aggregate_results.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analysis\n",
    "Collect, aggregate and evaluate causal links. Produces causal (correlation) matrix plots."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Main program"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Specifications"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import sys, yaml, numpy as np\n",
    "from os.path import exists\n",
    "\n",
    "from utils.setup import SetupPCMCIAggregation\n",
    "import utils.pcmci_aggregation as aggregation\n",
    "\n",
    "argv = ['-c', './nn_config/cfg_aggregate_results.yml']\n",
    "# argv = ['-c', './nn_config/cfg_aggregate_results_pearsonr.yml']\n",
    "\n",
    "setup = SetupPCMCIAggregation(argv)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Options"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "threshold_dict = False\n",
    "num_parents    = True\n",
    "insets         = False\n",
    "masking        = False\n",
    "save           = False"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Collect results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Variable: tphystnd\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-3.64_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-7.59_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-14.36_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-24.61_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-38.27_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-54.6_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-72.01_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-87.82_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-103.32_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-121.55_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-142.99_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-168.23_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-197.91_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-232.83_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-273.91_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-322.24_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-379.1_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-445.99_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-524.69_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-609.78_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-691.39_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-763.4_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-820.86_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-859.53_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-887.02_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-912.64_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-936.2_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-957.49_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-976.33_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/tphystnd-992.56_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "Variable: phq\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-3.64_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-7.59_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-14.36_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-24.61_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-38.27_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-54.6_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-72.01_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-87.82_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-103.32_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-121.55_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-142.99_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-168.23_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-197.91_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-232.83_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-273.91_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-322.24_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-379.1_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-445.99_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-524.69_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-609.78_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-691.39_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-763.4_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-820.86_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-859.53_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-887.02_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-912.64_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-936.2_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-957.49_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-976.33_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/phq-992.56_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "Variable: fsnt\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/fsnt_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "Variable: fsns\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/fsns_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "Variable: flnt\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/flnt_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "Variable: flns\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/flns_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "Variable: prect\n",
      "SPCAM_Rasp_causal_links_3mth_v2/aggregate_pdf/prect_parcorr_2021_09_02_TRAIN_For_Nando exists; loading...\n",
      "\n",
      "No errors were found\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# collected_results, errors = aggregation.collect_results(setup)\n",
    "collected_results, errors = aggregation.collect_results(setup, reuse=True)\n",
    "print(aggregation.print_errors(errors))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Analyze collected results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Caution! For threshold (0.59): phq-379.1 not normaly distributed\n"
     ]
    }
   ],
   "source": [
    "setup.thrs_argv = ['-c', './nn_config/quantile_dicts/001_quantile_thresholds_mse-6.yml']\n",
    "\n",
    "aggregated_results, var_names_parents = aggregation.aggregate_results(\n",
    "    collected_results, setup, threshold_dict=threshold_dict)\n",
    "# aggregation.print_aggregated_results(var_names_parents, aggregated_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calc_mean_num_parents(aggregated_results, save_numparents_dict=False):\n",
    "    numparents = []\n",
    "    numparents_dict = {}; numparents_dict['numparents_dict'] = {}\n",
    "    tot_inputs = len(var_names_parents)\n",
    "    for i, output in enumerate(aggregated_results.keys()):\n",
    "        threshold   = [setup.thresholds[0],'optimized'][threshold_dict==True]\n",
    "        num_parents = aggregated_results[output][str(setup.pc_alphas[0])]['num_parents'][str(threshold)]\n",
    "        numparents.append(num_parents)\n",
    "        numparents_dict['numparents_dict'][output] = num_parents\n",
    "        # print(f\"{i} {output} num. parents: {num_parents}\")\n",
    "    numparents = np.array(numparents)\n",
    "    print(f\"Mean Num. of Parents is: {numparents.mean()} ({numparents.mean()/tot_inputs * 100.} %)\\n\")\n",
    "    \n",
    "    if save_numparents_dict:\n",
    "        # Optmized threshold dictionary: Save?\n",
    "        if not exists(save_numparents_dict):\n",
    "            print(f\"Create {save_numparents_dict} dictionary.\")\n",
    "            with open(save_numparents_dict, 'w') as save_numparents_dict:\n",
    "                yaml.dump(numparents_dict, save_numparents_dict, default_flow_style=False)\n",
    "        else:\n",
    "            print(f\"{save_numparents_dict} dictionary exists.\")\n",
    "    \n",
    "    return numparents, numparents_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plots"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot Outputs-Inputs metrics "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "aggregation.plot_matrix_results(\n",
    "    var_names_parents, \n",
    "    aggregated_results, \n",
    "    setup, \n",
    "    values='percentage',\n",
    "    insets=insets,\n",
    "    threshold_dict=threshold_dict,\n",
    "    num_parents=num_parents,\n",
    "    save=save,\n",
    "    masking=masking,\n",
    "    cmap='Reds',\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Num. of Parents is: 36.353846153846156 (38.67430441898527 %)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "save_numparents_dict = False\n",
    "# save_numparents_dict = f\"./nn_config/quantile_dicts/001_quantile-{setup.thresholds[0]}_numparents.yml\"\n",
    "\n",
    "numparents, numparents_dict = calc_mean_num_parents(aggregated_results, save_numparents_dict=save_numparents_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "causalnncam",
   "language": "python",
   "name": "causalnncam"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}