Harmonixset parsing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "8d19f0b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import jams\n",
    "from tqdm import tqdm\n",
    "import mir_eval\n",
    "from glob import glob\n",
    "from pathlib import Path\n",
    "import warnings\n",
    "import numpy as np\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "39d4823b",
   "metadata": {},
   "outputs": [],
   "source": [
    "jamses = jams.util.find_with_extension('/home/bmcfee/data/harmonixset/dataset/jams/', 'jams')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e7215256",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 912/912 [00:06<00:00, 141.69it/s]\n"
     ]
    }
   ],
   "source": [
    "unique_labels = set()\n",
    "for fn in tqdm(jamses):\n",
    "    j = jams.load(fn, validate=False)\n",
    "    # get the segment annotation\n",
    "    ann = j.annotations['segment', 0]\n",
    "    \n",
    "    # get the intervals and labels\n",
    "    intervals, labels = ann.to_interval_values()\n",
    "    unique_labels |= set(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "acf38c83",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'altchorus',\n",
       " 'bigoutro',\n",
       " 'bre',\n",
       " 'break',\n",
       " 'break1',\n",
       " 'break2',\n",
       " 'break3',\n",
       " 'breakdown',\n",
       " 'breakdown2',\n",
       " 'bridge',\n",
       " 'bridge1',\n",
       " 'bridge2',\n",
       " 'bridge3',\n",
       " 'build',\n",
       " 'chorus',\n",
       " 'chorus1',\n",
       " 'chorus2',\n",
       " 'chorus3',\n",
       " 'chorus_instrumental',\n",
       " 'chorushalf',\n",
       " 'chorusinst',\n",
       " 'choruspart',\n",
       " 'drumroll',\n",
       " 'end',\n",
       " 'fadein',\n",
       " 'fast',\n",
       " 'gtr',\n",
       " 'gtr2',\n",
       " 'gtrbreak',\n",
       " 'guitar',\n",
       " 'guitarsolo',\n",
       " 'inst',\n",
       " 'inst2',\n",
       " 'instbridge',\n",
       " 'instchorus',\n",
       " 'instintro',\n",
       " 'instrumental',\n",
       " 'instrumental2',\n",
       " 'instrumental3',\n",
       " 'instrumentalverse',\n",
       " 'intchorus',\n",
       " 'intro',\n",
       " 'intro2',\n",
       " 'intro3',\n",
       " 'intro4',\n",
       " 'intro5',\n",
       " 'intro6',\n",
       " 'intro7',\n",
       " 'intro8',\n",
       " 'introchorus',\n",
       " 'intropt2',\n",
       " 'introverse',\n",
       " 'mainriff',\n",
       " 'mainriff2',\n",
       " 'miniverse',\n",
       " 'oddriff',\n",
       " 'opening',\n",
       " 'outro',\n",
       " 'outro1',\n",
       " 'outro2',\n",
       " 'outro3',\n",
       " 'outroa',\n",
       " 'postchorus',\n",
       " 'postchorus2',\n",
       " 'postverse',\n",
       " 'prechorus',\n",
       " 'prechorus2',\n",
       " 'prechorus3',\n",
       " 'prechorus5',\n",
       " 'preverse',\n",
       " 'quiet',\n",
       " 'quietchorus',\n",
       " 'raps',\n",
       " 'refrain',\n",
       " 'rhythmlessintro',\n",
       " 'saxobeat',\n",
       " 'section',\n",
       " 'section1',\n",
       " 'section10',\n",
       " 'section11',\n",
       " 'section12',\n",
       " 'section13',\n",
       " 'section14',\n",
       " 'section15',\n",
       " 'section16',\n",
       " 'section17',\n",
       " 'section2',\n",
       " 'section3',\n",
       " 'section4',\n",
       " 'section5',\n",
       " 'section6',\n",
       " 'section7',\n",
       " 'section8',\n",
       " 'section9',\n",
       " 'silence',\n",
       " 'slow',\n",
       " 'slow2',\n",
       " 'slowverse',\n",
       " 'solo',\n",
       " 'solo2',\n",
       " 'solo3',\n",
       " 'stutter',\n",
       " 'synth',\n",
       " 'transition',\n",
       " 'transition1',\n",
       " 'transition2',\n",
       " 'transition2a',\n",
       " 'transition3',\n",
       " 'verse',\n",
       " 'verse1',\n",
       " 'verse10',\n",
       " 'verse11',\n",
       " 'verse1a',\n",
       " 'verse2',\n",
       " 'verse3',\n",
       " 'verse4',\n",
       " 'verse5',\n",
       " 'verse6',\n",
       " 'verse7',\n",
       " 'verse8',\n",
       " 'verse9',\n",
       " 'verse_slow',\n",
       " 'verseinst',\n",
       " 'versepart',\n",
       " 'vocaloutro',\n",
       " 'worstthingever'}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unique_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "93544703",
   "metadata": {},
   "outputs": [],
   "source": [
    "MAPPING = {\n",
    "    'altchorus': 'chorus',\n",
    "    'bigoutro': 'outro',\n",
    "    r'(.*)\\d+.*': r'\\1',\n",
    "    '(.*)_instrumental': r'\\1',\n",
    "    'chorus.*': 'chorus',\n",
    "    'instbridge': 'bridge',\n",
    "    'instchorus': 'chorus',\n",
    "    'instintro': 'intro',\n",
    "    'instrumentalverse': 'verse',\n",
    "    'intropt2': 'intro',\n",
    "    'miniverse': 'verse',\n",
    "    'quietchorus': 'chorus',\n",
    "    'rhythmlessintro': 'intro',\n",
    "    'verse_slow': 'verse',\n",
    "    'verseinst': 'verse',\n",
    "    'versepart': 'verse',\n",
    "    'vocaloutro': 'outro'\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f4137ab8",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('../code/labelmapping.pickle', 'wb') as fd:\n",
    "    pickle.dump(MAPPING, fd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8d7ad578",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "lrr",
   "language": "python",
   "name": "lrr"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}