visualize_results.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import os,sys\n",
    "import pickle\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "# add project root dir to sys.path so that all packages can be found by python.\n",
    "root_dir = os.path.dirname(os.path.realpath(\"__file__\"))\n",
    "sys.path.append(root_dir)\n",
    "from data.musdb_loader import get_musdb_folds\n",
    "from data.eval import create_method_store\n",
    "from data.dataset import RawSeparationDataset\n",
    "from museval.aggregate import MethodStore"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# DO NOT EXECUTE UNLESS YOU HAVE BASELINE AND ORACLE ESTIMATES IN JSON FILES!\n",
    "def get_method_store(dataset, track_names=None):\n",
    "    db_path = os.path.join(root_dir,'data', dataset)\n",
    "    if (track_names is None):\n",
    "        data = get_musdb_folds(db_path)\n",
    "        path_list = data['test']\n",
    "        track_names = list(map(lambda target_dict: os.path.basename(os.path.dirname(target_dict['mix'])), path_list))\n",
    "    store = create_method_store(os.path.join(db_path,'estimates'), ['baseline', 'ibm','irm'], 'test')\n",
    "    store.df = store.df[store.df.track.isin(track_names)]\n",
    "    return store   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading train set...\n",
      "Loading test set...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Creating eval store: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3/3 [00:05<00:00,  1.85s/it]\n"
     ]
    }
   ],
   "source": [
    "# DO NOT EXECUTE UNLESS YOU HAVE BASELINE AND ORACLE ESTIMATES IN JSON FILES!\n",
    "musdb_store = get_method_store(\"musdb\")\n",
    "musdb_store.save(\"musdb_store\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Creating eval store: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3/3 [00:15<00:00,  5.04s/it]\n"
     ]
    }
   ],
   "source": [
    "# DO NOT EXECUTE UNLESS YOU HAVE BASELINE AND ORACLE ESTIMATES IN JSON FILES!\n",
    "musdb_extended_store = get_method_store(\"musdb_extended\", list(map(str,list(range(736,921)))))\n",
    "musdb_extended_store.save(\"musdb_ext_store\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def _filter_df_col(df, col_name, col_values):\n",
    "    if (col_values is not None):\n",
    "        col_values = [col_values] if isinstance(col_values, str) else col_values\n",
    "        df = df[df[col_name].isin(col_values)]\n",
    "    if (len(df[col_name].unique()) == 1):\n",
    "        df = df.drop(col_name, axis=1)\n",
    "    return df\n",
    "\n",
    "def get_method_metric_for_target(method_store, method, metric, target):\n",
    "    # method = baseline, ibm, irm\n",
    "    # metric = SDR, ISR, SIR, SAR\n",
    "    # target = vocals, accompaniment, drums...\n",
    "    \n",
    "    # Get median scores for each track\n",
    "    df = method_store.agg_frames_scores().reset_index(name='score').dropna()\n",
    "\n",
    "    # Get lists of scores per method-metric-target\n",
    "    df = df.groupby(['method','metric','target'])['score'].apply(list).reset_index(name='score')\n",
    "\n",
    "    # filter by provided column values\n",
    "    df = _filter_df_col( _filter_df_col( _filter_df_col(df, 'method', method),'metric', metric), 'target', target)\n",
    "    return df['score'].values[0]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# This loads the metrics store from the notebook kernel and allows to get metrics for any method, metric and target.\n",
    "musdb_store = MethodStore()\n",
    "musdb_store.load(\"musdb_store\")\n",
    "vals = get_method_metric_for_target(musdb_store, 'ibm', 'SDR', 'vocals')\n",
    "fig, ax = plt.subplots()\n",
    "ax.violinplot(vals,showmeans=True, showextrema=True, showmedians=True);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_metrics(filepath, instruments=[\"vocals\", \"accompaniment\"]):\n",
    "    with open(filepath, \"rb\") as f:\n",
    "        test_metrics = pickle.load(f)\n",
    "    print(test_metrics)\n",
    "    SDRs = {inst: [np.nanmean(song[inst][\"SDR\"]) for song in test_metrics] for inst in instruments}\n",
    "    SIRs = {inst: [np.nanmean(song[inst][\"SIR\"]) for song in test_metrics] for inst in instruments}\n",
    "    avg_SDRs = {inst: np.mean(SDRs[inst]) for inst in instruments}\n",
    "    avg_SIRs = {inst: np.mean(SIRs[inst]) for inst in instruments}\n",
    "    return SDRs, SIRs, avg_SDRs, avg_SIRs\n",
    "\n",
    "def plot_metric(metric, metric_name, instruments=[\"vocals\", \"accompaniment\"], axs=None):\n",
    "    if not axs:\n",
    "        fig, axs = plt.subplots(nrows=len(instruments), ncols=1, figsize=(10, 6))\n",
    "    for i, instrument in enumerate(instruments):\n",
    "        axs[i].violinplot(metric[instrument],\n",
    "                     showmeans=True, showextrema=True, showmedians=True)\n",
    "        axs[i].set_title(f\"{metric_name} for {instrument}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "SDRs, SIRs, avg_SDRs, avg_SIRs = get_metrics(\"results.pkl\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "avg_SDRs={'vocals': 2.7900325247439413, 'accompaniment': 7.700905718452066}, avg_SIRs={'vocals': 8.39428707334839, 'accompaniment': 18.718657266971196}\n"
     ]
    }
   ],
   "source": [
    "print(f\"{avg_SDRs=}, {avg_SIRs=}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_metric(SDRs, \"SDR\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_metric(SIRs, \"SIR\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}