Kernel_t-SNE_porting.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import sklearn.neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "euc = sklearn.neighbors.DistanceMetric.get_metric('euclidean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from jaratoolbox import loadopenephys\n",
    "from jaratoolbox import spikesorting\n",
    "from jaratoolbox.test.nick import clustercutting\n",
    "from matplotlib import pyplot as plt\n",
    "import os\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "animalName='pinp013'\n",
    "ephysLoc = '/home/nick/data/ephys/'\n",
    "ephysPath = os.path.join(ephysLoc, animalName)\n",
    "ephysFn='2016-05-27_14-13-26'\n",
    "tetrode=3\n",
    "spikesFn = os.path.join(ephysPath, ephysFn, 'Tetrode{}.spikes'.format(tetrode))\n",
    "dataSpikes = loadopenephys.DataSpikes(spikesFn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10111"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(dataSpikes.samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#GAIN = 5000.0\n",
    "#SAMPLING_RATE=30000.0\n",
    "#dataSpikes.samples = ((dataSpikes.samples - 32768.0) / GAIN) * 1000.0\n",
    "#dataSpikes.timestamps = dataSpikes.timestamps/SAMPLING_RATE\n",
    "\n",
    "(numSpikes, numChans, numSamples) = shape(dataSpikes.samples)\n",
    "allWaves = dataSpikes.samples.reshape(numSpikes, numChans*numSamples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "spikesToUse = random.randint(len(allWaves), size=1000) #Only use some of the spikes for now\n",
    "wavesToUse = allWaves[spikesToUse, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_dist = euc.pairwise(wavesToUse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.manifold import TSNE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[t-SNE] Computing pairwise distances...\n",
      "[t-SNE] Computing 91 nearest neighbors...\n",
      "[t-SNE] Computed conditional probabilities for sample 1000 / 1000\n",
      "[t-SNE] Mean sigma: 312.274894\n",
      "[t-SNE] Iteration 25: error = 1.3697388, gradient norm = 0.0179968\n",
      "[t-SNE] Iteration 50: error = 1.3613247, gradient norm = 0.0181911\n",
      "[t-SNE] Iteration 75: error = 1.0027137, gradient norm = 0.0101352\n",
      "[t-SNE] Iteration 100: error = 1.0242578, gradient norm = 0.0123768\n",
      "[t-SNE] KL divergence after 100 iterations with early exaggeration: 1.024258\n",
      "[t-SNE] Iteration 125: error = 0.9281195, gradient norm = 0.0076184\n",
      "[t-SNE] Iteration 150: error = 0.9502608, gradient norm = 0.0094917\n",
      "[t-SNE] Iteration 175: error = 0.9646636, gradient norm = 0.0107351\n",
      "[t-SNE] Iteration 175: did not make any progress during the last 30 episodes. Finished.\n",
      "[t-SNE] Error after 175 iterations: 1.024258\n",
      "ELAPSED TIME: 0.0862698833148 mins\n"
     ]
    }
   ],
   "source": [
    "import timeit\n",
    "start_time = timeit.default_timer()\n",
    "\n",
    "model = TSNE(n_components=2, method='barnes_hut', verbose=20, n_iter=1000)\n",
    "Y = model.fit_transform(wavesToUse)\n",
    "\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print 'ELAPSED TIME: {} mins'.format(elapsed/60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f42ea776d90>]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEACAYAAAC9Gb03AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f42eac1a290>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(Y[:,0], Y[:,1], '.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import sklearn.mixture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mixture = sklearn.mixture.GMM(n_components=12) #Gaussian mixture modeling looks like it might work well. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GMM(covariance_type='diag', init_params='wmc', min_covar=0.001,\n",
       "  n_components=12, n_init=1, n_iter=100, params='wmc', random_state=None,\n",
       "  tol=0.001, verbose=0)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mixture.fit(Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([10,  3, 10,  9,  1,  7, 10,  7, 11, 11,  9, 10,  9,  3,  9,  5,  7,\n",
       "       11, 10,  5,  3, 10,  9,  1, 10,  1,  9, 10,  8,  3,  4, 10, 10,  3,\n",
       "        7, 11,  9, 10, 10, 10, 11,  6,  3, 11, 10,  5,  2,  5,  2,  7, 10,\n",
       "       10, 11, 10,  4,  9, 10,  9, 10,  4,  8,  5, 10, 11,  9, 10, 10,  7,\n",
       "        4,  9,  9,  7,  7,  3,  9, 10,  6, 10,  3,  6,  8, 11,  6, 11,  7,\n",
       "        3, 11,  6, 11, 10,  8, 11,  4,  7, 10, 11, 10,  9,  7,  4, 11, 10,\n",
       "       10,  6, 10,  7,  8,  9,  8, 11,  9,  8, 10,  9,  9,  4,  9, 10,  9,\n",
       "       10, 10, 10,  9, 10,  1, 10,  3,  3, 10,  9,  9,  8, 11,  8,  9,  9,\n",
       "        1,  1,  1,  1, 10,  7,  0,  4,  2,  4,  7,  0,  9, 10,  7,  4, 10,\n",
       "        8,  4, 10,  4,  7,  9,  7,  8, 11,  9, 11, 11,  8,  7, 10,  9, 10,\n",
       "        7,  8,  9,  2, 10,  9,  9, 10,  7, 11,  9, 10,  3, 11,  4, 10,  3,\n",
       "       11,  6,  3, 10,  5,  9,  9,  3,  3, 11,  7, 10,  6,  9,  1,  1, 10,\n",
       "        2,  5,  7,  2,  0,  5, 10, 10,  7,  8,  0,  9, 11,  8,  3,  5,  5,\n",
       "        5,  4, 11, 11,  8,  5, 10, 10, 10,  7, 11, 11, 11, 11,  8,  4,  4,\n",
       "       10, 10, 10,  2,  7, 10, 10,  9,  9,  3,  2, 10,  6, 10,  8, 11,  8,\n",
       "        3,  2, 10, 10, 10,  9, 10,  4, 10, 10,  2,  1, 10,  4,  3, 10, 11,\n",
       "        2,  1, 10,  4,  3, 11,  4,  1,  7, 10,  9,  9, 10, 10, 10,  4,  4,\n",
       "       10,  8,  2,  7,  9, 10, 11,  3,  1, 10, 10, 11, 10,  3,  4,  3,  9,\n",
       "        3, 10, 10,  1,  4, 11, 10,  1, 10,  3,  9, 10,  8,  9,  7, 10,  6,\n",
       "        1,  6, 11,  3,  7,  7, 10,  8, 11,  9,  1,  9, 11,  9,  5,  5,  5,\n",
       "        8,  4, 11, 10, 11, 11,  3, 10,  9,  7,  1,  1,  1,  9,  9,  7,  7,\n",
       "       10, 10,  9,  9, 10,  1,  9,  1,  3,  7,  8,  8,  9,  7, 10, 10, 10,\n",
       "        1, 10,  5, 11, 11,  6, 11, 10,  9,  5, 11, 10,  9, 11,  9, 11, 10,\n",
       "       11,  0,  9, 10, 10,  9, 10,  9, 11,  8, 11, 11,  6, 11,  1,  6,  9,\n",
       "        7,  7,  4,  9,  7,  4,  1,  9,  1,  9, 10,  9, 11,  6,  8, 11,  8,\n",
       "        9,  4, 10, 10, 10, 10,  2, 11, 11,  7, 10, 10,  1,  7,  8,  8,  7,\n",
       "        2, 11,  4,  4,  4,  7,  8, 11, 11,  7,  7,  6,  6,  7,  4, 10, 10,\n",
       "        1,  9,  3,  9, 10,  5,  3,  7, 11, 10, 11,  9,  9, 10,  7,  8,  7,\n",
       "        9, 10,  7, 11, 10,  7, 11,  9,  4,  6,  5,  8, 10, 10, 10, 11,  9,\n",
       "       10,  7,  9,  9, 10, 11, 10,  9,  5, 11,  9, 10,  9,  7,  9,  7,  9,\n",
       "        6,  3, 10, 11,  9,  7,  7,  1,  9, 10,  9, 10,  2, 11, 10,  8,  9,\n",
       "        9, 10,  1,  9, 10, 10,  8, 11,  7,  9,  4,  1,  4,  5,  3,  9,  7,\n",
       "        9,  7,  7,  9, 10, 10,  7,  9,  7, 10,  7,  1,  1,  7, 11,  7, 11,\n",
       "        7,  9, 10,  9,  6,  8,  7, 11,  1, 10, 11, 11, 10, 10, 11, 10, 10,\n",
       "        1,  7,  9,  7, 11, 10, 10, 11,  9,  7,  3,  4, 10,  7, 10,  9,  6,\n",
       "        9,  9,  3,  4, 11, 11, 10,  5,  8, 10,  7, 11, 11,  7, 11, 11,  7,\n",
       "        7,  1,  5,  9, 11,  2,  8,  9, 11, 10,  4,  4,  3,  8, 11, 10, 10,\n",
       "        9,  5,  5,  3, 10,  4,  4,  9,  7,  7,  2, 10,  9,  9,  5,  6, 11,\n",
       "        7,  9, 11, 11,  1, 10, 10, 11, 10,  6,  7,  5, 10, 11, 10, 10,  7,\n",
       "       11,  6,  9, 10,  8, 10,  7, 10,  4,  3,  7,  7, 11, 10,  9, 10,  1,\n",
       "        3, 11, 11, 11, 10,  8, 11,  2,  3, 11,  5,  3, 11,  7, 10,  6,  4,\n",
       "        9,  8,  7,  8,  6, 11,  9, 10,  8, 11,  4,  9, 10, 10, 11,  3,  1,\n",
       "       10,  4, 11, 10,  6, 11,  9, 10,  8, 11,  6, 10, 10,  7, 10,  3,  5,\n",
       "        9, 10,  7,  6,  7, 10, 10,  2,  9,  2,  8,  2, 11,  7, 10,  6,  6,\n",
       "        9, 10,  4, 10, 10,  9,  5,  9,  3,  7,  9,  1,  7,  9, 10,  7, 10,\n",
       "       10,  7,  1, 11,  2,  7,  9,  9,  1,  7,  9, 10,  9,  3, 10,  9,  9,\n",
       "       11,  7, 11,  7,  4, 11, 10, 11,  2,  3,  9, 11, 11, 10, 10,  9,  9,\n",
       "        9,  8,  1, 10, 10, 10,  9,  8,  9,  5,  9,  1, 10,  1, 10,  8,  9,\n",
       "       10,  9,  9, 10,  0, 11, 11,  7,  7, 11,  7, 11,  7, 10,  2,  7, 10,\n",
       "        9,  8, 11,  3,  7, 10,  7, 10, 11, 10,  8, 10, 10,  9,  9,  4, 11,\n",
       "        7,  9, 11,  7, 10, 10,  9,  6,  9,  3,  9, 10,  9,  1,  1,  7, 10,\n",
       "        6,  9, 10, 10,  9,  2,  8, 11,  5, 11,  0,  4,  4, 10, 11,  7,  4,\n",
       "        7,  3,  9, 11, 10,  5,  9, 10, 10, 10,  8, 10,  7,  1, 11,  7,  9,\n",
       "        9,  9,  3,  4,  9,  9,  3, 10, 10,  6,  9,  7,  4,  6,  4,  2, 10,\n",
       "        7, 11, 10,  9, 11,  4,  3,  7,  4,  7, 11,  9,  9, 11,  7,  5, 10,\n",
       "        8,  7,  8,  9,  9, 10,  9,  9,  8,  7, 11, 10,  7,  4, 10, 10,  4,\n",
       "       11, 10, 10, 10,  1,  1,  9,  5,  7,  9,  7,  3,  7,  1,  3,  2,  8,\n",
       "        7,  4, 11, 10,  4,  4, 10,  3,  3,  5, 11, 10, 10,  4, 10,  9, 10,\n",
       "        7, 10,  8,  8,  8, 10,  1, 10,  7,  9, 10, 11,  9,  3])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mixture.predict(Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlkAAAJPCAYAAABPd8HBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f42ee2af150>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figure(figsize=(10, 10))\n",
    "uniqueLabels = np.unique(cluster_labels)\n",
    "colors = plt.cm.Paired(np.linspace(0, 1, len(uniqueLabels)))\n",
    "for indLabel, label in enumerate(unique(cluster_labels)):\n",
    "    hold(1)\n",
    "    indsThisLabel = np.flatnonzero(cluster_labels==label)\n",
    "    plot(Y[indsThisLabel, 0], Y[indsThisLabel, 1], '.', color=colors[indLabel])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[t-SNE] Computing 91 nearest neighbors...\n",
      "[t-SNE] Computed conditional probabilities for sample 1000 / 1000\n",
      "[t-SNE] Mean sigma: 7.944425\n",
      "[t-SNE] Iteration 25: error = 1.4007232, gradient norm = 0.0194313\n",
      "[t-SNE] Iteration 50: error = 1.3688711, gradient norm = 0.0202096\n",
      "[t-SNE] Iteration 75: error = 1.0311652, gradient norm = 0.0122536\n",
      "[t-SNE] Iteration 100: error = 1.0379034, gradient norm = 0.0141325\n",
      "[t-SNE] KL divergence after 100 iterations with early exaggeration: 1.037903\n",
      "[t-SNE] Iteration 125: error = 0.9480730, gradient norm = 0.0118740\n",
      "[t-SNE] Iteration 150: error = 1.0017443, gradient norm = 0.0117694\n",
      "[t-SNE] Iteration 175: error = 1.0400749, gradient norm = 0.0123166\n",
      "[t-SNE] Iteration 175: did not make any progress during the last 30 episodes. Finished.\n",
      "[t-SNE] Error after 175 iterations: 1.037903\n",
      "ELAPSED TIME: 0.0789584477743 mins\n"
     ]
    }
   ],
   "source": [
    "import timeit\n",
    "start_time = timeit.default_timer()\n",
    "\n",
    "model = TSNE(n_components=2, method='barnes_hut', verbose=20, n_iter=1000, metric='precomputed')\n",
    "Y = model.fit_transform(X_dist)\n",
    "\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print 'ELAPSED TIME: {} mins'.format(elapsed/60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7ff29e90fad0>]"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEACAYAAAC9Gb03AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff299c35a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(Y[:,0], Y[:,1], '.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sigma = 0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import sklearn.neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.        ,  1.41421356,  2.82842712],\n",
       "       [ 1.41421356,  0.        ,  1.41421356],\n",
       "       [ 2.82842712,  1.41421356,  0.        ]])"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = array([[1, 1], [2, 2], [3, 3]])\n",
    "dist = euc.pairwise(a)\n",
    "dist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 2.82842712,  1.41421356,  2.82842712])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nbrs = sklearn.neighbors.NearestNeighbors(n_neighbors=3, algorithm='auto').fit(a)\n",
    "distances, indices = nbrs.kneighbors(a)\n",
    "distances[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def determine_sigma_from_X(X, k_nb, local=True):\n",
    "    #Adapted from KMap Toolbox by Barbara Hammer's group\n",
    "    #Just needs the raw vectors, not a distance matrix\n",
    "    \n",
    "    k = k_nb + 1 ##### We add 1 to the number of the neighbor to return because the first will always be zero (dist to self)\n",
    "    \n",
    "    nbrs = sklearn.neighbors.NearestNeighbors(n_neighbors=k, algorithm='auto').fit(X)\n",
    "    distances, indices = nbrs.kneighbors(X)\n",
    "    \n",
    "    if local:\n",
    "        return distances[:,-1] #Return distances to last (kth)\n",
    "    else:\n",
    "        return mean(distances[:,-1]) #Return average distance to the kth neighbor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def determine_sigma(X_dist, k_nb, local=True):\n",
    "    #Uses numpy quicksort, not sure how fast it will be. \n",
    "    #FIXME: might need to work on this more, the toolbox updates to exclude more zeros\n",
    "    \n",
    "    sortedDist = np.sort(X_dist, axis=0)\n",
    "    \n",
    "    if local:\n",
    "        return sortedDist[k_nb, :]\n",
    "    else:\n",
    "        raise NotImplemented"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 2.82842712,  1.41421356,  2.82842712])"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sig = determine_sigma(dist, 2)\n",
    "sig"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0. ,  1. ,  1. ],\n",
       "       [ 0.5,  0. ,  0.5],\n",
       "       [ 1. ,  1. ,  0. ]])"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dist / sig #makes quick work of porting bsxfun(@rdivide, X_dist, Par.sig_nb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "Ker = exp(-1*(dist/sig))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 1.        ,  0.36787944,  0.36787944],\n",
       "       [ 0.60653066,  1.        ,  0.60653066],\n",
       "       [ 0.36787944,  0.36787944,  1.        ]])"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ker"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1.73575888,  2.21306132,  1.73575888])"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rowSum = np.sum(Ker, 1)\n",
    "rowSum"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 1.73575888,  2.21306132,  1.73575888]])"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rowSum.reshape((1, 3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.57611688,  0.21194156,  0.21194156],\n",
       "       [ 0.27406862,  0.45186276,  0.27406862],\n",
       "       [ 0.21194156,  0.21194156,  0.57611688]])"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ker / rowSum[:, np.newaxis]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 4.01768295, -1.95497378, -0.74858965],\n",
       "       [-1.98280532,  7.26914347, -1.98280532],\n",
       "       [-0.74858965, -1.95497378,  4.01768295]])"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ker = Ker / rowSum[:, np.newaxis]\n",
    "\n",
    "np.linalg.pinv(Ker)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def kmap_train(X_dist, Y, sigma_scale):\n",
    "    #Ported from KMap Toolbox by Barbara Hammer's group\n",
    "    \n",
    "    k_nb = 10 # Number of neighbors to use\n",
    "    f_local =1 # 0 for global, 1 for local\n",
    "\n",
    "    sig_nb=determine_sigma(X_dist, k_nb, f_local);\n",
    "    sig_nb = sig_nb**2\n",
    "    sib_nb = sig_nb * sigma_scale #Scale the sigma values\n",
    "    \n",
    "    #Compute the kernel\n",
    "    kernel = np.exp(-1*(X_dist/sig_nb))\n",
    "    \n",
    "    #Divide each row by the row sum (use newaxis trick to get broadcasting to behave)\n",
    "    kernel = kernel / np.sum(kernel, 1)[:,np.newaxis]\n",
    "    \n",
    "    A = np.dot(np.linalg.pinv(kernel), Y)\n",
    "    \n",
    "    return A, sig_nb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "A, sig_nb = kmap_train(X_dist, Y, 0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Test dataset\n",
    "spikesToUse = random.randint(len(allWaves), size=5000) #Only use some of the spikes for now\n",
    "testWaves = allWaves[spikesToUse, :]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10111, 1000)"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Dist between train and test datasets\n",
    "X_dist_ose = sklearn.metrics.pairwise.euclidean_distances(allWaves, wavesToUse)\n",
    "shape(X_dist_ose)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def kmap_test(X_dist_ose, A, sig_nb):\n",
    "    kernel_ose = np.exp(-1*(X_dist_ose/sig_nb))\n",
    "    \n",
    "    #Divide each row by the row sum (use newaxis trick to get broadcasting to behave)\n",
    "    kernel_ose = kernel_ose / np.sum(kernel_ose, 1)[:,np.newaxis]\n",
    "    \n",
    "    Y_ose = np.dot(kernel_ose, A)\n",
    "    \n",
    "    return Y_ose"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "Y_ose = kmap_test(X_dist_ose, A, sig_nb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7ff29e781610>]"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEACAYAAAC9Gb03AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n1TMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff299c50ad0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(Y_ose[:,0], Y_ose[:,1], '.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/pkg_resources/__init__.py:1268: UserWarning: /home/nick/.python-eggs is writable by group/others and vulnerable to attack when used with get_resource_filename. Consider a more secure location (set with .set_extraction_path or the PYTHON_EGG_CACHE environment variable).\n",
      "  warnings.warn(msg, UserWarning)\n"
     ]
    }
   ],
   "source": [
    "import hdbscan\n",
    "clus = hdbscan.HDBSCAN(min_cluster_size=10)\n",
    "cluster_labels = clus.fit_predict(Y_ose)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlkAAAJPCAYAAABPd8HBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2v8HEFU4rSkQ2i0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff29e6d4c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figure(figsize=(10, 10))\n",
    "uniqueLabels = np.unique(cluster_labels)\n",
    "colors = plt.cm.Paired(np.linspace(0, 1, len(uniqueLabels)))\n",
    "for indLabel, label in enumerate(unique(cluster_labels)):\n",
    "    hold(1)\n",
    "    indsThisLabel = np.flatnonzero(cluster_labels==label)\n",
    "    plot(Y_ose[indsThisLabel, 0], Y_ose[indsThisLabel, 1], '.', color=colors[indLabel])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "animalName='adap020'\n",
    "ephysLoc = '/home/nick/data/ephys/'\n",
    "ephysPath = os.path.join(ephysLoc, animalName)\n",
    "ephysFn='2016-05-25_16-33-09'\n",
    "tetrode=2\n",
    "spikesFn = os.path.join(ephysPath, ephysFn, 'Tetrode{}.spikes'.format(tetrode))\n",
    "dataSpikes = loadopenephys.DataSpikes(spikesFn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "(numSpikes, numChans, numSamples) = shape(dataSpikes.samples)\n",
    "allWaves = dataSpikes.samples.reshape(numSpikes, numChans*numSamples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "spikesToUse = random.randint(len(allWaves), size=5000) #Only use some of the spikes for now\n",
    "wavesToUse = allWaves[spikesToUse, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_dist = euc.pairwise(wavesToUse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[t-SNE] Computing 91 nearest neighbors...\n",
      "[t-SNE] Computed conditional probabilities for sample 1000 / 5000\n",
      "[t-SNE] Computed conditional probabilities for sample 2000 / 5000\n",
      "[t-SNE] Computed conditional probabilities for sample 3000 / 5000\n",
      "[t-SNE] Computed conditional probabilities for sample 4000 / 5000\n",
      "[t-SNE] Computed conditional probabilities for sample 5000 / 5000\n",
      "[t-SNE] Mean sigma: 6.650410\n",
      "[t-SNE] Iteration 25: error = 1.8181822, gradient norm = 0.0096474\n",
      "[t-SNE] Iteration 50: error = 1.7733963, gradient norm = 0.0054940\n",
      "[t-SNE] Iteration 75: error = 1.6334172, gradient norm = 0.0017734\n",
      "[t-SNE] Iteration 100: error = 1.6030693, gradient norm = 0.0014694\n",
      "[t-SNE] KL divergence after 100 iterations with early exaggeration: 1.603069\n",
      "[t-SNE] Iteration 125: error = 1.5512971, gradient norm = 0.0010687\n",
      "[t-SNE] Iteration 150: error = 1.5346128, gradient norm = 0.0009668\n",
      "[t-SNE] Iteration 150: gradient norm 0.000967. Finished.\n",
      "[t-SNE] Error after 150 iterations: 1.603069\n",
      "ELAPSED TIME: 0.562895484765 mins\n"
     ]
    }
   ],
   "source": [
    "import timeit\n",
    "start_time = timeit.default_timer()\n",
    "\n",
    "model = TSNE(n_components=2, method='barnes_hut', verbose=20, n_iter=1000, metric='precomputed')\n",
    "Y = model.fit_transform(X_dist)\n",
    "\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print 'ELAPSED TIME: {} mins'.format(elapsed/60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ELAPSED TIME: 3.26624031862 mins\n"
     ]
    }
   ],
   "source": [
    "start_time = timeit.default_timer()\n",
    "A, sig_nb = kmap_train(X_dist, Y, 0.1)\n",
    "elapsed = timeit.default_timer() - start_time\n",
    "print 'ELAPSED TIME: {} mins'.format(elapsed/60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "chunkWaves = allWaves[0:100000, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_dist_ose = sklearn.metrics.pairwise.euclidean_distances(chunkWaves, wavesToUse)\n",
    "shape(X_dist_ose)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10111"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(allWaves)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "200.0"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "1000000/5000.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "100"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "200*30/60"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fd"
   ]
  }
 ],
 "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": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}