results_plots.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mel_dir1 = \"/Users/rachelbittner/Desktop/exper_output/ver1/CC_exper_mel1\"\n",
    "mel_dir2 = \"/Users/rachelbittner/Desktop/exper_output/ver1/CC_exper_mel2\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import glob\n",
    "import os\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import seaborn as sns\n",
    "sns.set()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "olap_vals = np.arange(0, 1, 0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "meltype=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0.0</th>\n",
       "      <th>0.1</th>\n",
       "      <th>0.2</th>\n",
       "      <th>0.3</th>\n",
       "      <th>0.4</th>\n",
       "      <th>0.5</th>\n",
       "      <th>0.6</th>\n",
       "      <th>0.7</th>\n",
       "      <th>0.8</th>\n",
       "      <th>0.9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>VR</th>\n",
       "      <td>0.846552</td>\n",
       "      <td>0.864278</td>\n",
       "      <td>0.854450</td>\n",
       "      <td>0.846683</td>\n",
       "      <td>0.826180</td>\n",
       "      <td>0.841368</td>\n",
       "      <td>0.843426</td>\n",
       "      <td>0.848185</td>\n",
       "      <td>0.853018</td>\n",
       "      <td>0.891252</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VFA</th>\n",
       "      <td>0.534865</td>\n",
       "      <td>0.552351</td>\n",
       "      <td>0.542541</td>\n",
       "      <td>0.530622</td>\n",
       "      <td>0.501069</td>\n",
       "      <td>0.516117</td>\n",
       "      <td>0.526405</td>\n",
       "      <td>0.525216</td>\n",
       "      <td>0.536557</td>\n",
       "      <td>0.601873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RPA</th>\n",
       "      <td>0.538677</td>\n",
       "      <td>0.547301</td>\n",
       "      <td>0.541059</td>\n",
       "      <td>0.537255</td>\n",
       "      <td>0.527694</td>\n",
       "      <td>0.533120</td>\n",
       "      <td>0.534933</td>\n",
       "      <td>0.531150</td>\n",
       "      <td>0.524213</td>\n",
       "      <td>0.527995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RCA</th>\n",
       "      <td>0.656886</td>\n",
       "      <td>0.669491</td>\n",
       "      <td>0.661655</td>\n",
       "      <td>0.655415</td>\n",
       "      <td>0.646847</td>\n",
       "      <td>0.654268</td>\n",
       "      <td>0.657071</td>\n",
       "      <td>0.658486</td>\n",
       "      <td>0.656670</td>\n",
       "      <td>0.667304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OA</th>\n",
       "      <td>0.476696</td>\n",
       "      <td>0.475071</td>\n",
       "      <td>0.474770</td>\n",
       "      <td>0.477531</td>\n",
       "      <td>0.488738</td>\n",
       "      <td>0.486999</td>\n",
       "      <td>0.482549</td>\n",
       "      <td>0.480366</td>\n",
       "      <td>0.470254</td>\n",
       "      <td>0.443489</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0.0       0.1       0.2       0.3       0.4       0.5       0.6  \\\n",
       "VR   0.846552  0.864278  0.854450  0.846683  0.826180  0.841368  0.843426   \n",
       "VFA  0.534865  0.552351  0.542541  0.530622  0.501069  0.516117  0.526405   \n",
       "RPA  0.538677  0.547301  0.541059  0.537255  0.527694  0.533120  0.534933   \n",
       "RCA  0.656886  0.669491  0.661655  0.655415  0.646847  0.654268  0.657071   \n",
       "OA   0.476696  0.475071  0.474770  0.477531  0.488738  0.486999  0.482549   \n",
       "\n",
       "          0.7       0.8       0.9  \n",
       "VR   0.848185  0.853018  0.891252  \n",
       "VFA  0.525216  0.536557  0.601873  \n",
       "RPA  0.531150  0.524213  0.527995  \n",
       "RCA  0.658486  0.656670  0.667304  \n",
       "OA   0.480366  0.470254  0.443489  "
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "if meltype == 1:\n",
    "    mel_dir = mel_dir1\n",
    "else:\n",
    "    mel_dir = mel_dir2\n",
    "    \n",
    "all_cols = ['clf mean', 'clf std', 'melscore mean', 'melscore std',\n",
    "            'melodiness mean', 'melodiness std']\n",
    "\n",
    "olap_str = [str(olap) for olap in olap_vals]\n",
    "\n",
    "clf_dict_avg = {k: {} for k in olap_str}\n",
    "melscore_dict_avg = {k: {} for k in olap_str}\n",
    "melodiness_dict_avg = {k: {} for k in olap_str}\n",
    "\n",
    "clf_dict_std = {k: {} for k in olap_str}\n",
    "melscore_dict_std = {k: {} for k in olap_str}\n",
    "melodiness_dict_std = {k: {} for k in olap_str}\n",
    "\n",
    "for i, olap in enumerate(olap_vals):\n",
    "    olap_dir = \"olap_%s\" % olap\n",
    "\n",
    "    clf_dat = []\n",
    "    melscore_dat = []\n",
    "    melodiness_dat = []\n",
    "    for split in range(1, 6):\n",
    "        split_dir = \"splitnum_%s\" % split\n",
    "        temp_path = os.path.join(mel_dir, split_dir, olap_dir)\n",
    "        \n",
    "        clf_score_fpath = os.path.join(temp_path, \"classifier_scores.csv\")\n",
    "        temp_df = pd.read_csv(clf_score_fpath, index_col=0)\n",
    "        \n",
    "        x = np.fromstring(temp_df.loc['f1', 'test'].strip('[').strip(']'), sep=' ', dtype='float')\n",
    "        temp_df.loc['f1 nonmel', 'test'] = x[0]\n",
    "        temp_df.loc['f1 mel', 'test'] = x[1]\n",
    "        temp_df.loc['f1 mean', 'test'] = np.mean(x)\n",
    "        clf_cols = ['accuracy', 'auc score', 'f1 nonmel', 'f1 mel', 'f1 mean']\n",
    "        clf_dat.append(temp_df.loc[clf_cols, :].convert_objects(convert_numeric=True))\n",
    "        \n",
    "        mel_score_fpath = os.path.join(temp_path, \"validate_mel_score_summary.csv\")\n",
    "        temp_df = pd.read_csv(mel_score_fpath, index_col=0)\n",
    "        melscore_dat.append(temp_df.iloc[1:3, :].convert_objects(convert_numeric=True))\n",
    "        \n",
    "        melodiness_fpath = os.path.join(temp_path, \"melodiness_scores.csv\")\n",
    "        temp_df = pd.read_csv(melodiness_fpath, index_col=0)\n",
    "        x = np.fromstring(temp_df.loc['f1', 'test'].strip('[').strip(']'), sep=' ', dtype='float')\n",
    "        temp_df.loc['f1 nonmel', 'test'] = x[0]\n",
    "        temp_df.loc['f1 mel', 'test'] = x[1]\n",
    "        temp_df.loc['f1 mean', 'test'] = np.mean(x)\n",
    "        melodiness_dat.append(temp_df.loc[clf_cols, :].convert_objects(convert_numeric=True))\n",
    "    \n",
    "    clf_dict_avg[str(olap)] = pd.concat(clf_dat).groupby(level=0).mean()\n",
    "    clf_dict_std[str(olap)] = pd.concat(clf_dat).groupby(level=0).std()\n",
    "\n",
    "    melscore_dict_avg[str(olap)] = pd.concat(melscore_dat).groupby(level=0).mean()\n",
    "    melscore_dict_std[str(olap)] = pd.concat(melscore_dat).groupby(level=0).std()\n",
    "    \n",
    "    melodiness_dict_avg[str(olap)] = pd.concat(melodiness_dat).groupby(level=0).mean()\n",
    "    melodiness_dict_std[str(olap)] = pd.concat(melodiness_dat).groupby(level=0).std()\n",
    "\n",
    "clf_panel_avg = pd.Panel(data=clf_dict_avg)\n",
    "clf_panel_std = pd.Panel(data=clf_dict_std)\n",
    "\n",
    "melscore_panel_avg = pd.Panel(data=melscore_dict_avg)\n",
    "melscore_panel_std = pd.Panel(data=melscore_dict_std)\n",
    "\n",
    "melodiness_panel_avg = pd.Panel(data=melodiness_dict_avg)\n",
    "melodiness_panel_std = pd.Panel(data=melodiness_dict_std)\n",
    "\n",
    "melscore_panel_avg.loc[:, 'mean', :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAToAAAH0CAYAAACtoOGYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nnRBC8yTohBCaJ0EnhNA8CTohhOb9f6epN7i4dB9GAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e43ba90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "c1, c2, c3, c4, c5 = sns.color_palette('deep', 5)\n",
    "\n",
    "xlims = [-0.1, 1.0]\n",
    "ylims = [0, 1]\n",
    "\n",
    "fig = plt.figure(figsize=(5,8))\n",
    "plt.subplot(3,1,1)\n",
    "p1 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'VR'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'VR'], label='Voicing Recall', \n",
    "                  linestyle='-', marker='o', color=c1, alpha=1, capsize=0)\n",
    "\n",
    "p2 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'VFA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'VFA'], label='Voicing False Alarm', \n",
    "                  linestyle='-', marker='o', color=c2, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=4, fontsize=12)\n",
    "plt.title(\"Melody %s Metrics vs. Overlap Threshold\" % meltype, fontsize=12)\n",
    "plt.tick_params(axis='x', which='both', bottom='off', top='off', labelbottom='off')\n",
    "\n",
    "plt.subplot(3,1,2)\n",
    "p3 = plt.errorbar(olap_vals+0.01, melscore_panel_avg.loc[:, 'mean', 'RPA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'RPA'], label='Raw Pitch Accuracy', \n",
    "                  linestyle='-', marker='o', color=c3, alpha=1, capsize=0)\n",
    "p4 = plt.errorbar(olap_vals-0.01, melscore_panel_avg.loc[:, 'mean', 'RCA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'RCA'], label='Raw Chroma Accuracy', \n",
    "                  linestyle='-', marker='o', color=c4, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=4, fontsize=12)\n",
    "plt.tick_params(axis='x', which='both', bottom='off', top='off', labelbottom='off')\n",
    "\n",
    "ax = plt.subplot(3,1,3)\n",
    "p5 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'OA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'OA'], label='Overall Accuracy', \n",
    "                  linestyle='-', marker='o', color=c5, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=4, fontsize=12)\n",
    "plt.xlabel('Overlap Threshold', fontsize=12)\n",
    "\n",
    "# plt.show()\n",
    "plt.savefig('/Users/rachelbittner/Desktop/mel%s_metrics_v_olap.eps' % meltype, format='eps')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAGoCAYAAABG05KeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10de5a310>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mel_dir = mel_dir1\n",
    "    \n",
    "all_cols = ['clf mean', 'clf std', 'melscore mean', 'melscore std',\n",
    "            'melodiness mean', 'melodiness std']\n",
    "\n",
    "olap_str = [str(olap) for olap in olap_vals]\n",
    "\n",
    "clf_dict_avg = {k: {} for k in olap_str}\n",
    "melscore_dict_avg = {k: {} for k in olap_str}\n",
    "melodiness_dict_avg = {k: {} for k in olap_str}\n",
    "\n",
    "clf_dict_std = {k: {} for k in olap_str}\n",
    "melscore_dict_std = {k: {} for k in olap_str}\n",
    "melodiness_dict_std = {k: {} for k in olap_str}\n",
    "\n",
    "for i, olap in enumerate(olap_vals):\n",
    "    olap_dir = \"olap_%s\" % olap\n",
    "\n",
    "    clf_dat = []\n",
    "    melscore_dat = []\n",
    "    melodiness_dat = []\n",
    "    for split in range(1, 6):\n",
    "        split_dir = \"splitnum_%s\" % split\n",
    "        temp_path = os.path.join(mel_dir, split_dir, olap_dir)\n",
    "        \n",
    "        clf_score_fpath = os.path.join(temp_path, \"classifier_scores.csv\")\n",
    "        temp_df = pd.read_csv(clf_score_fpath, index_col=0)\n",
    "        \n",
    "        x = np.fromstring(temp_df.loc['f1', 'test'].strip('[').strip(']'), sep=' ', dtype='float')\n",
    "        temp_df.loc['f1 nonmel', 'test'] = x[0]\n",
    "        temp_df.loc['f1 mel', 'test'] = x[1]\n",
    "        temp_df.loc['f1 mean', 'test'] = np.mean(x)\n",
    "        clf_cols = ['accuracy', 'auc score', 'f1 nonmel', 'f1 mel', 'f1 mean']\n",
    "        clf_dat.append(temp_df.loc[clf_cols, :].convert_objects(convert_numeric=True))\n",
    "        \n",
    "        mel_score_fpath = os.path.join(temp_path, \"validate_mel_score_summary.csv\")\n",
    "        temp_df = pd.read_csv(mel_score_fpath, index_col=0)\n",
    "        melscore_dat.append(temp_df.iloc[1:3, :].convert_objects(convert_numeric=True))\n",
    "        \n",
    "        melodiness_fpath = os.path.join(temp_path, \"melodiness_scores.csv\")\n",
    "        temp_df = pd.read_csv(melodiness_fpath, index_col=0)\n",
    "        x = np.fromstring(temp_df.loc['f1', 'test'].strip('[').strip(']'), sep=' ', dtype='float')\n",
    "        temp_df.loc['f1 nonmel', 'test'] = x[0]\n",
    "        temp_df.loc['f1 mel', 'test'] = x[1]\n",
    "        temp_df.loc['f1 mean', 'test'] = np.mean(x)\n",
    "        melodiness_dat.append(temp_df.loc[clf_cols, :].convert_objects(convert_numeric=True))\n",
    "    \n",
    "    clf_dict_avg[str(olap)] = pd.concat(clf_dat).groupby(level=0).mean()\n",
    "    clf_dict_std[str(olap)] = pd.concat(clf_dat).groupby(level=0).std()\n",
    "\n",
    "    melscore_dict_avg[str(olap)] = pd.concat(melscore_dat).groupby(level=0).mean()\n",
    "    melscore_dict_std[str(olap)] = pd.concat(melscore_dat).groupby(level=0).std()\n",
    "    \n",
    "    melodiness_dict_avg[str(olap)] = pd.concat(melodiness_dat).groupby(level=0).mean()\n",
    "    melodiness_dict_std[str(olap)] = pd.concat(melodiness_dat).groupby(level=0).std()\n",
    "\n",
    "clf_panel_avg = pd.Panel(data=clf_dict_avg)\n",
    "clf_panel_std = pd.Panel(data=clf_dict_std)\n",
    "\n",
    "melscore_panel_avg = pd.Panel(data=melscore_dict_avg)\n",
    "melscore_panel_std = pd.Panel(data=melscore_dict_std)\n",
    "\n",
    "melodiness_panel_avg = pd.Panel(data=melodiness_dict_avg)\n",
    "melodiness_panel_std = pd.Panel(data=melodiness_dict_std)\n",
    "\n",
    "melscore_panel_avg.loc[:, 'mean', :]\n",
    "\n",
    "c1, c2, c3, c4, c5 = sns.color_palette('deep', 5)\n",
    "\n",
    "xlims = [-0.1, 1.0]\n",
    "ylims = [0, 1]\n",
    "\n",
    "fig = plt.figure(figsize=(5.5,6))\n",
    "plt.subplot(3,2,1)\n",
    "p1 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'VR'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'VR'], label='Voicing Recall', \n",
    "                  linestyle='-', marker='o', color=c1, alpha=1, capsize=0)\n",
    "\n",
    "p2 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'VFA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'VFA'], label='Voicing False Alarm', \n",
    "                  linestyle='-', marker='o', color=c2, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=3, fontsize=11)\n",
    "plt.title(\"Melody 1\", fontsize=12)\n",
    "plt.tick_params(axis='x', which='both', bottom='off', top='off', labelbottom='off')\n",
    "\n",
    "plt.subplot(3,2,3)\n",
    "p3 = plt.errorbar(olap_vals+0.01, melscore_panel_avg.loc[:, 'mean', 'RPA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'RPA'], label='Raw Pitch Accuracy', \n",
    "                  linestyle='-', marker='o', color=c3, alpha=1, capsize=0)\n",
    "p4 = plt.errorbar(olap_vals-0.01, melscore_panel_avg.loc[:, 'mean', 'RCA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'RCA'], label='Raw Chroma Accuracy', \n",
    "                  linestyle='-', marker='o', color=c4, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=3, fontsize=11)\n",
    "plt.tick_params(axis='x', which='both', bottom='off', top='off', labelbottom='off')\n",
    "\n",
    "ax = plt.subplot(3,2,5)\n",
    "p5 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'OA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'OA'], label='Overall Accuracy', \n",
    "                  linestyle='-', marker='o', color=c5, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=3, fontsize=11)\n",
    "plt.xlabel(r'Overlap Threshold $\\alpha$', fontsize=12)\n",
    "\n",
    "############################\n",
    "mel_dir = mel_dir2\n",
    "    \n",
    "all_cols = ['clf mean', 'clf std', 'melscore mean', 'melscore std',\n",
    "            'melodiness mean', 'melodiness std']\n",
    "\n",
    "olap_str = [str(olap) for olap in olap_vals]\n",
    "\n",
    "clf_dict_avg = {k: {} for k in olap_str}\n",
    "melscore_dict_avg = {k: {} for k in olap_str}\n",
    "melodiness_dict_avg = {k: {} for k in olap_str}\n",
    "\n",
    "clf_dict_std = {k: {} for k in olap_str}\n",
    "melscore_dict_std = {k: {} for k in olap_str}\n",
    "melodiness_dict_std = {k: {} for k in olap_str}\n",
    "\n",
    "for i, olap in enumerate(olap_vals):\n",
    "    olap_dir = \"olap_%s\" % olap\n",
    "\n",
    "    clf_dat = []\n",
    "    melscore_dat = []\n",
    "    melodiness_dat = []\n",
    "    for split in range(1, 6):\n",
    "        split_dir = \"splitnum_%s\" % split\n",
    "        temp_path = os.path.join(mel_dir, split_dir, olap_dir)\n",
    "        \n",
    "        clf_score_fpath = os.path.join(temp_path, \"classifier_scores.csv\")\n",
    "        temp_df = pd.read_csv(clf_score_fpath, index_col=0)\n",
    "        \n",
    "        x = np.fromstring(temp_df.loc['f1', 'test'].strip('[').strip(']'), sep=' ', dtype='float')\n",
    "        temp_df.loc['f1 nonmel', 'test'] = x[0]\n",
    "        temp_df.loc['f1 mel', 'test'] = x[1]\n",
    "        temp_df.loc['f1 mean', 'test'] = np.mean(x)\n",
    "        clf_cols = ['accuracy', 'auc score', 'f1 nonmel', 'f1 mel', 'f1 mean']\n",
    "        clf_dat.append(temp_df.loc[clf_cols, :].convert_objects(convert_numeric=True))\n",
    "        \n",
    "        mel_score_fpath = os.path.join(temp_path, \"validate_mel_score_summary.csv\")\n",
    "        temp_df = pd.read_csv(mel_score_fpath, index_col=0)\n",
    "        melscore_dat.append(temp_df.iloc[1:3, :].convert_objects(convert_numeric=True))\n",
    "        \n",
    "        melodiness_fpath = os.path.join(temp_path, \"melodiness_scores.csv\")\n",
    "        temp_df = pd.read_csv(melodiness_fpath, index_col=0)\n",
    "        x = np.fromstring(temp_df.loc['f1', 'test'].strip('[').strip(']'), sep=' ', dtype='float')\n",
    "        temp_df.loc['f1 nonmel', 'test'] = x[0]\n",
    "        temp_df.loc['f1 mel', 'test'] = x[1]\n",
    "        temp_df.loc['f1 mean', 'test'] = np.mean(x)\n",
    "        melodiness_dat.append(temp_df.loc[clf_cols, :].convert_objects(convert_numeric=True))\n",
    "    \n",
    "    clf_dict_avg[str(olap)] = pd.concat(clf_dat).groupby(level=0).mean()\n",
    "    clf_dict_std[str(olap)] = pd.concat(clf_dat).groupby(level=0).std()\n",
    "\n",
    "    melscore_dict_avg[str(olap)] = pd.concat(melscore_dat).groupby(level=0).mean()\n",
    "    melscore_dict_std[str(olap)] = pd.concat(melscore_dat).groupby(level=0).std()\n",
    "    \n",
    "    melodiness_dict_avg[str(olap)] = pd.concat(melodiness_dat).groupby(level=0).mean()\n",
    "    melodiness_dict_std[str(olap)] = pd.concat(melodiness_dat).groupby(level=0).std()\n",
    "\n",
    "clf_panel_avg = pd.Panel(data=clf_dict_avg)\n",
    "clf_panel_std = pd.Panel(data=clf_dict_std)\n",
    "\n",
    "melscore_panel_avg = pd.Panel(data=melscore_dict_avg)\n",
    "melscore_panel_std = pd.Panel(data=melscore_dict_std)\n",
    "\n",
    "melodiness_panel_avg = pd.Panel(data=melodiness_dict_avg)\n",
    "melodiness_panel_std = pd.Panel(data=melodiness_dict_std)\n",
    "\n",
    "melscore_panel_avg.loc[:, 'mean', :]\n",
    "\n",
    "c1, c2, c3, c4, c5 = sns.color_palette('deep', 5)\n",
    "\n",
    "xlims = [-0.1, 1.0]\n",
    "ylims = [0, 1]\n",
    "\n",
    "plt.subplot(3,2,2)\n",
    "p1 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'VR'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'VR'], label='Voicing Recall', \n",
    "                  linestyle='-', marker='o', color=c1, alpha=1, capsize=0)\n",
    "\n",
    "p2 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'VFA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'VFA'], label='Voicing False Alarm', \n",
    "                  linestyle='-', marker='o', color=c2, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=3, fontsize=11)\n",
    "plt.title(\"Melody 2\", fontsize=12)\n",
    "plt.tick_params(axis='x', which='both', bottom='off', top='off', labelbottom='off')\n",
    "plt.tick_params(axis='y', which='both', bottom='off', top='off', labelleft='off' )\n",
    "\n",
    "plt.subplot(3,2,4)\n",
    "p3 = plt.errorbar(olap_vals+0.01, melscore_panel_avg.loc[:, 'mean', 'RPA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'RPA'], label='Raw Pitch Accuracy', \n",
    "                  linestyle='-', marker='o', color=c3, alpha=1, capsize=0)\n",
    "p4 = plt.errorbar(olap_vals-0.01, melscore_panel_avg.loc[:, 'mean', 'RCA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'RCA'], label='Raw Chroma Accuracy', \n",
    "                  linestyle='-', marker='o', color=c4, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=3, fontsize=11)\n",
    "plt.tick_params(axis='x', which='both', bottom='off', top='off', labelbottom='off')\n",
    "plt.tick_params(axis='y', which='both', bottom='off', top='off', labelleft='off' )\n",
    "\n",
    "ax = plt.subplot(3,2,6)\n",
    "p5 = plt.errorbar(olap_vals, melscore_panel_avg.loc[:, 'mean', 'OA'], \n",
    "                  melscore_panel_std.loc[:, 'mean', 'OA'], label='Overall Accuracy', \n",
    "                  linestyle='-', marker='o', color=c5, alpha=1, capsize=0)\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend(loc=3, fontsize=11)\n",
    "plt.xlabel(r'Overlap Threshold $\\alpha$', fontsize=12)\n",
    "plt.tick_params(axis='y', which='both', bottom='off', top='off', labelleft='off' )\n",
    "############################\n",
    "\n",
    "# plt.show()\n",
    "fig.tight_layout()\n",
    "plt.savefig('/Users/rachelbittner/Desktop/mel_metrics_v_olap.eps', format='eps', bbox_inches='tight', pad_inches=0.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x10d902ad0>"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAecAAAFtCAYAAAA9JqLqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10de5c750>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xlims = [-0.1, 1.0]\n",
    "ylims = [0, 1.0]\n",
    "\n",
    "fig = plt.figure()\n",
    "# plt.subplot(3,2,1)\n",
    "plt.errorbar(olap_vals + 0.005, clf_panel_avg.loc[:, 'f1 mel', 'test'], \n",
    "             clf_panel_std.loc[:, 'f1 mel', 'test'], label='f1 melody: Random Forest', \n",
    "             linestyle='-', marker='s', alpha=1)\n",
    "plt.errorbar(olap_vals+ 0.005, melodiness_panel_avg.loc[:, 'f1 mel', 'test'], \n",
    "             melodiness_panel_std.loc[:, 'f1 mel', 'test'], label='f1 melody: Melodiness', \n",
    "             linestyle='-', marker='o', alpha=1)\n",
    "# plt.errorbar(olap_vals + 0.005, clf_panel_avg.loc[:, 'auc score', 'test'], \n",
    "#              clf_panel_std.loc[:, 'auc score', 'test'], label='AUC ROC: Random Forest', \n",
    "#              linestyle='-', marker='s', alpha=1)\n",
    "# plt.errorbar(olap_vals+ 0.005, clf_panel_avg.loc[:, 'accuracy', 'test'], \n",
    "#              clf_panel_std.loc[:, 'accuracy', 'test'], label='Classifier Accuracy: Random Forest', \n",
    "#              linestyle='-', marker='o', alpha=1)\n",
    "# plt.errorbar(olap_vals + 0.005, melodiness_panel_avg.loc[:, 'auc score', 'test'], \n",
    "#              melodiness_panel_std.loc[:, 'auc score', 'test'], label='AUC ROC: Melodiness', \n",
    "#              linestyle='-', marker='s', alpha=1)\n",
    "# plt.errorbar(olap_vals+ 0.005, melodiness_panel_avg.loc[:, 'accuracy', 'test'], \n",
    "#              melodiness_panel_std.loc[:, 'accuracy', 'test'], label='Classifier Accuracy: Melodiness', \n",
    "#              linestyle='-', marker='o', alpha=1)\n",
    "\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend()\n",
    "plt.xlabel('Overlap Threshold')\n",
    "plt.title('Classifier Scores (Test Set)')\n",
    "# plt.show()\n",
    "# plt.savefig('/Users/rachelbittner/Desktop/clf_metrics_mel2.eps', format='eps')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAewAAAFtCAYAAADFxlkdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10f309dd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xlims = [-0.1, 1.0]\n",
    "ylims = [0.6, 1.0]\n",
    "\n",
    "fig = plt.figure()\n",
    "plt.errorbar(olap_vals + 0.005, melodiness_panel_avg.loc[:, 'auc score', 'test'], \n",
    "             melodiness_panel_std.loc[:, 'auc score', 'test'], label='Area under ROC curve', \n",
    "             linestyle='-', marker='s', alpha=1)\n",
    "plt.errorbar(olap_vals+ 0.005, melodiness_panel_avg.loc[:, 'accuracy', 'test'], \n",
    "             melodiness_panel_std.loc[:, 'accuracy', 'test'], label='Classifier Accuracy', \n",
    "             linestyle='-', marker='o', alpha=1)\n",
    "\n",
    "plt.xlim(xlims)\n",
    "plt.ylim(ylims)\n",
    "plt.legend()\n",
    "plt.xlabel('Overlap Threshold')\n",
    "plt.title('Melodiness Scores (Test Set)')\n",
    "plt.show()\n",
    "# plt.savefig('/Users/rachelbittner/Desktop/melodiness_metrics_mel1.eps', format='eps')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "clf_score_fpath1 = \"bestf_classifier_mel1.csv\"\n",
    "clf_df1 = pd.read_csv(clf_score_fpath1)\n",
    "melodiness_score_fpath1 = \"bestf_melodiness_mel1.csv\"\n",
    "mds_df1 = pd.read_csv(melodiness_score_fpath1)\n",
    "\n",
    "clf_score_fpath2 = \"bestf_classifier_mel2.csv\"\n",
    "clf_df2 = pd.read_csv(clf_score_fpath2)\n",
    "melodiness_score_fpath2 = \"bestf_melodiness_mel2.csv\"\n",
    "mds_df2 = pd.read_csv(melodiness_score_fpath2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.750</td>\n",
       "      <td>0.020000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.750</td>\n",
       "      <td>0.022361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.750</td>\n",
       "      <td>0.015811</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.748</td>\n",
       "      <td>0.016432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.748</td>\n",
       "      <td>0.019235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.744</td>\n",
       "      <td>0.016733</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.748</td>\n",
       "      <td>0.024900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.742</td>\n",
       "      <td>0.023875</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.734</td>\n",
       "      <td>0.023022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.714</td>\n",
       "      <td>0.021909</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean       std\n",
       "0  0.750  0.020000\n",
       "1  0.750  0.022361\n",
       "2  0.750  0.015811\n",
       "3  0.748  0.016432\n",
       "4  0.748  0.019235\n",
       "5  0.744  0.016733\n",
       "6  0.748  0.024900\n",
       "7  0.742  0.023875\n",
       "8  0.734  0.023022\n",
       "9  0.714  0.021909"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEBCAYAAABrF5JMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEEIIkRApGEIIIRLyF+5hqQtkr4BEAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112878c90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "c1, c2 = sns.color_palette('deep', 2)\n",
    "\n",
    "fig = plt.figure(figsize=(6,3.5))\n",
    "plt.subplot(1,2,1)\n",
    "plt.errorbar(olap_vals-0.01, clf_df1['mean'], clf_df1['std'], label='Random Forest', \n",
    "             linestyle='-', marker='o', color=c1, alpha=1, capsize=0)\n",
    "plt.errorbar(olap_vals+0.01, mds_df1['mean'], mds_df1['std'], label='Melodiness', \n",
    "             linestyle='-', marker='o', color=c2, alpha=1, capsize=0)\n",
    "plt.xlim([-0.1, 1])\n",
    "plt.ylim([0.63, 0.78])\n",
    "plt.legend(loc=3, fontsize=12)\n",
    "plt.xlabel(r'Overlap Threshold $\\alpha$', fontsize=12)\n",
    "plt.ylabel('F1 Score', fontsize=12)\n",
    "plt.title('Melody 1', fontsize=12)\n",
    "\n",
    "plt.subplot(1,2,2)\n",
    "plt.errorbar(olap_vals-0.01, clf_df2['mean'], clf_df2['std'], label='Random Forest', \n",
    "             linestyle='-', marker='o', color=c1, alpha=1, capsize=0)\n",
    "plt.errorbar(olap_vals+0.01, mds_df2['mean'], mds_df2['std'], label='Melodiness', \n",
    "             linestyle='-', marker='o', color=c2, alpha=1, capsize=0)\n",
    "plt.xlim([-0.1, 1])\n",
    "plt.ylim([0.63, 0.78])\n",
    "plt.legend(loc=3, fontsize=12)\n",
    "plt.xlabel(r'Overlap Threshold $\\alpha$', fontsize=12)\n",
    "plt.tick_params(axis='y', which='both', bottom='off', top='off', labelleft='off' )\n",
    "#plt.ylabel('F1 Score', fontsize=12)\n",
    "plt.title('Melody 2', fontsize=12)\n",
    "\n",
    "plt.savefig('/Users/rachelbittner/Desktop/clf_v_mds_f1.eps', format='eps')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[0.0, 0.0, 0.0],\n",
       " [0.0654042722902678, 0.17637446680033858, 0.08773649777777773],\n",
       " [0.15456460348327078, 0.34410299533078853, 0.19270830615916945],\n",
       " [0.2658737501614362, 0.4977976168883662, 0.3125473286889657],\n",
       " [0.3952112563745892, 0.6336627816021115, 0.4431984952249134],\n",
       " [0.5501852107235036, 0.7596905579845981, 0.5923471685044214],\n",
       " [0.7273079805575802, 0.8716844273942121, 0.7563630559323337],\n",
       " [0.9265795658768193, 0.9696443898309538, 0.9352461575086505]]"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sns.cubehelix_palette(8, start=2, rot=0, dark=0, light=.95, reverse=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "x = sns.diverging_palette(145, 280, s=85, l=25, n=7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['#12432b', '#5c8270', '#a9c2b6', '#f2f2f2', '#c2abd7', '#8a62ae', '#531b86']"
      ]
     },
     "execution_count": 214,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "['#%02x%02x%02x' % (s[0], s[1], s[2]) for s in 255*x]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0.071064747955066704, 0.26573340015411195, 0.17177438979976059),\n",
       " (0.3644102814050737, 0.51184576068995702, 0.44068437148712447),\n",
       " (0.6644227587971262, 0.76355158396525313, 0.71570594366738294),\n",
       " (0.94999999999999996, 0.94999999999999996, 0.94999999999999996),\n",
       " (0.76295149595245304, 0.6726194002713497, 0.84602388211117907),\n",
       " (0.5427050946785329, 0.38706232814016617, 0.68583933624791582),\n",
       " (0.32735305787736652, 0.10785096872300895, 0.52921444695939179)]"
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[(s[0], s[1], s[2]) for s in x]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "olap_vals = np.arange(0, 1, 0.1)\n",
    "mel1_alpha_mean = [0.104,\n",
    "0.088,\n",
    "0.086,\n",
    "0.092,\n",
    "0.102,\n",
    "0.092,\n",
    "0.078,\n",
    "0.072,\n",
    "0.054,\n",
    "0.024]\n",
    "mel1_alpha_std = [0.04079215611,\n",
    "0.02315167381,\n",
    "0.02727636339,\n",
    "0.02925747768,\n",
    "0.02481934729,\n",
    "0.03544009029,\n",
    "0.02481934729,\n",
    "0.02925747768,\n",
    "0.02332380758,\n",
    "0.01356465997]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFVCAYAAADCLbfjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10d559710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.errorbar(olap_vals, mel1_alpha_mean, \n",
    "                  mel1_alpha_std, label='Melody 1', \n",
    "                  linestyle='-', marker='o', alpha=1, capsize=0)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}