examples/pzmm_generate_complete_model_card.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright © 2024, SAS Institute Inc., Cary, NC, USA.  All Rights Reserved.\n",
    "SPDX-License-Identifier: Apache-2.0\n",
    "\n",
    "# Creating a Complete Model Card for a Python Model using Python-sasctl\n",
    "This notebook walks through the steps for building a complete model card for a Python model. \n",
    "\n",
    "***\n",
    "##  Table of Contents\n",
    "1. [Introduction](#Introduction)\n",
    "1. [Gather Resources](#Gather-Resources)\n",
    "1. [Explore and Prepare Data](#Explore-and-Prepare-Data)\n",
    "1. [Train and Assess Model](#Train-and-Assess-Model)\n",
    "1. [Prepare Model Files and Register into SAS Model Manager](#Prepare-Model-Files-and-Register-into-SAS-Model-Manager)\n",
    "1. [Conclusion](#Conclusion)\n",
    "\n",
    "***\n",
    "## Introduction\n",
    "Model cards were introduced with the SAS Viya 2024.07 release and act as a nutrition label for AI models. The model cards in SAS Viya feature easy-to-understand visuals and actionable takeaways. When designing the model card, SAS made sure it contained useful and digestible information for a variety of stakeholders and would become a natural extension of the model’s lifecycle. As you develop and manage models within SAS Viya, the model card starts to populate. Complete model cards can be built for Python models using the python-sasctl package and tools on SAS Viya. This notebook will review the steps performed via Python-sasctl. \n",
    "\n",
    "*** \n",
    "## Gather Resources \n",
    "First, we need to import the packages we will use. Make sure you are using the latest version of python-sasctl!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Standard Library\n",
    "from pathlib import Path\n",
    "import warnings\n",
    "import os\n",
    "import json\n",
    "\n",
    "# Third Party\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "import getpass\n",
    "\n",
    "# Application Specific\n",
    "import sasctl.pzmm as pzmm\n",
    "from sasctl import Session\n",
    "from sasctl.services import model_repository as mr\n",
    "\n",
    "# Global Package Options\n",
    "pd.options.mode.chained_assignment = None  # default=\"warn\"\n",
    "plt.rc(\"font\", size=14)\n",
    "\n",
    "# Ignore warnings from pandas about SWAT using a feature that will be depreciated soon\n",
    "warnings.simplefilter(action=\"ignore\", category=FutureWarning)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we need to import our data. We are using the [Adult dataset](https://archive.ics.uci.edu/dataset/2/adult) from the UC Irvine Machine Learning Repository. Before bringing the data into this notebook, I removed the fnlwgt and education-num columns and created a binary indicator for the income variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of data: (32561, 11)\n"
     ]
    }
   ],
   "source": [
    "adult = pd.read_csv(\"data/adult.csv\", sep= \",\")\n",
    "\n",
    "print(\"Shape of data:\", adult.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***\n",
    "## Explore and Prepare Data\n",
    "Before moving further, let's explore the dataset. We'll start by looking at a sample of the data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>Age</th>\n",
       "      <th>WorkClass</th>\n",
       "      <th>Education</th>\n",
       "      <th>MartialStatus</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>Relationship</th>\n",
       "      <th>Race</th>\n",
       "      <th>Sex</th>\n",
       "      <th>HoursPerWeek</th>\n",
       "      <th>NativeCountry</th>\n",
       "      <th>GE50K</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>13</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Divorced</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>Wife</td>\n",
       "      <td>Black</td>\n",
       "      <td>Female</td>\n",
       "      <td>40</td>\n",
       "      <td>Cuba</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Age         WorkClass  Education       MartialStatus         Occupation  \\\n",
       "0   39         State-gov  Bachelors       Never-married       Adm-clerical   \n",
       "1   50  Self-emp-not-inc  Bachelors  Married-civ-spouse    Exec-managerial   \n",
       "2   38           Private    HS-grad            Divorced  Handlers-cleaners   \n",
       "3   53           Private       11th  Married-civ-spouse  Handlers-cleaners   \n",
       "4   28           Private  Bachelors  Married-civ-spouse     Prof-specialty   \n",
       "\n",
       "    Relationship   Race     Sex  HoursPerWeek  NativeCountry  GE50K  \n",
       "0  Not-in-family  White    Male            40  United-States      0  \n",
       "1        Husband  White    Male            13  United-States      0  \n",
       "2  Not-in-family  White    Male            40  United-States      0  \n",
       "3        Husband  Black    Male            40  United-States      0  \n",
       "4           Wife  Black  Female            40           Cuba      0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adult.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we'll look at missing values. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Age                 0\n",
       "WorkClass        1836\n",
       "Education           0\n",
       "MartialStatus       0\n",
       "Occupation       1843\n",
       "Relationship        0\n",
       "Race                0\n",
       "Sex                 0\n",
       "HoursPerWeek        0\n",
       "NativeCountry     583\n",
       "GE50K               0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adult.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And at the distribution of the continuous variables.   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x1500 with 16 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "adult.hist(figsize=(15,15), layout=(4, 4));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's also look at the unique values and value counts for the categorical variables. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unique Values: \n",
      "WorkClass         8\n",
      "Education        16\n",
      "MartialStatus     7\n",
      "Occupation       14\n",
      "Relationship      6\n",
      "Race              5\n",
      "Sex               2\n",
      "NativeCountry    41\n",
      "dtype: int64\n",
      "--------------------------------\n",
      "Value Counts: \n",
      "WorkClass\n",
      "Private             22696\n",
      "Self-emp-not-inc     2541\n",
      "Local-gov            2093\n",
      "State-gov            1298\n",
      "Self-emp-inc         1116\n",
      "Federal-gov           960\n",
      "Without-pay            14\n",
      "Never-worked            7\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n",
      "Education\n",
      "HS-grad         10501\n",
      "Some-college     7291\n",
      "Bachelors        5355\n",
      "Masters          1723\n",
      "Assoc-voc        1382\n",
      "11th             1175\n",
      "Assoc-acdm       1067\n",
      "10th              933\n",
      "7th-8th           646\n",
      "Prof-school       576\n",
      "9th               514\n",
      "12th              433\n",
      "Doctorate         413\n",
      "5th-6th           333\n",
      "1st-4th           168\n",
      "Preschool          51\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n",
      "MartialStatus\n",
      "Married-civ-spouse       14976\n",
      "Never-married            10683\n",
      "Divorced                  4443\n",
      "Separated                 1025\n",
      "Widowed                    993\n",
      "Married-spouse-absent      418\n",
      "Married-AF-spouse           23\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n",
      "Occupation\n",
      "Prof-specialty       4140\n",
      "Craft-repair         4099\n",
      "Exec-managerial      4066\n",
      "Adm-clerical         3770\n",
      "Sales                3650\n",
      "Other-service        3295\n",
      "Machine-op-inspct    2002\n",
      "Transport-moving     1597\n",
      "Handlers-cleaners    1370\n",
      "Farming-fishing       994\n",
      "Tech-support          928\n",
      "Protective-serv       649\n",
      "Priv-house-serv       149\n",
      "Armed-Forces            9\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n",
      "Relationship\n",
      "Husband           13193\n",
      "Not-in-family      8305\n",
      "Own-child          5068\n",
      "Unmarried          3446\n",
      "Wife               1568\n",
      "Other-relative      981\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n",
      "Race\n",
      "White                 27816\n",
      "Black                  3124\n",
      "Asian-Pac-Islander     1039\n",
      "Amer-Indian-Eskimo      311\n",
      "Other                   271\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n",
      "Sex\n",
      "Male      21790\n",
      "Female    10771\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n",
      "NativeCountry\n",
      "United-States                 29170\n",
      "Mexico                          643\n",
      "Philippines                     198\n",
      "Germany                         137\n",
      "Canada                          121\n",
      "Puerto-Rico                     114\n",
      "El-Salvador                     106\n",
      "India                           100\n",
      "Cuba                             95\n",
      "England                          90\n",
      "Jamaica                          81\n",
      "South                            80\n",
      "China                            75\n",
      "Italy                            73\n",
      "Dominican-Republic               70\n",
      "Vietnam                          67\n",
      "Guatemala                        64\n",
      "Japan                            62\n",
      "Poland                           60\n",
      "Columbia                         59\n",
      "Taiwan                           51\n",
      "Haiti                            44\n",
      "Iran                             43\n",
      "Portugal                         37\n",
      "Nicaragua                        34\n",
      "Peru                             31\n",
      "France                           29\n",
      "Greece                           29\n",
      "Ecuador                          28\n",
      "Ireland                          24\n",
      "Hong                             20\n",
      "Cambodia                         19\n",
      "Trinadad&Tobago                  19\n",
      "Laos                             18\n",
      "Thailand                         18\n",
      "Yugoslavia                       16\n",
      "Outlying-US(Guam-USVI-etc)       14\n",
      "Honduras                         13\n",
      "Hungary                          13\n",
      "Scotland                         12\n",
      "Holand-Netherlands                1\n",
      "Name: count, dtype: int64\n",
      "--------------------------------\n"
     ]
    }
   ],
   "source": [
    "print(\"Unique Values: \")\n",
    "print(adult[adult.select_dtypes(include=['object']).columns].nunique())\n",
    "print(\"--------------------------------\")\n",
    "print(\"Value Counts: \")\n",
    "for i in adult.select_dtypes(include=['object']).columns:\n",
    "     print(adult[i].value_counts())\n",
    "     print(\"--------------------------------\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is a lot of unique values for the categorical variables. Let's combine similar categorical values and one-hot-encode our categorical variables. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = adult.copy()\n",
    "df = df.dropna().reset_index()\n",
    "cat_vals = df[[\"WorkClass\", \"Education\", \"MartialStatus\", \"Relationship\", \"Race\", \"Sex\"]]\n",
    "df = pd.get_dummies(df, columns=[\"WorkClass\", \"Education\", \"MartialStatus\", \"Relationship\", \"Race\", \"Sex\"])\n",
    "df.columns = df.columns.str.replace(' ', '')\n",
    "df.columns = df.columns.str.replace('-', '_')\n",
    "df = df.drop(['Sex_Male'], axis=1)\n",
    "df = pd.concat([df, cat_vals], axis=1).drop('index', axis=1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data is looking better, but the martial status, education and work class statuses are a bit too granular. Lets combine some of them to make the job easier for our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Age', 'Occupation', 'HoursPerWeek', 'NativeCountry', 'GE50K',\n",
       "       'WorkClass_Federal_gov', 'WorkClass_Local_gov', 'WorkClass_Private',\n",
       "       'WorkClass_Self_emp_inc', 'WorkClass_Self_emp_not_inc',\n",
       "       'WorkClass_State_gov', 'WorkClass_Without_pay', 'Education_10th',\n",
       "       'Education_11th', 'Education_12th', 'Education_1st_4th',\n",
       "       'Education_5th_6th', 'Education_7th_8th', 'Education_9th',\n",
       "       'Education_Assoc_acdm', 'Education_Assoc_voc', 'Education_Bachelors',\n",
       "       'Education_Doctorate', 'Education_HS_grad', 'Education_Masters',\n",
       "       'Education_Preschool', 'Education_Prof_school',\n",
       "       'Education_Some_college', 'MartialStatus_Divorced',\n",
       "       'MartialStatus_Married_AF_spouse', 'MartialStatus_Married_civ_spouse',\n",
       "       'MartialStatus_Married_spouse_absent', 'MartialStatus_Never_married',\n",
       "       'MartialStatus_Separated', 'MartialStatus_Widowed',\n",
       "       'Relationship_Husband', 'Relationship_Not_in_family',\n",
       "       'Relationship_Other_relative', 'Relationship_Own_child',\n",
       "       'Relationship_Unmarried', 'Relationship_Wife',\n",
       "       'Race_Amer_Indian_Eskimo', 'Race_Asian_Pac_Islander', 'Race_Black',\n",
       "       'Race_Other', 'Race_White', 'Sex_Female', 'WorkClass', 'Education',\n",
       "       'MartialStatus', 'Relationship', 'Race', 'Sex', 'Education_Some_HS',\n",
       "       'Education_Assoc', 'Education_Adv_Degree', 'Education_No_HS',\n",
       "       'WorkClass_Self', 'WorkClass_Gov', 'WorkClass_Other',\n",
       "       'MartialStatus_Other'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"Education_Some_HS\"] = df[\"Education_9th\"] | df[\"Education_10th\"] | df[\"Education_11th\"] | df[\"Education_12th\"]\n",
    "df[\"Education_Assoc\"] = df[\"Education_Assoc_voc\"] | df[\"Education_Assoc_acdm\"]\n",
    "df[\"Education_Adv_Degree\"] = df[\"Education_Masters\"] | df[\"Education_Prof_school\"] | df[\"Education_Doctorate\"]\n",
    "df[\"Education_No_HS\"] = df[\"Education_Preschool\"] | df[\"Education_1st_4th\"] | df[\"Education_5th_6th\"] | df[\"Education_7th_8th\"]\n",
    "\n",
    "df[\"WorkClass_Self\"] = df[\"WorkClass_Self_emp_inc\"] | df[\"WorkClass_Self_emp_not_inc\"]\n",
    "df[\"WorkClass_Gov\"] = df[\"WorkClass_Federal_gov\"] | df[\"WorkClass_Local_gov\"] | df[\"WorkClass_State_gov\"]\n",
    "df[\"WorkClass_Other\"] = df[\"WorkClass_Without_pay\"] # df[\"WorkClass_Never_worked\"] is fully dropped when all NA values removed\n",
    "\n",
    "df[\"MartialStatus_Other\"] = df[\"MartialStatus_Married_spouse_absent\"] | df[\"MartialStatus_Married_AF_spouse\"]\n",
    "\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we want the automatically generated score code to leaverage these steps when scoring new data, we can put them in a preprocessing function and pass them into our import_model function call. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def preprocess_function(df):\n",
    "    cat_vals = df[[\"WorkClass\", \"Education\", \"MartialStatus\", \"Relationship\", \"Race\", \"Sex\"]]\n",
    "    df = pd.get_dummies(df, columns=[\"WorkClass\", \"Education\", \"MartialStatus\", \"Relationship\", \"Race\", \"Sex\"])\n",
    "    df.columns = df.columns.str.replace(' ', '')\n",
    "    df.columns = df.columns.str.replace('-', '_')\n",
    "    df = df.drop(['Sex_Male'], axis=1)\n",
    "    if 'index' in df.columns or 'index' in cat_vals.columns:\n",
    "        df = pd.concat([df, cat_vals], axis=1).drop('index', axis=1)\n",
    "    # For the model to score correctly, all OHE columns must exist\n",
    "    input_cols = [\n",
    "        \"Education_9th\", \"Education_10th\", \"Education_11th\", \"Education_12th\", \"Education_Assoc_voc\", \"Education_Assoc_acdm\", \"Education_Masters\", \"Education_Prof_school\",\n",
    "        \"Education_Doctorate\", \"Education_Preschool\", \"Education_1st_4th\", \"Education_5th_6th\", \"Education_7th_8th\", \"WorkClass_Self_emp_inc\", \"WorkClass_Self_emp_not_inc\",\n",
    "        \"WorkClass_Federal_gov\", \"WorkClass_Local_gov\", \"WorkClass_State_gov\", \"WorkClass_Without_pay\", \"WorkClass_Never_worked\", \"MartialStatus_Married_spouse_absent\",\n",
    "        \"MartialStatus_Married_AF_spouse\", 'MartialStatus_Married_civ_spouse', 'MartialStatus_Never_married', 'MartialStatus_Divorced', 'MartialStatus_Separated', \n",
    "        'MartialStatus_Widowed', 'Race_White', 'Race_Black', 'Race_Asian_Pac_Islander', 'Race_Amer_Indian_Eskimo', 'Race_Other', 'Relationship_Husband', \n",
    "        'Relationship_Not_in_family', 'Relationship_Own_child', 'Relationship_Unmarried', 'Relationship_Wife', 'Relationship_Other_relative', 'WorkClass_Private',\n",
    "        'Education_Bachelors'\n",
    "    ]\n",
    "    # OHE columns must be removed after data combination\n",
    "    predictor_columns = ['Age', 'HoursPerWeek', 'WorkClass_Private', 'WorkClass_Self', 'WorkClass_Gov', \n",
    "       'WorkClass_Other', 'Education_HS_grad', 'Education_Some_HS', 'Education_Assoc', 'Education_Some_college',\n",
    "       'Education_Bachelors', 'Education_Adv_Degree', 'Education_No_HS', 'MartialStatus_Married_civ_spouse',\n",
    "       'MartialStatus_Never_married', 'MartialStatus_Divorced', 'MartialStatus_Separated', 'MartialStatus_Widowed',\n",
    "       'MartialStatus_Other', 'Relationship_Husband', 'Relationship_Not_in_family', 'Relationship_Own_child', 'Relationship_Unmarried',\n",
    "       'Relationship_Wife', 'Relationship_Other_relative', 'Race_White', 'Race_Black', 'Race_Asian_Pac_Islander',\n",
    "       'Race_Amer_Indian_Eskimo', 'Race_Other', 'Sex_Female']\n",
    "\n",
    "    for col in input_cols:\n",
    "        if col not in df.columns:\n",
    "            df[col] = 0\n",
    "    \n",
    "\n",
    "    df[\"Education_Some_HS\"] = df[\"Education_9th\"] | df[\"Education_10th\"] | df[\"Education_11th\"] | df[\"Education_12th\"]\n",
    "    df[\"Education_Assoc\"] = df[\"Education_Assoc_voc\"] | df[\"Education_Assoc_acdm\"]\n",
    "    df[\"Education_Adv_Degree\"] = df[\"Education_Masters\"] | df[\"Education_Prof_school\"] | df[\"Education_Doctorate\"]\n",
    "    df[\"Education_No_HS\"] = df[\"Education_Preschool\"] | df[\"Education_1st_4th\"] | df[\"Education_5th_6th\"] | df[\"Education_7th_8th\"]\n",
    "\n",
    "    df[\"WorkClass_Self\"] = df[\"WorkClass_Self_emp_inc\"] | df[\"WorkClass_Self_emp_not_inc\"]\n",
    "    df[\"WorkClass_Gov\"] = df[\"WorkClass_Federal_gov\"] | df[\"WorkClass_Local_gov\"] | df[\"WorkClass_State_gov\"]\n",
    "    df[\"WorkClass_Other\"] = df[\"WorkClass_Without_pay\"] | df[\"WorkClass_Never_worked\"]\n",
    "\n",
    "    df[\"MartialStatus_Other\"] = df[\"MartialStatus_Married_spouse_absent\"] | df[\"MartialStatus_Married_AF_spouse\"]\n",
    "\n",
    "    df = df[predictor_columns]\n",
    "\n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, let's create a training, testing, and validation set. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "original_columns = ['Age', 'WorkClass', 'Education', 'MartialStatus', \n",
    "       'Relationship', 'Race', 'Sex', 'HoursPerWeek', 'GE50K']\n",
    "\n",
    "original_inputs = ['Age', 'WorkClass', 'Education', 'MartialStatus', \n",
    "       'Relationship', 'Race', 'Sex', 'HoursPerWeek']\n",
    "\n",
    "predictor_columns = ['Age', 'HoursPerWeek', 'WorkClass_Private', 'WorkClass_Self', 'WorkClass_Gov', \n",
    "       'WorkClass_Other', 'Education_HS_grad', 'Education_Some_HS', 'Education_Assoc', 'Education_Some_college',\n",
    "       'Education_Bachelors', 'Education_Adv_Degree', 'Education_No_HS', 'MartialStatus_Married_civ_spouse',\n",
    "       'MartialStatus_Never_married', 'MartialStatus_Divorced', 'MartialStatus_Separated', 'MartialStatus_Widowed',\n",
    "       'MartialStatus_Other', 'Relationship_Husband', 'Relationship_Not_in_family', 'Relationship_Own_child', 'Relationship_Unmarried',\n",
    "       'Relationship_Wife', 'Relationship_Other_relative', 'Race_White', 'Race_Black', 'Race_Asian_Pac_Islander',\n",
    "       'Race_Amer_Indian_Eskimo', 'Race_Other', 'Sex_Female']\n",
    "\n",
    "target_column = \"GE50K\"\n",
    "\n",
    "x = df[list(dict.fromkeys(original_columns + predictor_columns))]\n",
    "\n",
    "y = df[target_column]\n",
    "\n",
    "x_train_full, x_test_full, y_train, y_test = train_test_split(x, y, test_size=.2, random_state=42)\n",
    "x_train_full, x_val_full, y_train, y_val = train_test_split(x_train_full, y_train, test_size=.2, random_state=42)\n",
    "\n",
    "x_train = x_train_full[predictor_columns]\n",
    "x_test = x_test_full[predictor_columns]\n",
    "x_val = x_val_full[predictor_columns]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***\n",
    "## Train and Assess Model\n",
    "Our data is ready for modeling, so let's train a simple decision tree model!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtc = DecisionTreeClassifier(max_depth=7, min_samples_split=2, min_samples_leaf=2, max_leaf_nodes=500)\n",
    "dtc = dtc.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Great, let's look at how well our model performed. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_axis =np.array([\"Training\", \"Testing\", \"Validation\"])\n",
    "x_axis = np.array([(dtc.score(x_train, y_train)), (dtc.score(x_test, y_test)), (dtc.score(x_val, y_val))])\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "p = ax.bar(x = y_axis, height = x_axis)\n",
    "ax.bar_label(p, label_type='center')\n",
    "ax.set_title('Decision Tree Accuarcy Across Each Set')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare Model Files and Register into SAS Model Manager\n",
    "Since I'm happy with my model, let's add it into SAS Model Manager with all the metadata it needs to build a Model Card. Once the model is within SAS Model Manager, I can build a complete card by updating the model usage properties, specifying thresholds for Key Performance Indicators (KPIs), and running performance monitoring reports in SAS Model Manager.\n",
    "But from our Python development environment, I need to do the following:\n",
    "1. Update the variables in the block below to make sure everything matches my current use case. \n",
    "1. Connect to SAS Viya. \n",
    "1. Create a folder for my output files. \n",
    "1. Pickle my model.  \n",
    "1. Write the input variables to a file. \n",
    "1. Write the output variables to a file. \n",
    "1. Write the model properties to a file. \n",
    "1. Write the file metadata.\n",
    "1. Score training, testing, and validation data. \n",
    "1. Write model performance statistics to a file. \n",
    "1. Assess model bias and fairness (if a potentially sensitive variable is available for assessment). \n",
    "1. Generate the last few model card files. \n",
    "1. Generate requirements file. \n",
    "1. Import model to SAS Model Manager and automatically generate the score code. \n",
    "1. Open the model in SAS Model Manager and begin managing the model lifecycle there. \n",
    "\n",
    "So first, be sure that the variables in the block below match your use case."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 1: Update variables below to fit your needs\n",
    "output_directory = 'outputs'\n",
    "model_prefix = 'dtc'\n",
    "model_object = dtc\n",
    "target= \"GE50K\"\n",
    "algorithm = \"Decision Tree\"\n",
    "description = \"SKLearn Decision Tree Model\"\n",
    "assess_bias_var = 'Sex'\n",
    "cat_columns = ['WorkClass', 'Education', 'MartialStatus', 'Relationship', 'Race', 'Sex']\n",
    "interval_columns = ['Age', 'HoursPerWeek']\n",
    "mm_project = \"Salary Classification\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, let's build our connection to SAS Viya. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 2: Connect to SAS Viya\n",
    "username = input(\"Username: \")\n",
    "password = getpass.getpass(\"Password: \")\n",
    "host = input(\"Viya Environment URL: \")\n",
    "sess = Session(host, username, password, protocol=\"http\") # For TLS-enabled servers, change protocol value to \"https\"\n",
    "conn = sess.as_swat() # Connect to SWAT through the sasctl authenticated connection\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's build our metadata files. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 3: Create folder for output files. \n",
    "output_path = Path.cwd() / output_directory / model_prefix\n",
    "if not os.path.exists(output_path):\n",
    "    os.makedirs(output_path)\n",
    "\n",
    "# Step 4: Pickle model\n",
    "pzmm.PickleModel.pickle_trained_model(\n",
    "        model_prefix=model_prefix,\n",
    "        trained_model=model_object,\n",
    "        pickle_path=output_path)\n",
    "\n",
    "# Step 5: Write input variables file\n",
    "pzmm.JSONFiles.write_var_json(input_data=x[original_inputs], is_input=True, json_path=output_path)\n",
    "\n",
    "# Step 6: Write output variables file\n",
    "output_var = pd.DataFrame(columns=[\"EM_CLASSIFICATION\", \"EM_EVENTPROBABILITY\"], data=[[\"A\", 0.5]])\n",
    "pzmm.JSONFiles.write_var_json(output_var, is_input=False, json_path=output_path)\n",
    "\n",
    "# Step 7: Write model properties files\n",
    "pzmm.JSONFiles.write_model_properties_json(\n",
    "    model_name=model_prefix, \n",
    "    target_variable=target, \n",
    "    target_values=[\"1\", \"0\"], \n",
    "    json_path=output_path, \n",
    "    model_desc=description,\n",
    "    model_algorithm=algorithm,\n",
    "    modeler=username,\n",
    ")\n",
    "\n",
    "# Step 8: Write file metadata file\n",
    "pzmm.JSONFiles.write_file_metadata_json(model_prefix=model_prefix, json_path=output_path)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, let's score our training, testing, and validation data sets using our model. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 9: Score training, testing, and validation data \n",
    "t1 = 'P_' + target + '1'\n",
    "t0 = 'P_' + target + '0'\n",
    "ti = 'I_'+ target\n",
    "\n",
    "train_scored = pd.DataFrame({t1: model_object.predict_proba(x_train)[:,1], \n",
    "                                 t0: model_object.predict_proba(x_train)[:,0], \n",
    "                                 ti: model_object.predict(x_train), \n",
    "                                 target: y_train, \n",
    "                                 assess_bias_var: x_train_full[assess_bias_var], \n",
    "                                 'Split': 'Train'})\n",
    "\n",
    "test_scored = pd.DataFrame({t1: model_object.predict_proba(x_test)[:,1], \n",
    "                                 t0: model_object.predict_proba(x_test)[:,0], \n",
    "                                 ti: model_object.predict(x_test),\n",
    "                                 target: y_test, \n",
    "                                 assess_bias_var: x_test_full[assess_bias_var], \n",
    "                                 'Split': 'Testing'})\n",
    "\n",
    "val_scored = pd.DataFrame({t1: model_object.predict_proba(x_val)[:,1], \n",
    "                                 t0: model_object.predict_proba(x_val)[:,0], \n",
    "                                 ti: model_object.predict(x_val),\n",
    "                                 target: y_val, \n",
    "                                 assess_bias_var: x_val_full[assess_bias_var], \n",
    "                                 'Split': 'Validation'})\n",
    "\n",
    "scored = pd.concat([train_scored, test_scored, val_scored])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And write the performance metrics to a file so we can preserve how our model was performing at the time of training. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 10: Write model statistics files\n",
    "pzmm.JSONFiles.calculate_model_statistics(\n",
    "        target_value=1, \n",
    "        prob_value=0.5, \n",
    "        train_data=train_scored[[target, ti, t1]], \n",
    "        test_data=test_scored[[target, ti, t1]],\n",
    "        validate_data=test_scored[[target, ti, t1]],\n",
    "        json_path=output_path\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since we also have variables that contain information about protected classes, we can look at bias and fairness information to determine if our model is scoring differently across these classes or if our model performs better on one class over the other. When generating these values, I recommend using a variable where the group values are in a string format. Let's also print the results below since there is a lot of interesting information returned. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 11: Assess Model Bias\n",
    "dfs = pzmm.JSONFiles.assess_model_bias(\n",
    "        score_table=test_scored,\n",
    "        actual_values=target, \n",
    "        sensitive_values=assess_bias_var, \n",
    "        prob_values=[t1, t0], \n",
    "        levels=['1', '0'], \n",
    "        json_path=output_path, \n",
    "        return_dataframes=True \n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Group Metrics: \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<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>INTO_EVENT</th>\n",
       "      <th>LEVEL</th>\n",
       "      <th>PREDICTED_EVENT</th>\n",
       "      <th>P_GE50K0</th>\n",
       "      <th>P_GE50K1</th>\n",
       "      <th>VLABEL</th>\n",
       "      <th>_DATAROLE_</th>\n",
       "      <th>_VARIABLE_</th>\n",
       "      <th>_acc_</th>\n",
       "      <th>_ase_</th>\n",
       "      <th>...</th>\n",
       "      <th>_mcll_</th>\n",
       "      <th>_misccutoff_</th>\n",
       "      <th>_miscks_</th>\n",
       "      <th>_nobs_</th>\n",
       "      <th>_rase_</th>\n",
       "      <th>_resp_</th>\n",
       "      <th>_tn_</th>\n",
       "      <th>_tnr_</th>\n",
       "      <th>_tp_</th>\n",
       "      <th>_tpr_</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.052524</td>\n",
       "      <td>Female</td>\n",
       "      <td>0.110370</td>\n",
       "      <td>0.889630</td>\n",
       "      <td>0.110370</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.906680</td>\n",
       "      <td>0.068113</td>\n",
       "      <td>...</td>\n",
       "      <td>0.280479</td>\n",
       "      <td>0.093320</td>\n",
       "      <td>0.148394</td>\n",
       "      <td>1961.0</td>\n",
       "      <td>0.260985</td>\n",
       "      <td>22.078538</td>\n",
       "      <td>1707.0</td>\n",
       "      <td>0.981599</td>\n",
       "      <td>71.0</td>\n",
       "      <td>0.319820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.180255</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.312586</td>\n",
       "      <td>0.687414</td>\n",
       "      <td>0.312586</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.777996</td>\n",
       "      <td>0.149287</td>\n",
       "      <td>...</td>\n",
       "      <td>0.501074</td>\n",
       "      <td>0.222004</td>\n",
       "      <td>0.255648</td>\n",
       "      <td>4072.0</td>\n",
       "      <td>0.386377</td>\n",
       "      <td>12.219715</td>\n",
       "      <td>2599.0</td>\n",
       "      <td>0.940304</td>\n",
       "      <td>569.0</td>\n",
       "      <td>0.435015</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   INTO_EVENT   LEVEL  PREDICTED_EVENT  P_GE50K0  P_GE50K1 VLABEL _DATAROLE_  \\\n",
       "0    0.052524  Female         0.110370  0.889630  0.110370              TEST   \n",
       "1    0.180255    Male         0.312586  0.687414  0.312586              TEST   \n",
       "\n",
       "  _VARIABLE_     _acc_     _ase_  ...    _mcll_  _misccutoff_  _miscks_  \\\n",
       "0        Sex  0.906680  0.068113  ...  0.280479      0.093320  0.148394   \n",
       "1        Sex  0.777996  0.149287  ...  0.501074      0.222004  0.255648   \n",
       "\n",
       "   _nobs_    _rase_     _resp_    _tn_     _tnr_   _tp_     _tpr_  \n",
       "0  1961.0  0.260985  22.078538  1707.0  0.981599   71.0  0.319820  \n",
       "1  4072.0  0.386377  12.219715  2599.0  0.940304  569.0  0.435015  \n",
       "\n",
       "[2 rows x 35 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------\n",
      "Max Differences: \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<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>BASE</th>\n",
       "      <th>COMPARE</th>\n",
       "      <th>Metric</th>\n",
       "      <th>MetricLabel</th>\n",
       "      <th>VLABEL</th>\n",
       "      <th>_DATAROLE_</th>\n",
       "      <th>_VARIABLE_</th>\n",
       "      <th>maxdiff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>P_GE50K1</td>\n",
       "      <td>Average Predicted: GE50K=1</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.20221505042769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>P_GE50K0</td>\n",
       "      <td>Average Predicted: GE50K=0</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.20221505042769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>TPR</td>\n",
       "      <td>True Positive Rate</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.11519547070005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>FPR</td>\n",
       "      <td>False Positive Rate</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.04129471251588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>TNR</td>\n",
       "      <td>True Negative Rate</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.04129471251588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>FNR</td>\n",
       "      <td>False Negative Rate</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.11519547070005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>FDR</td>\n",
       "      <td>False Discovery Rate</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.0858839713235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>ACC</td>\n",
       "      <td>Accuracy</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.12868419444391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>C</td>\n",
       "      <td>Area under ROC</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.04113329073887</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>F1</td>\n",
       "      <td>F1 Score</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.12037369095155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>GINI</td>\n",
       "      <td>Gini Coefficient</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.08226658147774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>MISCEVENT</td>\n",
       "      <td>Event Misclassification Rate at Cutoff</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.12868419444391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>MISCEVENTKS</td>\n",
       "      <td>Event Misclassification Rate at Maximum KS</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.10725465336337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>MCE</td>\n",
       "      <td>Misclassification Rate</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.12868419444391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>ASE</td>\n",
       "      <td>Average Squared Error</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.08117390649308</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>RASE</td>\n",
       "      <td>Root Average Squared Error</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.1253917503238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>MCLL</td>\n",
       "      <td>Multiclass Log Loss</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.22059451243813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>maxKS</td>\n",
       "      <td>Best Kolmogorov-Smirnov along ROC</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.12528871354234</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>cutoffKS</td>\n",
       "      <td>Kolmogorov-Smirnov at Cutoff</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.07390075818417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>GAIN</td>\n",
       "      <td>Gain</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>2.71511983391718</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>LIFT</td>\n",
       "      <td>Lift</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>1.97176451211151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>RESP</td>\n",
       "      <td>%% Captured Response</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>9.85882256055754</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>CUMRESP</td>\n",
       "      <td>Cumulative %% Captured Response</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>27.1511983391718</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td>CUMLIFT</td>\n",
       "      <td>Cumulative Lift</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>2.71511983391718</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>PREDICTED_EVENT</td>\n",
       "      <td>Average Prediction for Event</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.20221505042769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>INTO_EVENT</td>\n",
       "      <td>Proportion into Event Level</td>\n",
       "      <td></td>\n",
       "      <td>TEST</td>\n",
       "      <td>Sex</td>\n",
       "      <td>0.12773118041494</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      BASE COMPARE           Metric  \\\n",
       "0     Male  Female         P_GE50K1   \n",
       "1   Female    Male         P_GE50K0   \n",
       "2     Male  Female              TPR   \n",
       "3     Male  Female              FPR   \n",
       "4   Female    Male              TNR   \n",
       "5   Female    Male              FNR   \n",
       "6   Female    Male              FDR   \n",
       "7   Female    Male              ACC   \n",
       "8   Female    Male                C   \n",
       "9     Male  Female               F1   \n",
       "10  Female    Male             GINI   \n",
       "11    Male  Female        MISCEVENT   \n",
       "12    Male  Female      MISCEVENTKS   \n",
       "13    Male  Female              MCE   \n",
       "14    Male  Female              ASE   \n",
       "15    Male  Female             RASE   \n",
       "16    Male  Female             MCLL   \n",
       "17  Female    Male            maxKS   \n",
       "18    Male  Female         cutoffKS   \n",
       "19  Female    Male             GAIN   \n",
       "20  Female    Male             LIFT   \n",
       "21  Female    Male             RESP   \n",
       "22  Female    Male          CUMRESP   \n",
       "23  Female    Male          CUMLIFT   \n",
       "24    Male  Female  PREDICTED_EVENT   \n",
       "25    Male  Female       INTO_EVENT   \n",
       "\n",
       "                                   MetricLabel VLABEL _DATAROLE_ _VARIABLE_  \\\n",
       "0                   Average Predicted: GE50K=1              TEST        Sex   \n",
       "1                   Average Predicted: GE50K=0              TEST        Sex   \n",
       "2                           True Positive Rate              TEST        Sex   \n",
       "3                          False Positive Rate              TEST        Sex   \n",
       "4                           True Negative Rate              TEST        Sex   \n",
       "5                          False Negative Rate              TEST        Sex   \n",
       "6                         False Discovery Rate              TEST        Sex   \n",
       "7                                     Accuracy              TEST        Sex   \n",
       "8                               Area under ROC              TEST        Sex   \n",
       "9                                     F1 Score              TEST        Sex   \n",
       "10                            Gini Coefficient              TEST        Sex   \n",
       "11      Event Misclassification Rate at Cutoff              TEST        Sex   \n",
       "12  Event Misclassification Rate at Maximum KS              TEST        Sex   \n",
       "13                      Misclassification Rate              TEST        Sex   \n",
       "14                       Average Squared Error              TEST        Sex   \n",
       "15                  Root Average Squared Error              TEST        Sex   \n",
       "16                         Multiclass Log Loss              TEST        Sex   \n",
       "17           Best Kolmogorov-Smirnov along ROC              TEST        Sex   \n",
       "18                Kolmogorov-Smirnov at Cutoff              TEST        Sex   \n",
       "19                                        Gain              TEST        Sex   \n",
       "20                                        Lift              TEST        Sex   \n",
       "21                        %% Captured Response              TEST        Sex   \n",
       "22             Cumulative %% Captured Response              TEST        Sex   \n",
       "23                             Cumulative Lift              TEST        Sex   \n",
       "24                Average Prediction for Event              TEST        Sex   \n",
       "25                 Proportion into Event Level              TEST        Sex   \n",
       "\n",
       "             maxdiff  \n",
       "0   0.20221505042769  \n",
       "1   0.20221505042769  \n",
       "2   0.11519547070005  \n",
       "3   0.04129471251588  \n",
       "4   0.04129471251588  \n",
       "5   0.11519547070005  \n",
       "6    0.0858839713235  \n",
       "7   0.12868419444391  \n",
       "8   0.04113329073887  \n",
       "9   0.12037369095155  \n",
       "10  0.08226658147774  \n",
       "11  0.12868419444391  \n",
       "12  0.10725465336337  \n",
       "13  0.12868419444391  \n",
       "14  0.08117390649308  \n",
       "15   0.1253917503238  \n",
       "16  0.22059451243813  \n",
       "17  0.12528871354234  \n",
       "18  0.07390075818417  \n",
       "19  2.71511983391718  \n",
       "20  1.97176451211151  \n",
       "21  9.85882256055754  \n",
       "22  27.1511983391718  \n",
       "23  2.71511983391718  \n",
       "24  0.20221505042769  \n",
       "25  0.12773118041494  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------\n",
      "Bias Metrics : \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<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>Metric</th>\n",
       "      <th>MetricLabel</th>\n",
       "      <th>Value</th>\n",
       "      <th>Base</th>\n",
       "      <th>Compare</th>\n",
       "      <th>Note</th>\n",
       "      <th>_VARIABLE_</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DemographicParity</td>\n",
       "      <td>Demographic Parity (Statistical Parity)</td>\n",
       "      <td>0.127731</td>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td></td>\n",
       "      <td>Sex</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PredictiveParity</td>\n",
       "      <td>Predictive Parity</td>\n",
       "      <td>0.202215</td>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td></td>\n",
       "      <td>Sex</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>EqualAccuracy</td>\n",
       "      <td>Equal Accuracy</td>\n",
       "      <td>0.128684</td>\n",
       "      <td>Female</td>\n",
       "      <td>Male</td>\n",
       "      <td></td>\n",
       "      <td>Sex</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>EqualizedOdds</td>\n",
       "      <td>Equalized Odds</td>\n",
       "      <td>0.115195</td>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td>The maximum TPR difference is greater than the...</td>\n",
       "      <td>Sex</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>EqualOpportunity</td>\n",
       "      <td>Equal Opportunity</td>\n",
       "      <td>0.115195</td>\n",
       "      <td>Male</td>\n",
       "      <td>Female</td>\n",
       "      <td></td>\n",
       "      <td>Sex</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Metric                              MetricLabel     Value  \\\n",
       "0  DemographicParity  Demographic Parity (Statistical Parity)  0.127731   \n",
       "1   PredictiveParity                        Predictive Parity  0.202215   \n",
       "2      EqualAccuracy                           Equal Accuracy  0.128684   \n",
       "3      EqualizedOdds                           Equalized Odds  0.115195   \n",
       "4   EqualOpportunity                        Equal Opportunity  0.115195   \n",
       "\n",
       "     Base Compare                                               Note  \\\n",
       "0    Male  Female                                                      \n",
       "1    Male  Female                                                      \n",
       "2  Female    Male                                                      \n",
       "3    Male  Female  The maximum TPR difference is greater than the...   \n",
       "4    Male  Female                                                      \n",
       "\n",
       "  _VARIABLE_  \n",
       "0        Sex  \n",
       "1        Sex  \n",
       "2        Sex  \n",
       "3        Sex  \n",
       "4        Sex  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"Group Metrics: \")\n",
    "display(dfs['groupMetricsData'])\n",
    "print(\"-------------------------------------------\")\n",
    "print(\"Max Differences: \")\n",
    "display(dfs['maxDifferencesData'])\n",
    "print(\"-------------------------------------------\")\n",
    "print(\"Bias Metrics : \")\n",
    "display(dfs['biasMetricsData'])\n",
    "\n",
    "print(\"-------------------------------------------\")\n",
    "y_axis =np.array([\"Male\", \"Female\"])\n",
    "x_axis = np.array([(dfs['groupMetricsData']['_acc_'][0]), (dfs['groupMetricsData']['_acc_'][1])])\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "p = ax.bar(x = y_axis, height = x_axis)\n",
    "ax.set_title('Model Accuracy')\n",
    "plt.show()\n",
    "\n",
    "print(\"-------------------------------------------\")\n",
    "y_axis =np.array([\"Male\", \"Female\"])\n",
    "x_axis = np.array([(dfs['groupMetricsData']['PREDICTED_EVENT'][0]), (dfs['groupMetricsData']['PREDICTED_EVENT'][1])])\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "p = ax.bar(x = y_axis, height = x_axis)\n",
    "ax.set_title('Predicted Probability for Salary > $50K')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our model card is built using a few files new to sasctl and requires that data reside in CAS for SAS Information Catalog, which this new function will handle."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 12: Generate Model Card Function \n",
    "train_predict = model_object.predict(x_train)\n",
    "\n",
    "pzmm.JSONFiles.generate_model_card(\n",
    "    model_prefix = model_prefix,\n",
    "    model_files = output_path,\n",
    "    algorithm = algorithm,\n",
    "    train_data = x_train_full[original_columns], # changed to x_train_full with original variables (non-OHE'd)\n",
    "    train_predictions = train_scored[ti],\n",
    "    target_type='classification',\n",
    "    target_value=1,\n",
    "    class_vars=cat_columns, \n",
    "    interval_vars=interval_columns\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To document the requirements for our score code, lets create a requirements file. Since scikit-learn is installed as scikit-learn but imported as sklearn, we're going to tweak the requirements file to make sure we are referencing scikit-learn correctly. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[\n",
      "    {\n",
      "        \"command\": \"pip install numpy==1.24.3\",\n",
      "        \"step\": \"install numpy\"\n",
      "    },\n",
      "    {\n",
      "        \"command\": \"pip install pandas==2.1.4\",\n",
      "        \"step\": \"install pandas\"\n",
      "    },\n",
      "    {\n",
      "        \"command\": \"pip install sklearn==1.3.2\",\n",
      "        \"step\": \"install sklearn\"\n",
      "    }\n",
      "]\n",
      "[\n",
      "    {\n",
      "        \"command\": \"pip install numpy==1.24.3\",\n",
      "        \"step\": \"install numpy\"\n",
      "    },\n",
      "    {\n",
      "        \"command\": \"pip install pandas==2.1.4\",\n",
      "        \"step\": \"install pandas\"\n",
      "    },\n",
      "    {\n",
      "        \"command\": \"pip install scikit-learn==1.3.2\",\n",
      "        \"step\": \"install scikit-learn\"\n",
      "    }\n",
      "]\n"
     ]
    }
   ],
   "source": [
    "# Step 13: Generate requirements files\n",
    "requirements_json = pzmm.JSONFiles.create_requirements_json(output_path)\n",
    "\n",
    "import json\n",
    "print(json.dumps(requirements_json, sort_keys=True, indent=4))\n",
    "\n",
    "for requirement in requirements_json:\n",
    "    if 'sklearn' in requirement['step']:\n",
    "        requirement['command'] = requirement[\"command\"].replace('sklearn', 'scikit-learn')\n",
    "        requirement['step'] = requirement['step'].replace('sklearn', 'scikit-learn')\n",
    "\n",
    "print(json.dumps(requirements_json, sort_keys=True, indent=4))\n",
    "\n",
    "with open(Path(output_path) / \"requirements.json\", \"w\") as req_file:\n",
    "    req_file.write(json.dumps(requirements_json, indent=4))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Alright, we should have all our files now, so let's import our model and its files into SAS Model Manager! With the following function, we can automatically generate the score code and import the model into SAS Model Manager in one step."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 14: Automatically generate score code and import model \n",
    "pzmm.ImportModel.import_model(\n",
    "        model_files=output_path, # Where are the model files?\n",
    "        model_prefix=model_prefix, # What is the model name?\n",
    "        project=mm_project, # What is the project name?\n",
    "        input_data=x_train_full[['Age', 'WorkClass', 'Education', 'MartialStatus', \n",
    "       'Relationship', 'Race', 'Sex', 'HoursPerWeek']], # What does example input data look like?\n",
    "        predict_method=[dtc.predict_proba, [int, int]], # What is the predict method and what does it return?\n",
    "        score_metrics=[\"EM_CLASSIFICATION\", \"EM_EVENTPROBABILITY\"], # What are the output variables?\n",
    "        overwrite_model=True, # Overwrite the model if it already exists?\n",
    "        target_values=[\"0\", \"1\"], # What are the expected values of the target variable?\n",
    "        target_index=1, # What is the index of the target value in target_values?\n",
    "        model_file_name=model_prefix + \".pickle\", # How was the model file serialized?\n",
    "        missing_values=False, # Does the data include missing values?\n",
    "        preprocess_function=preprocess_function # What do we want to do to the data before we score it?\n",
    "    )\n",
    "    # Reinitialize the score_code variable when writing more than one model's score code\n",
    "pzmm.ScoreCode.score_code = \"\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And now your model is available in SAS Model Manager where you can run the scoring test, update the model usage properties, add Key Performance Indicators (KPIs) thresholds, and run performance monitoring against the model. Managing the rest of the model lifecycle in SAS Model Manager will complete the rest of the model card."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***\n",
    "## Conclusion\n",
    "Now you should have the tools you need to create a model card for Python classification and prediction models! \n",
    "***"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}