline.yml\n",
        "name: Week 3 MLS CI/CD Pipeline\n",
        "\n",
        "on:\n",
        "  push:\n",
        "    branches:\n",
        "      - main\n",
        "  workflow_dispatch:\n",
        "\n",
        "permissions:\n",
        "  contents: write\n",
        "\n",
        "jobs:\n",
        "  register-dataset:\n",
        "    name: Register Dataset\n",
        "    runs-on: ubuntu-latest\n",
        "\n",
        "    steps:\n",
        "      - name: Checkout Repository\n",
        "        uses: actions/checkout@v4\n",
        "\n",
        "      - name: Setup Python\n",
        "        uses: actions/setup-python@v5\n",
        "        with:\n",
        "          python-version: \"3.11\"\n",
        "\n",
        "      - name: Install Dependencies\n",
        "        run: |\n",
        "          python -m pip install --upgrade pip\n",
        "          pip install -r week_3_mls/requirements.txt\n",
        "\n",
        "      - name: Register Dataset\n",
        "        run: python week_3_mls/model_building/data_register.py\n",
        "\n",
        "      - name: Upload Registered Dataset\n",
        "        uses: actions/upload-artifact@v4\n",
        "        with:\n",
        "          name: registered-data\n",
        "          path: week_3_mls/data/machine-failure-prediction.csv\n",
        "\n",
        "  data-preparation:\n",
        "    name: Data Preparation\n",
        "    needs: register-dataset\n",
        "    runs-on: ubuntu-latest\n",
        "\n",
        "    steps:\n",
        "      - name: Checkout Repository\n",
        "        uses: actions/checkout@v4\n",
        "\n",
        "      - name: Setup Python\n",
        "        uses: actions/setup-python@v5\n",
        "        with:\n",
        "          python-version: \"3.11\"\n",
        "\n",
        "      - name: Install Dependencies\n",
        "        run: |\n",
        "          python -m pip install --upgrade pip\n",
        "          pip install -r week_3_mls/requirements.txt\n",
        "\n",
        "      - name: Prepare Data\n",
        "        run: python week_3_mls/model_building/prep.py\n",
        "\n",
        "      - name: Upload Train/Test Splits\n",
        "        uses: actions/upload-artifact@v4\n",
        "        with:\n",
        "          name: data-splits\n",
        "          path: |\n",
        "            Xtrain.csv\n",
        "            Xtest.csv\n",
        "            ytrain.csv\n",
        "            ytest.csv\n",
        "\n",
        "  model-training:\n",
        "    name: Model Training\n",
        "    needs: data-preparation\n",
        "    runs-on: ubuntu-latest\n",
        "\n",
        "    steps:\n",
        "      - name: Checkout Repository\n",
        "        uses: actions/checkout@v4\n",
        "\n",
        "      - name: Setup Python\n",
        "        uses: actions/setup-python@v5\n",
        "        with:\n",
        "          python-version: \"3.11\"\n",
        "\n",
        "      - name: Install Dependencies\n",
        "        run: |\n",
        "          python -m pip install --upgrade pip\n",
        "          pip install -r week_3_mls/requirements.txt\n",
        "\n",
        "      - name: Start MLflow Tracking Server\n",
        "        run: |\n",
        "          nohup mlflow ui --host 0.0.0.0 --port 5000 &\n",
        "          sleep 5\n",
        "\n",
        "      - name: Download Train/Test Splits\n",
        "        uses: actions/download-artifact@v4\n",
        "        with:\n",
        "          name: data-splits\n",
        "\n",
        "      - name: Train Model\n",
        "        run: python week_3_mls/model_building/train.py\n",
        "\n",
        "      - name: Commit Trained Model\n",
        "        run: |\n",
        "          git config user.name \"github-actions[bot]\"\n",
        "          git config user.email \"github-actions[bot]@users.noreply.github.com\"\n",
        "\n",
        "          git add week_3_mls/deployment/best_machine_failure_model_v1.joblib\n",
        "\n",
        "          git diff --cached --quiet || git commit -m \"Add trained model [skip ci]\"\n",
        "          git push"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4sMZOH8jrqYc"
      },
      "source": [
        "**Note:** The cells above already created `.github/workflows/pipeline.yml` locally. There's nothing to copy-paste manually — the push step later in this notebook will upload it to your repo along with the rest of the project."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PvEUJ-t5kdxH"
      },
      "source": [
        "## Requirements file for the Github Actions Workflow"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 141,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bEr0Lly9rqYc",
        "outputId": "3fb98058-257b-48b1-dcb4-a8a91abade33"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overwriting week_3_mls/requirements.txt\n"
          ]
        }
      ],
      "source": [
        "%%writefile week_3_mls/requirements.txt\n",
        "pandas==2.2.2\n",
        "numpy==2.2.3\n",
        "scikit-learn==1.6.0\n",
        "xgboost==2.1.4\n",
        "joblib==1.5.1\n",
        "mlflow==3.0.1\n",
        "streamlit==1.43.2\n",
        "PyGithub==2.3.0"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BA6mP-Ebkm3O"
      },
      "source": [
        "## Github Authentication and Push Files"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "T1cH6gpOrqYc"
      },
      "source": [
        "We already created a GitHub PAT and stored it in Colab secrets during the **Prerequisites** step above (as `GITHUB_TOKEN`), and loaded it into `GH_TOKEN` in the **Configuration** section. The cell below uses that token (via `PyGithub`) to create the repo if needed and push every file — no manual `git clone`/`git push`, and no token hardcoded in this notebook."
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "os.makedirs(\"week_3_mls/model_building\", exist_ok=True)\n",
        "os.makedirs(\"week_3_mls/data\", exist_ok=True)"
      ],
      "metadata": {
        "id": "kuDxUMgO8xP7"
      },
      "execution_count": 142,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4cc3a423"
      },
      "source": [
        "# Edit these three values, then run every cell top to bottom.\n",
        "GITHUB_USERNAME   = \"Swetha1929\"                 # Your GitHub username\n",
        "REPO_NAME         = \"predictive-maintenance-mlops\"  # Repository name\n",
        "COLAB_SECRET_NAME = \"GITHUB_TOKEN\"                # Name of the secret in Colab\n",
        "\n",
        "REPO   = f\"{GITHUB_USERNAME}/{REPO_NAME}\"\n",
        "BRANCH = \"main\""
      ],
      "execution_count": 143,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "import os\n",
        "\n",
        "# Dataset path\n",
        "CSV_PATH = \"/content/week_3_mls/data/machine-failure-prediction.csv\"\n",
        "\n",
        "# Ensure data directory exists\n",
        "os.makedirs(\"/content/week_3_mls/data\", exist_ok=True)\n",
        "\n",
        "# Check dataset exists\n",
        "if not os.path.exists(CSV_PATH):\n",
        "    raise FileNotFoundError(\n",
        "        f\"Dataset not found at {CSV_PATH}. \"\n",
        "        \"Please upload machine-failure-prediction.csv into week_3_mls/data.\"\n",
        "    )\n",
        "\n",
        "# Load dataset\n",
        "df = pd.read_csv(CSV_PATH)\n",
        "\n",
        "# Remove any extra spaces from column names\n",
        "df.columns = df.columns.str.strip()\n",
        "\n",
        "print(\"Dataset loaded successfully.\")\n",
        "print(f\"Path: {CSV_PATH}\")\n",
        "print(f\"Shape: {df.shape}\")\n",
        "\n",
        "print(\"\\nColumns:\")\n",
        "print(df.columns.tolist())\n",
        "\n",
        "print(\"\\nFirst 5 rows:\")\n",
        "print(df.head())\n",
        "\n",
        "print(\"\\nMissing Values:\")\n",
        "print(df.isnull().sum())\n",
        "\n",
        "print(\"\\nTarget Distribution:\")\n",
        "print(df[\"Machine failure\"].value_counts())\n",
        "\n",
        "# Save the validated dataset\n",
        "df.to_csv(CSV_PATH, index=False)\n",
        "\n",
        "print(\"\\nDataset registered successfully.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tJcezWODOV1N",
        "outputId": "556a1c8b-bb62-4071-eb60-b43c728689a1"
      },
      "execution_count": 144,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Dataset loaded successfully.\n",
            "Path: /content/week_3_mls/data/machine-failure-prediction.csv\n",
            "Shape: (10000, 14)\n",
            "\n",
            "Columns:\n",
            "['UDI', 'Product ID', 'Type', 'Air temperature [K]', 'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]', 'Tool wear [min]', 'Machine failure', 'TWF', 'HDF', 'PWF', 'OSF', 'RNF']\n",
            "\n",
            "First 5 rows:\n",
            "   UDI Product ID Type  Air temperature [K]  Process temperature [K]  \\\n",
            "0    1     M14860    M                298.1                    308.6   \n",
            "1    2     L47181    L                298.2                    308.7   \n",
            "2    3     L47182    L                298.1                    308.5   \n",
            "3    4     L47183    L                298.2                    308.6   \n",
            "4    5     L47184    L                298.2                    308.7   \n",
            "\n",
            "   Rotational speed [rpm]  Torque [Nm]  Tool wear [min]  Machine failure  TWF  \\\n",
            "0                    1551         42.8                0                0    0   \n",
            "1                    1408         46.3                3                0    0   \n",
            "2                    1498         49.4                5                0    0   \n",
            "3                    1433         39.5                7                0    0   \n",
            "4                    1408         40.0                9                0    0   \n",
            "\n",
            "   HDF  PWF  OSF  RNF  \n",
            "0    0    0    0    0  \n",
            "1    0    0    0    0  \n",
            "2    0    0    0    0  \n",
            "3    0    0    0    0  \n",
            "4    0    0    0    0  \n",
            "\n",
            "Missing Values:\n",
            "UDI                        0\n",
            "Product ID                 0\n",
            "Type                       0\n",
            "Air temperature [K]        0\n",
            "Process temperature [K]    0\n",
            "Rotational speed [rpm]     0\n",
            "Torque [Nm]                0\n",
            "Tool wear [min]            0\n",
            "Machine failure            0\n",
            "TWF                        0\n",
            "HDF                        0\n",
            "PWF                        0\n",
            "OSF                        0\n",
            "RNF                        0\n",
            "dtype: int64\n",
            "\n",
            "Target Distribution:\n",
            "Machine failure\n",
            "0    9661\n",
            "1     339\n",
            "Name: count, dtype: int64\n",
            "\n",
            "Dataset registered successfully.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 145,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F4BrFFPTrqYc",
        "outputId": "f1714171-815c-4166-95d4-44343c35b8ee"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/tmp/ipykernel_1044/4257184302.py:35: DeprecationWarning: Argument login_or_token is deprecated, please use auth=github.Auth.Token(...) instead\n",
            "  gh = Github(os.environ[\"GH_TOKEN\"])\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Repo already exists: Swetha1929/predictive-maintenance-mlops\n",
            "updated week_3_mls/requirements.txt\n",
            "updated week_3_mls/data/machine-failure-prediction.csv\n",
            "added   week_3_mls/deployment/best_machine_failure_model_v1.joblib\n",
            "updated week_3_mls/deployment/app.py\n",
            "updated week_3_mls/deployment/requirements.txt\n",
            "updated week_3_mls/model_building/prep.py\n",
            "updated week_3_mls/model_building/train.py\n",
            "updated week_3_mls/model_building/data_register.py\n",
            "updated .github/workflows/pipeline.yml\n",
            "\n",
            "Verifying repo layout...\n",
            "  ok       week_3_mls/requirements.txt\n",
            "  ok       week_3_mls/model_building/data_register.py\n",
            "  ok       week_3_mls/model_building/prep.py\n",
            "  ok       week_3_mls/model_building/train.py\n",
            "  ok       week_3_mls/data/machine-failure-prediction.csv\n",
            "  ok       .github/workflows/pipeline.yml\n",
            "  ok       week_3_mls/deployment/app.py\n",
            "  ok       week_3_mls/deployment/requirements.txt\n",
            "\n",
            "All files pushed.\n"
          ]
        }
      ],
      "source": [
        "import os\n",
        "from pathlib import Path\n",
        "from github import Github, GithubException\n",
        "\n",
        "# -----------------------------\n",
        "# Paths\n",
        "# -----------------------------\n",
        "ROOT = Path(\"/content\")\n",
        "PROJECT_DIR = ROOT / \"week_3_mls\"\n",
        "DATA_DIR = PROJECT_DIR / \"data\"\n",
        "\n",
        "LEGACY_CSV = PROJECT_DIR / \"machine failure.csv\"\n",
        "TARGET_CSV = DATA_DIR / \"machine-failure-prediction.csv\"\n",
        "\n",
        "# -----------------------------\n",
        "# Make sure folders exist\n",
        "# -----------------------------\n",
        "DATA_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "# -----------------------------\n",
        "# Move/rename dataset if needed\n",
        "# -----------------------------\n",
        "if LEGACY_CSV.exists() and not TARGET_CSV.exists():\n",
        "    LEGACY_CSV.replace(TARGET_CSV)\n",
        "    print(f\"Moved dataset to: {TARGET_CSV}\")\n",
        "\n",
        "assert TARGET_CSV.exists(), (\n",
        "    f\"{TARGET_CSV} not found. Place the dataset in week_3_mls/data/ \"\n",
        "    \"and name it machine-failure-prediction.csv before running this cell.\"\n",
        ")\n",
        "\n",
        "# -----------------------------\n",
        "# Connect to GitHub\n",
        "# -----------------------------\n",
        "gh = Github(os.environ[\"GH_TOKEN\"])\n",
        "user = gh.get_user()\n",
        "\n",
        "try:\n",
        "    repo = gh.get_repo(REPO)\n",
        "    print(\"Repo already exists:\", repo.full_name)\n",
        "except GithubException:\n",
        "    repo = user.create_repo(\n",
        "        REPO_NAME,\n",
        "        private=False,\n",
        "        auto_init=True,\n",
        "    )\n",
        "    print(\"Repo created:\", repo.full_name)\n",
        "\n",
        "# -----------------------------\n",
        "# Push local folders to repo\n",
        "# -----------------------------\n",
        "def push_folder(local_dir):\n",
        "    for root, _, files in os.walk(local_dir):\n",
        "        for fname in files:\n",
        "            local_path = os.path.join(root, fname)\n",
        "            repo_path = local_path  # keep same relative path in GitHub\n",
        "            content = open(local_path, \"rb\").read()\n",
        "\n",
        "            try:\n",
        "                sha = repo.get_contents(repo_path, ref=BRANCH).sha\n",
        "                repo.update_file(\n",
        "                    repo_path,\n",
        "                    f\"update {repo_path}\",\n",
        "                    content,\n",
        "                    sha,\n",
        "                    branch=BRANCH,\n",
        "                )\n",
        "                print(\"updated\", repo_path)\n",
        "            except GithubException:\n",
        "                repo.create_file(\n",
        "                    repo_path,\n",
        "                    f\"add {repo_path}\",\n",
        "                    content,\n",
        "                    branch=BRANCH,\n",
        "                )\n",
        "                print(\"added  \", repo_path)\n",
        "\n",
        "push_folder(\"week_3_mls\")\n",
        "push_folder(\".github\")\n",
        "\n",
        "# -----------------------------\n",
        "# Verify repo layout\n",
        "# -----------------------------\n",
        "print(\"\\nVerifying repo layout...\")\n",
        "required = [\n",
        "    \"week_3_mls/requirements.txt\",\n",
        "    \"week_3_mls/model_building/data_register.py\",\n",
        "    \"week_3_mls/model_building/prep.py\",\n",
        "    \"week_3_mls/model_building/train.py\",\n",
        "    \"week_3_mls/data/machine-failure-prediction.csv\",\n",
        "    \".github/workflows/pipeline.yml\",\n",
        "    \"week_3_mls/deployment/app.py\",\n",
        "    \"week_3_mls/deployment/requirements.txt\",\n",
        "]\n",
        "\n",
        "all_ok = True\n",
        "for f in required:\n",
        "    try:\n",
        "        repo.get_contents(f, ref=BRANCH)\n",
        "        print(\"  ok      \", f)\n",
        "    except GithubException:\n",
        "        print(\"  MISSING \", f)\n",
        "        all_ok = False\n",
        "\n",
        "print(\"\\nAll files pushed.\" if all_ok else \"\\nSome files are missing - check the list above.\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from github import Github\n",
        "import os\n",
        "\n",
        "# GitHub connection\n",
        "gh = Github(os.environ[\"GH_TOKEN\"])\n",
        "repo = gh.get_repo(REPO)\n",
        "\n",
        "# File to push\n",
        "file_path = \"week_3_mls/model_building/train.py\"\n",
        "\n",
        "# Read local file\n",
        "with open(file_path, \"rb\") as f:\n",
        "    content = f.read()\n",
        "\n",
        "try:\n",
        "    # Update existing file\n",
        "    file = repo.get_contents(file_path, ref=BRANCH)\n",
        "    repo.update_file(\n",
        "        path=file_path,\n",
        "        message=\"Update train.py\",\n",
        "        content=content,\n",
        "        sha=file.sha,\n",
        "        branch=BRANCH\n",
        "    )\n",
        "    print(\"✅ train.py updated successfully.\")\n",
        "\n",
        "except Exception:\n",
        "    # Create file if it doesn't exist\n",
        "    repo.create_file(\n",
        "        path=file_path,\n",
        "        message=\"Add train.py\",\n",
        "        content=content,\n",
        "        branch=BRANCH\n",
        "    )\n",
        "    print(\"✅ train.py created successfully.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_fdMrG73eThz",
        "outputId": "4fc4345d-e411-4768-a80c-8f007125e1c6"
      },
      "execution_count": 147,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/tmp/ipykernel_1044/793259524.py:5: DeprecationWarning: Argument login_or_token is deprecated, please use auth=github.Auth.Token(...) instead\n",
            "  gh = Github(os.environ[\"GH_TOKEN\"])\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "✅ train.py updated successfully.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from github import Github\n",
        "import os\n",
        "\n",
        "# Connect to GitHub\n",
        "gh = Github(os.environ[\"GH_TOKEN\"])\n",
        "repo = gh.get_repo(REPO)\n",
        "\n",
        "# File to update\n",
        "file_path = \"week_3_mls/model_building/data_register.py\"\n",
        "\n",
        "# Read the local file\n",
        "with open(file_path, \"rb\") as f:\n",
        "    content = f.read()\n",
        "\n",
        "try:\n",
        "    # Update if it already exists\n",
        "    file = repo.get_contents(file_path, ref=BRANCH)\n",
        "\n",
        "    repo.update_file(\n",
        "        path=file_path,\n",
        "        message=\"Update data_register.py\",\n",
        "        content=content,\n",
        "        sha=file.sha,\n",
        "        branch=BRANCH\n",
        "    )\n",
        "\n",
        "    print(\"✅ data_register.py updated successfully.\")\n",
        "\n",
        "except Exception:\n",
        "    # Create if it doesn't exist\n",
        "    repo.create_file(\n",
        "        path=file_path,\n",
        "        message=\"Add data_register.py\",\n",
        "        content=content,\n",
        "        branch=BRANCH\n",
        "    )\n",
        "\n",
        "    print(\"✅ data_register.py created successfully.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FFiLXwvVfqjG",
        "outputId": "51b6fdea-f5a9-453f-8af7-d29ebb9c7c61"
      },
      "execution_count": 150,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/tmp/ipykernel_1044/1283134696.py:5: DeprecationWarning: Argument login_or_token is deprecated, please use auth=github.Auth.Token(...) instead\n",
            "  gh = Github(os.environ[\"GH_TOKEN\"])\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "✅ data_register.py updated successfully.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from github import Github\n",
        "import os\n",
        "\n",
        "gh = Github(os.environ[\"GH_TOKEN\"])\n",
        "repo = gh.get_repo(REPO)\n",
        "\n",
        "file_path = \"week_3_mls/model_building/prep.py\"\n",
        "\n",
        "with open(file_path, \"rb\") as f:\n",
        "    content = f.read()\n",
        "\n",
        "try:\n",
        "    file = repo.get_contents(file_path, ref=BRANCH)\n",
        "\n",
        "    repo.update_file(\n",
        "        path=file_path,\n",
        "        message=\"Update prep.py\",\n",
        "        content=content,\n",
        "        sha=file.sha,\n",
        "        branch=BRANCH\n",
        "    )\n",
        "\n",
        "    print(\"✅ prep.py updated successfully.\")\n",
        "\n",
        "except Exception:\n",
        "    repo.create_file(\n",
        "        path=file_path,\n",
        "        message=\"Add prep.py\",\n",
        "        content=content,\n",
        "        branch=BRANCH\n",
        "    )\n",
        "\n",
        "    print(\"✅ prep.py created successfully.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fxDWGf7AgOUI",
        "outputId": "d0d9ab59-ef3e-4bee-c224-e0383624e5e6"
      },
      "execution_count": 152,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/tmp/ipykernel_1044/2163703070.py:4: DeprecationWarning: Argument login_or_token is deprecated, please use auth=github.Auth.Token(...) instead\n",
            "  gh = Github(os.environ[\"GH_TOKEN\"])\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "✅ prep.py updated successfully.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EHl1Wm1CrqYc"
      },
      "source": [
        "# Run the Pipeline, then Deploy on Streamlit Community Cloud\n",
        "\n",
        "## 1. Run the GitHub Actions pipeline\n",
        "After the push above finishes, open your repo → **Actions** tab. Select **Week 3 MLS CI/CD Pipeline**\n",
        "→ **Run workflow** → **Run workflow** (it will also run automatically on every push to `main`).\n",
        "\n",
        "You'll see three jobs run in order: **Register Dataset → Data Preparation → Model Training**.\n",
        "The **Model Training** job starts a local MLflow server, runs the hyperparameter search while logging\n",
        "every trial to it, and finishes by committing the trained model to\n",
        "`week_3_mls/deployment/best_machine_failure_model_v1.joblib` on `main`.\n",
        "\n",
        "## 2. Deploy the app on Streamlit Community Cloud\n",
        "1. Go to **https://share.streamlit.io** and sign in with the GitHub account that owns the repo.\n",
        "2. Click **Create app**\n",
        "3. Fill in:\n",
        "   - **Repository:** `your-username/Machine_Failure_Prediction`\n",
        "   - **Branch:** `main`\n",
        "   - **Main file path:** `week_3_mls/deployment/app.py`\n",
        "4. Open **Advanced settings** and set **Python version** to **3.11** (this matches the version that\n",
        "   trained the model in GitHub Actions, so the saved `.joblib` loads cleanly).\n",
        "5. Click **Deploy**.\n",
        "\n",
        "Streamlit installs dependencies from `week_3_mls/deployment/requirements.txt` (it sits next to\n",
        "`app.py`, so it takes precedence) and launches your app at a public `…streamlit.app` URL.\n",
        "\n",
        "**Note:**\n",
        "*   Refer to the Model Deployment via Streamlit and GitHub guide to set up Streamlit Community Cloud deployment.\n",
        "*   Refer to the GitHub Account and Token Guide to create and configure your GitHub Personal Access Token (PAT)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fN8j9-3nW8G9"
      },
      "source": [
        "<font size=6 color=\"navyblue\">Power Ahead!</font>\n",
        "___"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}