{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SHAP-Style Plots\n",
    "Beyond waterfall plots, SHAP values can be visualized as summary (beeswarm) plots and dependence plots."
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "import skexplain\n",
    "from skexplain.common.importance_utils import to_skexplain_importance\n",
    "import plotting_config\n",
    "import numpy as np\n",
    "import shap"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "# Load data and models\n",
    "estimators = skexplain.load_models()\n",
    "X, y = skexplain.load_data()\n",
    "\n",
    "# Create a random subset of 1000 examples\n",
    "random_state = np.random.RandomState(42)\n",
    "N = 1000\n",
    "ind = random_state.choice(len(X), size=N, replace=False)\n",
    "X_subset = X.iloc[ind]\n",
    "y_subset = y[ind]\n",
    "X_subset.reset_index(inplace=True, drop=True)"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "explainer = skexplain.ExplainToolkit(estimators[0], X=X_subset)\n",
    "explainer.set_plotting_config(\n",
    "    display_feature_names=plotting_config.display_feature_names,\n",
    "    display_units=plotting_config.display_units,\n",
    "    feature_colors=plotting_config.color_dict,\n",
    ")"
   ],
   "outputs": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Computing SHAP Values"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "shap_kws = {\n",
    "    'masker': shap.maskers.Partition(X, max_samples=100, clustering='correlation'),\n",
    "    'algorithm': 'permutation',\n",
    "}\n",
    "\n",
    "# Compute SHAP values for the subset\n",
    "# To save time, you can load pre-computed results:\n",
    "#   results = explainer.load('../tutorial_data/attr_values.nc')\n",
    "results = explainer.local_attributions(\n",
    "    method='shap',\n",
    "    shap_kws=shap_kws,\n",
    "    n_jobs=8,\n",
    ")"
   ],
   "outputs": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary (Beeswarm) Plot"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "explainer.scatter_plot(\n",
    "    plot_type='summary',\n",
    "    dataset=results,\n",
    "    method='shap',\n",
    "    estimator_name='Random Forest',\n",
    ")"
   ],
   "outputs": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dependence Plots\n",
    "Show how SHAP values relate to feature values."
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "explainer.scatter_plot(\n",
    "    features=['sfc_temp', 'temp2m', 'dwpt2m', 'sat_irbt'],\n",
    "    plot_type='dependence',\n",
    "    dataset=results,\n",
    "    method=['shap'],\n",
    "    estimator_name='Random Forest',\n",
    "    interaction_index=None,\n",
    ")"
   ],
   "outputs": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Color-Coded Dependence\n",
    "Color by interaction feature to reveal interactions."
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "explainer.scatter_plot(\n",
    "    features=['sfc_temp', 'temp2m', 'dwpt2m', 'sat_irbt'],\n",
    "    plot_type='dependence',\n",
    "    dataset=results,\n",
    "    method=['shap'],\n",
    "    estimator_name='Random Forest',\n",
    "    interaction_index='auto',\n",
    "    colorbar_pad=0.2,\n",
    "    wspace=0.7,\n",
    "    hspace=0.6,\n",
    ")"
   ],
   "outputs": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Converting Attributions to Importance Rankings"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "source": [
    "# Convert SHAP values to importance scores for the bar-style ranking plot\n",
    "shap_rank = to_skexplain_importance(\n",
    "    results['shap_values__Random Forest'].values,\n",
    "    estimator_name='Random Forest',\n",
    "    feature_names=X.columns,\n",
    "    method='shap_sum',\n",
    ")\n",
    "\n",
    "explainer.plot_importance(\n",
    "    data=[shap_rank],\n",
    "    panels=[('shap_sum', 'Random Forest')],\n",
    "    xlabels=['SHAP'],\n",
    ")"
   ],
   "outputs": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python",
   "version": "3.8.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}