ch05_clustering/Concept01_clustering.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "# Ch `05`: Concept `01`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "We're going to need a Python library called the BregmanToolkit. It's available here: https://github.com/BinRoot/BregmanToolkit\n",
    "\n",
    "One way to install it is by downloading from the GitHub link above, and then running `sudo python setup.py install` in the downloaded directory. You will also need scipy: `sudo pip2 install scipy`. Unfortunately, this BregmanToolkit library isn't supported on Python 3, so you'll have to use Python 2.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "from bregman.suite import *"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Define some hyper-parameters for clustering:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "k = 2\n",
    "max_iterations = 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Select the location for the audio files:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "filenames = tf.train.match_filenames_once('./audio_dataset/*.wav')\n",
    "count_num_files = tf.size(filenames)\n",
    "filename_queue = tf.train.string_input_producer(filenames)\n",
    "reader = tf.WholeFileReader()\n",
    "filename, file_contents = reader.read(filename_queue)\n",
    "\n",
    "chromo = tf.placeholder(tf.float32)\n",
    "max_freqs = tf.argmax(chromo, 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Create a helper function to get the next audio file's Chromogram:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def get_next_chromogram(sess):\n",
    "    audio_file = sess.run(filename)\n",
    "    print('Loading {}'.format(audio_file))\n",
    "    F = Chromagram(audio_file, nfft=16384, wfft=8192, nhop=2205)\n",
    "    return F.X, audio_file"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "And create a helper function to extract a feature vector from the Chromogram data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def extract_feature_vector(sess, chromo_data):\n",
    "    num_features, num_samples = np.shape(chromo_data)\n",
    "    freq_vals = sess.run(max_freqs, feed_dict={chromo: chromo_data})\n",
    "    hist, bins = np.histogram(freq_vals, bins=range(num_features + 1))\n",
    "    normalized_hist = hist.astype(float) / num_samples\n",
    "    return normalized_hist"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Extract a dataset of feature vectors by calling our helper functions above:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def get_dataset(sess):\n",
    "    num_files = sess.run(count_num_files)\n",
    "    coord = tf.train.Coordinator()\n",
    "    threads = tf.train.start_queue_runners(coord=coord)\n",
    "    xs = list()\n",
    "    names = list()\n",
    "    \n",
    "    for _ in range(num_files):\n",
    "        plt.figure()\n",
    "        chromo_data, filename = get_next_chromogram(sess)\n",
    "\n",
    "        plt.subplot(1, 2, 1)\n",
    "        plt.imshow(chromo_data, cmap='Greys', interpolation='nearest')\n",
    "        plt.title('Visualization of Sound Spectrum')\n",
    "\n",
    "        plt.subplot(1, 2, 2)\n",
    "        freq_vals = sess.run(max_freqs, feed_dict={chromo: chromo_data})\n",
    "        plt.hist(freq_vals)\n",
    "        plt.title('Histogram of Notes')\n",
    "        plt.xlabel('Musical Note')\n",
    "        plt.ylabel('Count')\n",
    "        plt.savefig('{}.png'.format(filename))\n",
    "        plt.show()\n",
    "\n",
    "        names.append(filename)\n",
    "        x = extract_feature_vector(sess, chromo_data)\n",
    "        xs.append(x)\n",
    "    xs = np.asmatrix(xs)\n",
    "    coord.request_stop()\n",
    "    coord.join(threads)\n",
    "    return xs, names"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "k-means clustering requires 2 main phases: clustering and assignment. We'll also throw in an initialization helper function for good measure. Here are the three components:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def initial_cluster_centroids(X, k):\n",
    "    return X[0:k, :]\n",
    "\n",
    "\n",
    "def assign_cluster(X, centroids):\n",
    "    expanded_vectors = tf.expand_dims(X, 0)\n",
    "    expanded_centroids = tf.expand_dims(centroids, 1)\n",
    "    distances = tf.reduce_sum(tf.square(tf.subtract(expanded_vectors, expanded_centroids)), 2)\n",
    "    mins = tf.argmin(distances, 0)\n",
    "    return mins\n",
    "\n",
    "\n",
    "def recompute_centroids(X, Y):\n",
    "    sums = tf.unsorted_segment_sum(X, Y, k)\n",
    "    counts = tf.unsorted_segment_sum(tf.ones_like(X), Y, k)\n",
    "    return sums / counts\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Open a session, obtain a dataset, and cluster the data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading ./audio_dataset/cough_1.wav\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/bregman/features_base.py:353: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future\n",
      "  mxnorm = P.empty(self._cqtN) # Normalization coefficients\n",
      "/usr/local/lib/python2.7/dist-packages/bregman/features_base.py:357: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future\n",
      "  for i in P.arange(self._cqtN)])\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAEZCAYAAABy91VnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nA8n1UAt/SamFAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4cef2dc590>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading ./audio_dataset/cough_2.wav\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAEZCAYAAABy91VnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nB3kzswL7X0UqhEbhhh4VAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4cef21b810>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading ./audio_dataset/scream_1.wav\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEZCAYAAABoy7XpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIoS9EgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4ce47ee3d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading ./audio_dataset/scream_2.wav\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAEZCAYAAACD/A7qAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4cbdf11ed0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading ./audio_dataset/scream_3.wav\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEZCAYAAABoy7XpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4cbddfe090>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[(0, './audio_dataset/cough_1.wav'), (1, './audio_dataset/cough_2.wav'), (0, './audio_dataset/scream_1.wav'), (0, './audio_dataset/scream_2.wav'), (0, './audio_dataset/scream_3.wav')]\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    X, names = get_dataset(sess)\n",
    "    centroids = initial_cluster_centroids(X, k)\n",
    "    i, converged = 0, False\n",
    "    while not converged and i < max_iterations:\n",
    "        i += 1\n",
    "        Y = assign_cluster(X, centroids)\n",
    "        centroids = sess.run(recompute_centroids(X, Y))\n",
    "    print(zip(sess.run(Y), names))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}