MAT 240E homework1 data collection

{
 "metadata": {
  "name": ""
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "############################################################\n",
      "## Based on code from Python online documention for os.walk \n",
      "## and the Essentia 'Python tutorial' \n",
      "############################################################\n",
      "\n",
      "import os\n",
      "import essentia\n",
      "import essentia.standard\n",
      "import essentia.streaming\n",
      "\n",
      "from os.path import join, getsize\n",
      "\n",
      "extensions = dict()\n",
      "durations = []\n",
      "with open('durations.txt', 'wb') as f:\n",
      "    for root, dirs, files in os.walk('/Users/ogc/Music/iTunes'):\n",
      "        if 'CVS' in dirs:\n",
      "            dirs.remove('CVS')  # don't visit CVS directories\n",
      "        #print sum(getsize(join(root, name)) for name in files),\n",
      "        print 'album: {0}\\n'.format(root)\n",
      "        for name in files:\n",
      "            #used this code to get a complete list of different audio formats in my music library\n",
      "            #extension = os.path.splitext(name)[1]\n",
      "            #if extension not in extensions:\n",
      "            #    extensions[extension] = 1\n",
      "            #else:\n",
      "            #    extensions[extension] += 1\n",
      "            if name.endswith(('.wav', '.aif', 'aiff', '.flac', '.mp3', '.m4a')):\n",
      "                #print root\n",
      "                # and then we actually perform the loading:\n",
      "                try:\n",
      "                    loader = essentia.standard.MonoLoader(filename = join(root, name))\n",
      "                    #print loader.paramValue('sr') # How can I check for the sampling rate of the file?\n",
      "                    audio = loader()\n",
      "                    #durations.append(len(audio)/44100.) # Using hard-coded 44100 sr for now\n",
      "                    duration_in_seconds = len(audio)/44100.\n",
      "                    f.write('{0}, '.format(duration_in_seconds))\n",
      "                    print '{0},'.format(duration_in_seconds),\n",
      "                    del loader\n",
      "                    del audio\n",
      "                    gc.collect()\n",
      "                except (RuntimeError, TypeError, NameError):\n",
      "                    #print 'could not parse file $s\\n' % name\n",
      "                    continue\n",
      "        print\n",
      "                        \n",
      "#print extensions.keys()\n",
      "#print durations\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}