DT_experiments.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dense Trajectories (DT features) with Fisher encoding\n",
    "\n",
    "\n",
    "## A short summary\n",
    "\n",
    "\n",
    "<font color='red'>This feature extraction process removes the **human detector**, **SURF features**, **RANSAC estimation**.</font>\n",
    "\n",
    "This jupyter notebook shows the feature analysis on the iCetanaPrivate dataset. To visualize the feature space, we use two experimental settings, which are t-SNE with PCA applied for feature dimensionality reduction and t-SNE only for feature dimensionality reduction. There are 4 csv files created when running this jupyter notebook:\n",
    "\n",
    "- `DT_features.csv`: a concatenation of a subset of DT features (around 120 videos)\n",
    "- `sampled_features.csv`: a random sampling of 100,000 DT features extracted from the dataset for a GMM estimation\n",
    "- `fisher_vectors.csv`: a concatenation of all the fisher encoded DT features for the iCetanaPrivate dataset\n",
    "- `all_labels.csv`: a list of labels for the whole dataset (total 270 videos)\n",
    "\n",
    "The fisher kernel implementation is based on the ICCV 2011 paper: *J. Krapac, J. Verbeek, F.Jurie. Modeling Spatial Layout with Fisher Vectors for Image Categorization.*\n",
    "\n",
    "Most of the parameters we used in our experiments are the default values given by the author, such as trajectory length $L=5$, and spatiotemporal grid size $n_\\sigma \\times n_\\sigma \\times n_t = 2 \\times 2 \\times 3$. Other parameters that we updated in our experiments are shown in the following:\n",
    "\n",
    "- The number of Gaussians $K$: 128\n",
    "- The dimension of fisher vector: $K + 2DK$, where $D$ is the descriptor dimension\n",
    "- Sampled feature vectors for building a GMM estimation: 100,000\n",
    "\n",
    "\n",
    "## 1 Data preprocessing\n",
    "In this step, we concatenate all the DT features extracted from the iCetanaPrivateDataset. Moreover, we also get the labels for the videos and concatenate them together for further analysis (e.g., plot data points in t-SNE space with different colors specified by labels)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Concatenating features from:  DT_features/0040005.features.gz\n",
      "Concatenating features from:  DT_features/0011035.features.gz\n",
      "Concatenating features from:  DT_features/0041042.features.gz\n",
      "Concatenating features from:  DT_features/0001023.features.gz\n",
      "Concatenating features from:  DT_features/0050013.features.gz\n",
      "Concatenating features from:  DT_features/0001053.features.gz\n",
      "Concatenating features from:  DT_features/0010002.features.gz\n",
      "Concatenating features from:  DT_features/0040047.features.gz\n",
      "Concatenating features from:  DT_features/0000056.features.gz\n",
      "Concatenating features from:  DT_features/0040037.features.gz\n",
      "Concatenating features from:  DT_features/0011007.features.gz\n",
      "Concatenating features from:  DT_features/0001061.features.gz\n",
      "Concatenating features from:  DT_features/0000008.features.gz\n",
      "Concatenating features from:  DT_features/0011029.features.gz\n",
      "Concatenating features from:  DT_features/0001011.features.gz\n",
      "Concatenating features from:  DT_features/0051048.features.gz\n",
      "Concatenating features from:  DT_features/0040019.features.gz\n",
      "Concatenating features from:  DT_features/0010040.features.gz\n",
      "Concatenating features from:  DT_features/0000078.features.gz\n",
      "Concatenating features from:  DT_features/0050021.features.gz\n",
      "Concatenating features from:  DT_features/0001074.features.gz\n",
      "Concatenating features from:  DT_features/0050044.features.gz\n",
      "Concatenating features from:  DT_features/0050034.features.gz\n",
      "Concatenating features from:  DT_features/0000033.features.gz\n",
      "Concatenating features from:  DT_features/0040052.features.gz\n",
      "Concatenating features from:  DT_features/0051003.features.gz\n",
      "Concatenating features from:  DT_features/0040022.features.gz\n",
      "Concatenating features from:  DT_features/0050006.features.gz\n",
      "Concatenating features from:  DT_features/0001046.features.gz\n",
      "Concatenating features from:  DT_features/0050028.features.gz\n",
      "Concatenating features from:  DT_features/0040010.features.gz\n",
      "Concatenating features from:  DT_features/0001018.features.gz\n",
      "Concatenating features from:  DT_features/0021006.features.gz\n",
      "Concatenating features from:  DT_features/0011020.features.gz\n",
      "Concatenating features from:  DT_features/0050058.features.gz\n",
      "Concatenating features from:  DT_features/0010039.features.gz\n",
      "Concatenating features from:  DT_features/0001068.features.gz\n",
      "Concatenating features from:  DT_features/0011017.features.gz\n",
      "Concatenating features from:  DT_features/0040027.features.gz\n",
      "Concatenating features from:  DT_features/0000036.features.gz\n",
      "Concatenating features from:  DT_features/0040057.features.gz\n",
      "Concatenating features from:  DT_features/0040009.features.gz\n",
      "Concatenating features from:  DT_features/0050031.features.gz\n",
      "Concatenating features from:  DT_features/0001001.features.gz\n",
      "Concatenating features from:  DT_features/0050041.features.gz\n",
      "Concatenating features from:  DT_features/0001071.features.gz\n",
      "Concatenating features from:  DT_features/0000004.features.gz\n",
      "Concatenating features from:  DT_features/0011025.features.gz\n",
      "Concatenating features from:  DT_features/0021003.features.gz\n",
      "Concatenating features from:  DT_features/0040015.features.gz\n",
      "Concatenating features from:  DT_features/0010012.features.gz\n",
      "Concatenating features from:  DT_features/0001043.features.gz\n",
      "Concatenating features from:  DT_features/0050066.features.gz\n",
      "Concatenating features from:  DT_features/0050016.features.gz\n",
      "Concatenating features from:  DT_features/0001026.features.gz\n",
      "Concatenating features from:  DT_features/0051051.features.gz\n",
      "Concatenating features from:  DT_features/0011030.features.gz\n",
      "Concatenating features from:  DT_features/0050038.features.gz\n",
      "Concatenating features from:  DT_features/0010045.features.gz\n",
      "Concatenating features from:  DT_features/0050024.features.gz\n",
      "Concatenating features from:  DT_features/0001014.features.gz\n",
      "Concatenating features from:  DT_features/0020013.features.gz\n",
      "Concatenating features from:  DT_features/0050054.features.gz\n",
      "Concatenating features from:  DT_features/0001064.features.gz\n",
      "Concatenating features from:  DT_features/0051063.features.gz\n",
      "Concatenating features from:  DT_features/0040032.features.gz\n",
      "Concatenating features from:  DT_features/0040031.features.gz\n",
      "Concatenating features from:  DT_features/0000050.features.gz\n",
      "Concatenating features from:  DT_features/0050009.features.gz\n",
      "Concatenating features from:  DT_features/0001039.features.gz\n",
      "Concatenating features from:  DT_features/0040041.features.gz\n",
      "Concatenating features from:  DT_features/0000020.features.gz\n",
      "Concatenating features from:  DT_features/0001049.features.gz\n",
      "Concatenating features from:  DT_features/0050027.features.gz\n",
      "Concatenating features from:  DT_features/0010046.features.gz\n",
      "Concatenating features from:  DT_features/0001067.features.gz\n",
      "Concatenating features from:  DT_features/0020010.features.gz\n",
      "Concatenating features from:  DT_features/0040003.features.gz\n",
      "Concatenating features from:  DT_features/0051052.features.gz\n",
      "Concatenating features from:  DT_features/0010004.features.gz\n",
      "Concatenating features from:  DT_features/0001025.features.gz\n",
      "Concatenating features from:  DT_features/0050015.features.gz\n",
      "Concatenating features from:  DT_features/0001040.features.gz\n",
      "Concatenating features from:  DT_features/0000029.features.gz\n",
      "Concatenating features from:  DT_features/0050070.features.gz\n",
      "Concatenating features from:  DT_features/0010011.features.gz\n",
      "Concatenating features from:  DT_features/0000059.features.gz\n",
      "Concatenating features from:  DT_features/0040038.features.gz\n",
      "Concatenating features from:  DT_features/0000007.features.gz\n",
      "Concatenating features from:  DT_features/0040016.features.gz\n",
      "Concatenating features from:  DT_features/0000077.features.gz\n",
      "Concatenating features from:  DT_features/0011026.features.gz\n",
      "Concatenating features from:  DT_features/0051047.features.gz\n",
      "Concatenating features from:  DT_features/0001002.features.gz\n",
      "Concatenating features from:  DT_features/0050032.features.gz\n",
      "Concatenating features from:  DT_features/0001072.features.gz\n",
      "Concatenating features from:  DT_features/0050042.features.gz\n",
      "Concatenating features from:  DT_features/0040024.features.gz\n",
      "Concatenating features from:  DT_features/0000045.features.gz\n",
      "Concatenating features from:  DT_features/0040054.features.gz\n",
      "Concatenating features from:  DT_features/0000035.features.gz\n",
      "Concatenating features from:  DT_features/0051005.features.gz\n",
      "Concatenating features from:  DT_features/0011023.features.gz\n",
      "Concatenating features from:  DT_features/0021005.features.gz\n",
      "Concatenating features from:  DT_features/0040013.features.gz\n",
      "Concatenating features from:  DT_features/0010014.features.gz\n",
      "Concatenating features from:  DT_features/0040051.features.gz\n",
      "Concatenating features from:  DT_features/0010008.features.gz\n",
      "Concatenating features from:  DT_features/0000030.features.gz\n",
      "Concatenating features from:  DT_features/0050069.features.gz\n",
      "Concatenating features from:  DT_features/0041048.features.gz\n",
      "Concatenating features from:  DT_features/0040021.features.gz\n",
      "Concatenating features from:  DT_features/0050019.features.gz\n",
      "Concatenating features from:  DT_features/0050037.features.gz\n",
      "Concatenating features from:  DT_features/0010033.features.gz\n",
      "Concatenating features from:  DT_features/0001062.features.gz\n",
      "Concatenating features from:  DT_features/0050022.features.gz\n",
      "Concatenating features from:  DT_features/0010043.features.gz\n",
      "Concatenating features from:  DT_features/0001012.features.gz\n",
      "Concatenating features from:  DT_features/0040044.features.gz\n",
      "Saving all features into a csv file...\n",
      "Done!\n"
     ]
    }
   ],
   "source": [
    "import glob\n",
    "import gzip\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "# define an empty numpy array for concatenating features\n",
    "all_features=np.array([])\n",
    "\n",
    "a=0\n",
    "# concatenate all the extracted features and video labels\n",
    "for filename in glob.iglob('DT_features/*.gz'):\n",
    "    # print(filename)\n",
    "    # use '\\s+' for the \\t and \\n\n",
    "    df = pd.read_table(gzip.open(filename),sep='\\s+', header=None)\n",
    "    \n",
    "    # turn pandas dataframe into array\n",
    "    df_array=np.round(df.values, decimals=3)\n",
    "    print(\"Concatenating features from: \", filename)\n",
    "    \n",
    "    a=a+1\n",
    "    if a == 120:\n",
    "        break\n",
    "        \n",
    "    # concatenate all the features\n",
    "    all_features=np.vstack([all_features, df_array]) if all_features.size else df_array\n",
    "    # print(all_features.shape)\n",
    "    # print(all_label.shape)\n",
    "    \n",
    "    \n",
    "print(\"Saving all features into a csv file...\")    \n",
    "df = pd.DataFrame(all_features)\n",
    "df.to_csv(\"DT_features.csv\", index=False)\n",
    "print(\"Done!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2 Gaussian Mixture Model (GMM)\n",
    "We use the concatenated features (random sampling 100,000 feature vectors) to get an estimation of a Gaussian Mixture Model for the whole dataset (here we use the iCetanaPrivateDataset as an example). At this stage, we set the number of Gaussians $K=128$, which is different from the original authors' in the DT paper. Moreover, at this stage, we do not apply the principle component analysis (PCA) on the concatenated features for feature dimensionality reduction.\n",
    "\n",
    "### 2.1 Random sampling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "descriptors = pd.read_csv('DT_features.csv')\n",
    "descriptors=descriptors.iloc[:, 10:]\n",
    "print(\"Read csv file successfully!\")\n",
    "# random sample 100,000 feature vectors\n",
    "print(\"Start sampling 100,000 feature vectors\")\n",
    "f_samp=descriptors.sample(n=100000, random_state=42)\n",
    "print(\"Save sampled feature vectors...\")\n",
    "f_samp.to_csv(\"sampled_features.csv\", index=False)\n",
    "print(\"Done!\")\n",
    "print(f_samp.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2 GMM estimation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Read csv file successfully!\n",
      "Start the GMM estimation...\n",
      "A GMM estimation for the whole dataset has been finished!\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "# import sklearn GMM model\n",
    "from sklearn.mixture import GMM\n",
    "# ignore all warning messages\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "# define K\n",
    "K = 128\n",
    "gmm = GMM(n_components=K, covariance_type='diag')\n",
    "\n",
    "# get a GMM estimation for the whole dataset\n",
    "samp_f = pd.read_csv('sampled_features.csv')\n",
    "print(\"Read csv file successfully!\")\n",
    "print(\"Start the GMM estimation...\")\n",
    "gmm.fit(samp_f)\n",
    "print(\"A GMM estimation for the whole dataset has been finished!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3 Feature encoding using fisher kernel\n",
    "\n",
    "Now, we use the fisher kernel to encoding the features. For more details about our fisher encoding algorithm, please refer to our code (`fisher_vector.py`)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start fisher encoding for DT_features/0000003.features.gz\n",
      "Start fisher encoding for DT_features/0000004.features.gz\n",
      "Start fisher encoding for DT_features/0000006.features.gz\n",
      "Start fisher encoding for DT_features/0000007.features.gz\n",
      "Start fisher encoding for DT_features/0000008.features.gz\n",
      "Start fisher encoding for DT_features/0000009.features.gz\n",
      "Start fisher encoding for DT_features/0000010.features.gz\n",
      "Start fisher encoding for DT_features/0000013.features.gz\n",
      "Start fisher encoding for DT_features/0000017.features.gz\n",
      "Start fisher encoding for DT_features/0000020.features.gz\n",
      "Start fisher encoding for DT_features/0000024.features.gz\n",
      "Start fisher encoding for DT_features/0000027.features.gz\n",
      "Start fisher encoding for DT_features/0000028.features.gz\n",
      "Start fisher encoding for DT_features/0000029.features.gz\n",
      "Start fisher encoding for DT_features/0000030.features.gz\n",
      "Start fisher encoding for DT_features/0000031.features.gz\n",
      "Start fisher encoding for DT_features/0000033.features.gz\n",
      "Start fisher encoding for DT_features/0000034.features.gz\n",
      "Start fisher encoding for DT_features/0000035.features.gz\n",
      "Start fisher encoding for DT_features/0000036.features.gz\n",
      "Start fisher encoding for DT_features/0000037.features.gz\n",
      "Start fisher encoding for DT_features/0000041.features.gz\n",
      "Start fisher encoding for DT_features/0000045.features.gz\n",
      "Start fisher encoding for DT_features/0000048.features.gz\n",
      "Start fisher encoding for DT_features/0000050.features.gz\n",
      "Start fisher encoding for DT_features/0000051.features.gz\n",
      "Start fisher encoding for DT_features/0000055.features.gz\n",
      "Start fisher encoding for DT_features/0000056.features.gz\n",
      "Start fisher encoding for DT_features/0000057.features.gz\n",
      "Start fisher encoding for DT_features/0000058.features.gz\n",
      "Start fisher encoding for DT_features/0000059.features.gz\n",
      "Start fisher encoding for DT_features/0000075.features.gz\n",
      "Start fisher encoding for DT_features/0000077.features.gz\n",
      "Start fisher encoding for DT_features/0000078.features.gz\n",
      "Start fisher encoding for DT_features/0000079.features.gz\n",
      "Start fisher encoding for DT_features/0001001.features.gz\n",
      "Start fisher encoding for DT_features/0001002.features.gz\n",
      "Start fisher encoding for DT_features/0001005.features.gz\n",
      "Start fisher encoding for DT_features/0001011.features.gz\n",
      "Start fisher encoding for DT_features/0001012.features.gz\n",
      "Start fisher encoding for DT_features/0001014.features.gz\n",
      "Start fisher encoding for DT_features/0001015.features.gz\n",
      "Start fisher encoding for DT_features/0001016.features.gz\n",
      "Start fisher encoding for DT_features/0001018.features.gz\n",
      "Start fisher encoding for DT_features/0001019.features.gz\n",
      "Start fisher encoding for DT_features/0001021.features.gz\n",
      "Start fisher encoding for DT_features/0001022.features.gz\n",
      "Start fisher encoding for DT_features/0001023.features.gz\n",
      "Start fisher encoding for DT_features/0001025.features.gz\n",
      "Start fisher encoding for DT_features/0001026.features.gz\n",
      "Start fisher encoding for DT_features/0001032.features.gz\n",
      "Start fisher encoding for DT_features/0001038.features.gz\n",
      "Start fisher encoding for DT_features/0001039.features.gz\n",
      "Start fisher encoding for DT_features/0001040.features.gz\n",
      "Start fisher encoding for DT_features/0001042.features.gz\n",
      "Start fisher encoding for DT_features/0001043.features.gz\n",
      "Start fisher encoding for DT_features/0001044.features.gz\n",
      "Start fisher encoding for DT_features/0001046.features.gz\n",
      "Start fisher encoding for DT_features/0001047.features.gz\n",
      "Start fisher encoding for DT_features/0001049.features.gz\n",
      "Start fisher encoding for DT_features/0001052.features.gz\n",
      "Start fisher encoding for DT_features/0001053.features.gz\n",
      "Start fisher encoding for DT_features/0001054.features.gz\n",
      "Start fisher encoding for DT_features/0001060.features.gz\n",
      "Start fisher encoding for DT_features/0001061.features.gz\n",
      "Start fisher encoding for DT_features/0001062.features.gz\n",
      "Start fisher encoding for DT_features/0001063.features.gz\n",
      "Start fisher encoding for DT_features/0001064.features.gz\n",
      "Start fisher encoding for DT_features/0001065.features.gz\n",
      "Start fisher encoding for DT_features/0001066.features.gz\n",
      "Start fisher encoding for DT_features/0001067.features.gz\n",
      "Start fisher encoding for DT_features/0001068.features.gz\n",
      "Start fisher encoding for DT_features/0001069.features.gz\n",
      "Start fisher encoding for DT_features/0001070.features.gz\n",
      "Start fisher encoding for DT_features/0001071.features.gz\n",
      "Start fisher encoding for DT_features/0001072.features.gz\n",
      "Start fisher encoding for DT_features/0001073.features.gz\n",
      "Start fisher encoding for DT_features/0001074.features.gz\n",
      "Start fisher encoding for DT_features/0001076.features.gz\n",
      "Start fisher encoding for DT_features/0010001.features.gz\n",
      "Start fisher encoding for DT_features/0010002.features.gz\n",
      "Start fisher encoding for DT_features/0010003.features.gz\n",
      "Start fisher encoding for DT_features/0010004.features.gz\n",
      "Start fisher encoding for DT_features/0010008.features.gz\n",
      "Start fisher encoding for DT_features/0010009.features.gz\n",
      "Start fisher encoding for DT_features/0010011.features.gz\n",
      "Start fisher encoding for DT_features/0010012.features.gz\n",
      "Start fisher encoding for DT_features/0010013.features.gz\n",
      "Start fisher encoding for DT_features/0010014.features.gz\n",
      "Start fisher encoding for DT_features/0010015.features.gz\n",
      "Start fisher encoding for DT_features/0010016.features.gz\n",
      "Start fisher encoding for DT_features/0010033.features.gz\n",
      "Start fisher encoding for DT_features/0010037.features.gz\n",
      "Start fisher encoding for DT_features/0010038.features.gz\n",
      "Start fisher encoding for DT_features/0010039.features.gz\n",
      "Start fisher encoding for DT_features/0010040.features.gz\n",
      "Start fisher encoding for DT_features/0010041.features.gz\n",
      "Start fisher encoding for DT_features/0010042.features.gz\n",
      "Start fisher encoding for DT_features/0010043.features.gz\n",
      "Start fisher encoding for DT_features/0010044.features.gz\n",
      "Start fisher encoding for DT_features/0010045.features.gz\n",
      "Start fisher encoding for DT_features/0010046.features.gz\n",
      "Start fisher encoding for DT_features/0011005.features.gz\n",
      "Start fisher encoding for DT_features/0011006.features.gz\n",
      "Start fisher encoding for DT_features/0011007.features.gz\n",
      "Start fisher encoding for DT_features/0011010.features.gz\n",
      "Start fisher encoding for DT_features/0011017.features.gz\n",
      "Start fisher encoding for DT_features/0011018.features.gz\n",
      "Start fisher encoding for DT_features/0011019.features.gz\n",
      "Start fisher encoding for DT_features/0011020.features.gz\n",
      "Start fisher encoding for DT_features/0011021.features.gz\n",
      "Start fisher encoding for DT_features/0011022.features.gz\n",
      "Start fisher encoding for DT_features/0011023.features.gz\n",
      "Start fisher encoding for DT_features/0011024.features.gz\n",
      "Start fisher encoding for DT_features/0011025.features.gz\n",
      "Start fisher encoding for DT_features/0011026.features.gz\n",
      "Start fisher encoding for DT_features/0011027.features.gz\n",
      "Start fisher encoding for DT_features/0011028.features.gz\n",
      "Start fisher encoding for DT_features/0011029.features.gz\n",
      "Start fisher encoding for DT_features/0011030.features.gz\n",
      "Start fisher encoding for DT_features/0011031.features.gz\n",
      "Start fisher encoding for DT_features/0011032.features.gz\n",
      "Start fisher encoding for DT_features/0011034.features.gz\n",
      "Start fisher encoding for DT_features/0011035.features.gz\n",
      "Start fisher encoding for DT_features/0011036.features.gz\n",
      "Start fisher encoding for DT_features/0020002.features.gz\n",
      "Start fisher encoding for DT_features/0020008.features.gz\n",
      "Start fisher encoding for DT_features/0020009.features.gz\n",
      "Start fisher encoding for DT_features/0020010.features.gz\n",
      "Start fisher encoding for DT_features/0020011.features.gz\n",
      "Start fisher encoding for DT_features/0020012.features.gz\n",
      "Start fisher encoding for DT_features/0020013.features.gz\n",
      "Start fisher encoding for DT_features/0020014.features.gz\n",
      "Start fisher encoding for DT_features/0021001.features.gz\n",
      "Start fisher encoding for DT_features/0021003.features.gz\n",
      "Start fisher encoding for DT_features/0021004.features.gz\n",
      "Start fisher encoding for DT_features/0021005.features.gz\n",
      "Start fisher encoding for DT_features/0021006.features.gz\n",
      "Start fisher encoding for DT_features/0021007.features.gz\n",
      "Start fisher encoding for DT_features/0030001.features.gz\n",
      "Start fisher encoding for DT_features/0030002.features.gz\n",
      "Start fisher encoding for DT_features/0030003.features.gz\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start fisher encoding for DT_features/0040001.features.gz\n",
      "Start fisher encoding for DT_features/0040002.features.gz\n",
      "Start fisher encoding for DT_features/0040003.features.gz\n",
      "Start fisher encoding for DT_features/0040004.features.gz\n",
      "Start fisher encoding for DT_features/0040005.features.gz\n",
      "Start fisher encoding for DT_features/0040006.features.gz\n",
      "Start fisher encoding for DT_features/0040008.features.gz\n",
      "Start fisher encoding for DT_features/0040009.features.gz\n",
      "Start fisher encoding for DT_features/0040010.features.gz\n",
      "Start fisher encoding for DT_features/0040011.features.gz\n",
      "Start fisher encoding for DT_features/0040012.features.gz\n",
      "Start fisher encoding for DT_features/0040013.features.gz\n",
      "Start fisher encoding for DT_features/0040014.features.gz\n",
      "Start fisher encoding for DT_features/0040015.features.gz\n",
      "Start fisher encoding for DT_features/0040016.features.gz\n",
      "Start fisher encoding for DT_features/0040017.features.gz\n",
      "Start fisher encoding for DT_features/0040019.features.gz\n",
      "Start fisher encoding for DT_features/0040020.features.gz\n",
      "Start fisher encoding for DT_features/0040021.features.gz\n",
      "Start fisher encoding for DT_features/0040022.features.gz\n",
      "Start fisher encoding for DT_features/0040023.features.gz\n",
      "Start fisher encoding for DT_features/0040024.features.gz\n",
      "Start fisher encoding for DT_features/0040025.features.gz\n",
      "Start fisher encoding for DT_features/0040026.features.gz\n",
      "Start fisher encoding for DT_features/0040027.features.gz\n",
      "Start fisher encoding for DT_features/0040028.features.gz\n",
      "Start fisher encoding for DT_features/0040029.features.gz\n",
      "Start fisher encoding for DT_features/0040030.features.gz\n",
      "Start fisher encoding for DT_features/0040031.features.gz\n",
      "Start fisher encoding for DT_features/0040032.features.gz\n",
      "Start fisher encoding for DT_features/0040033.features.gz\n",
      "Start fisher encoding for DT_features/0040034.features.gz\n",
      "Start fisher encoding for DT_features/0040035.features.gz\n",
      "Start fisher encoding for DT_features/0040036.features.gz\n",
      "Start fisher encoding for DT_features/0040037.features.gz\n",
      "Start fisher encoding for DT_features/0040038.features.gz\n",
      "Start fisher encoding for DT_features/0040039.features.gz\n",
      "Start fisher encoding for DT_features/0040040.features.gz\n",
      "Start fisher encoding for DT_features/0040041.features.gz\n",
      "Start fisher encoding for DT_features/0040043.features.gz\n",
      "Start fisher encoding for DT_features/0040044.features.gz\n",
      "Start fisher encoding for DT_features/0040045.features.gz\n",
      "Start fisher encoding for DT_features/0040046.features.gz\n",
      "Start fisher encoding for DT_features/0040047.features.gz\n",
      "Start fisher encoding for DT_features/0040050.features.gz\n",
      "Start fisher encoding for DT_features/0040051.features.gz\n",
      "Start fisher encoding for DT_features/0040052.features.gz\n",
      "Start fisher encoding for DT_features/0040053.features.gz\n",
      "Start fisher encoding for DT_features/0040054.features.gz\n",
      "Start fisher encoding for DT_features/0040055.features.gz\n",
      "Start fisher encoding for DT_features/0040056.features.gz\n",
      "Start fisher encoding for DT_features/0040057.features.gz\n",
      "Start fisher encoding for DT_features/0041007.features.gz\n",
      "Start fisher encoding for DT_features/0041042.features.gz\n",
      "Start fisher encoding for DT_features/0041048.features.gz\n",
      "Start fisher encoding for DT_features/0041049.features.gz\n",
      "Start fisher encoding for DT_features/0050002.features.gz\n",
      "Start fisher encoding for DT_features/0050006.features.gz\n",
      "Start fisher encoding for DT_features/0050007.features.gz\n",
      "Start fisher encoding for DT_features/0050008.features.gz\n",
      "Start fisher encoding for DT_features/0050009.features.gz\n",
      "Start fisher encoding for DT_features/0050010.features.gz\n",
      "Start fisher encoding for DT_features/0050011.features.gz\n",
      "Start fisher encoding for DT_features/0050012.features.gz\n",
      "Start fisher encoding for DT_features/0050013.features.gz\n",
      "Start fisher encoding for DT_features/0050014.features.gz\n",
      "Start fisher encoding for DT_features/0050015.features.gz\n",
      "Start fisher encoding for DT_features/0050016.features.gz\n",
      "Start fisher encoding for DT_features/0050017.features.gz\n",
      "Start fisher encoding for DT_features/0050018.features.gz\n",
      "Start fisher encoding for DT_features/0050019.features.gz\n",
      "Start fisher encoding for DT_features/0050020.features.gz\n",
      "Start fisher encoding for DT_features/0050021.features.gz\n",
      "Start fisher encoding for DT_features/0050022.features.gz\n",
      "Start fisher encoding for DT_features/0050023.features.gz\n",
      "Start fisher encoding for DT_features/0050024.features.gz\n",
      "Start fisher encoding for DT_features/0050025.features.gz\n",
      "Start fisher encoding for DT_features/0050026.features.gz\n",
      "Start fisher encoding for DT_features/0050027.features.gz\n",
      "Start fisher encoding for DT_features/0050028.features.gz\n",
      "Start fisher encoding for DT_features/0050029.features.gz\n",
      "Start fisher encoding for DT_features/0050030.features.gz\n",
      "Start fisher encoding for DT_features/0050031.features.gz\n",
      "Start fisher encoding for DT_features/0050032.features.gz\n",
      "Start fisher encoding for DT_features/0050033.features.gz\n",
      "Start fisher encoding for DT_features/0050034.features.gz\n",
      "Start fisher encoding for DT_features/0050035.features.gz\n",
      "Start fisher encoding for DT_features/0050036.features.gz\n",
      "Start fisher encoding for DT_features/0050037.features.gz\n",
      "Start fisher encoding for DT_features/0050038.features.gz\n",
      "Start fisher encoding for DT_features/0050039.features.gz\n",
      "Start fisher encoding for DT_features/0050040.features.gz\n",
      "Start fisher encoding for DT_features/0050041.features.gz\n",
      "Start fisher encoding for DT_features/0050042.features.gz\n",
      "Start fisher encoding for DT_features/0050043.features.gz\n",
      "Start fisher encoding for DT_features/0050044.features.gz\n",
      "Start fisher encoding for DT_features/0050053.features.gz\n",
      "Start fisher encoding for DT_features/0050054.features.gz\n",
      "Start fisher encoding for DT_features/0050055.features.gz\n",
      "Start fisher encoding for DT_features/0050058.features.gz\n",
      "Start fisher encoding for DT_features/0050059.features.gz\n",
      "Start fisher encoding for DT_features/0050060.features.gz\n",
      "Start fisher encoding for DT_features/0050061.features.gz\n",
      "Start fisher encoding for DT_features/0050066.features.gz\n",
      "Start fisher encoding for DT_features/0050067.features.gz\n",
      "Start fisher encoding for DT_features/0050068.features.gz\n",
      "Start fisher encoding for DT_features/0050069.features.gz\n",
      "Start fisher encoding for DT_features/0050070.features.gz\n",
      "Start fisher encoding for DT_features/0050071.features.gz\n",
      "Start fisher encoding for DT_features/0050072.features.gz\n",
      "Start fisher encoding for DT_features/0051001.features.gz\n",
      "Start fisher encoding for DT_features/0051003.features.gz\n",
      "Start fisher encoding for DT_features/0051004.features.gz\n",
      "Start fisher encoding for DT_features/0051005.features.gz\n",
      "Start fisher encoding for DT_features/0051045.features.gz\n",
      "Start fisher encoding for DT_features/0051046.features.gz\n",
      "Start fisher encoding for DT_features/0051047.features.gz\n",
      "Start fisher encoding for DT_features/0051048.features.gz\n",
      "Start fisher encoding for DT_features/0051049.features.gz\n",
      "Start fisher encoding for DT_features/0051050.features.gz\n",
      "Start fisher encoding for DT_features/0051051.features.gz\n",
      "Start fisher encoding for DT_features/0051052.features.gz\n",
      "Start fisher encoding for DT_features/0051056.features.gz\n",
      "Start fisher encoding for DT_features/0051057.features.gz\n",
      "Start fisher encoding for DT_features/0051062.features.gz\n",
      "Start fisher encoding for DT_features/0051063.features.gz\n",
      "Start fisher encoding for DT_features/0051064.features.gz\n",
      "Start fisher encoding for DT_features/0051065.features.gz\n",
      "Saving the fisher encoded features into a csv file...\n",
      "(270, 109184)\n",
      "Done!\n",
      "Saving all the labels into a csv file...\n",
      "Done!\n"
     ]
    }
   ],
   "source": [
    "# import fisher encoding function\n",
    "import numpy as np\n",
    "import glob\n",
    "import gzip\n",
    "import pandas as pd\n",
    "from fisher_vector import fisher_vector\n",
    "\n",
    "# define an empty numpy array to store all the fisher encoding\n",
    "fisher_vec=np.array([])\n",
    "all_label=np.array([])\n",
    "# loop through each video to get the fisher encoding\n",
    "for filename in sorted(glob.iglob('DT_features/*.gz')):\n",
    "    # use '\\s+' for the \\t and \\n\n",
    "    df = pd.read_table(gzip.open(filename),sep='\\s+', header=None)\n",
    "    df = df.iloc[:, 10:]\n",
    "    # concatenate all the labels\n",
    "    vlabel=np.asarray(int(filename[15:16]))\n",
    "    all_label=np.vstack([all_label, vlabel]) if all_label.size else vlabel\n",
    "    \n",
    "    # turn pandas dataframe into array\n",
    "    df_array=np.round(df.values, decimals=3)\n",
    "    \n",
    "    # get the fisher vector for each video sequence\n",
    "    print('Start fisher encoding for %s' % filename)\n",
    "    fv = np.round(fisher_vector(df_array, gmm), decimals=6)\n",
    "    \n",
    "    fisher_vec=np.vstack([fisher_vec, fv]) if fisher_vec.size else fv\n",
    "# save the fisher encoded features for the whole dataset\n",
    "print(\"Saving the fisher encoded features into a csv file...\")\n",
    "print(fisher_vec.shape)\n",
    "fv_storage= pd.DataFrame(fisher_vec)\n",
    "fv_storage.to_csv(\"fisher_vectors.csv\", index=False)\n",
    "print(\"Done!\")\n",
    "print(\"Saving all the labels into a csv file...\")\n",
    "df_label= pd.DataFrame(all_label)\n",
    "df_label.to_csv(\"all_labels.csv\", index=False)\n",
    "print(\"Done!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1 A glance of fisher encoded DT feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>109174</th>\n",
       "      <th>109175</th>\n",
       "      <th>109176</th>\n",
       "      <th>109177</th>\n",
       "      <th>109178</th>\n",
       "      <th>109179</th>\n",
       "      <th>109180</th>\n",
       "      <th>109181</th>\n",
       "      <th>109182</th>\n",
       "      <th>109183</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.005660</td>\n",
       "      <td>-0.011786</td>\n",
       "      <td>-0.014118</td>\n",
       "      <td>-0.006284</td>\n",
       "      <td>-0.003553</td>\n",
       "      <td>-0.005048</td>\n",
       "      <td>-0.005099</td>\n",
       "      <td>-0.006175</td>\n",
       "      <td>-0.004669</td>\n",
       "      <td>-0.013202</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.005618</td>\n",
       "      <td>-0.011514</td>\n",
       "      <td>-0.014186</td>\n",
       "      <td>-0.006267</td>\n",
       "      <td>-0.003553</td>\n",
       "      <td>-0.003380</td>\n",
       "      <td>-0.005049</td>\n",
       "      <td>-0.006169</td>\n",
       "      <td>-0.004669</td>\n",
       "      <td>-0.013199</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.005381</td>\n",
       "      <td>0.027035</td>\n",
       "      <td>-0.008809</td>\n",
       "      <td>-0.004444</td>\n",
       "      <td>-0.003553</td>\n",
       "      <td>-0.004703</td>\n",
       "      <td>-0.002357</td>\n",
       "      <td>0.034120</td>\n",
       "      <td>-0.004669</td>\n",
       "      <td>-0.011317</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.000033</td>\n",
       "      <td>-0.000045</td>\n",
       "      <td>0.000073</td>\n",
       "      <td>0.000007</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0.000017</td>\n",
       "      <td>0.000100</td>\n",
       "      <td>0.000085</td>\n",
       "      <td>-0.000007</td>\n",
       "      <td>0.000024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.005660</td>\n",
       "      <td>-0.006885</td>\n",
       "      <td>-0.013874</td>\n",
       "      <td>-0.005545</td>\n",
       "      <td>-0.003553</td>\n",
       "      <td>-0.005309</td>\n",
       "      <td>-0.005071</td>\n",
       "      <td>-0.004592</td>\n",
       "      <td>-0.004641</td>\n",
       "      <td>-0.003102</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000117</td>\n",
       "      <td>0.000162</td>\n",
       "      <td>0.000260</td>\n",
       "      <td>0.000236</td>\n",
       "      <td>0.000272</td>\n",
       "      <td>0.000201</td>\n",
       "      <td>0.000194</td>\n",
       "      <td>0.000229</td>\n",
       "      <td>0.000218</td>\n",
       "      <td>0.000239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.005660</td>\n",
       "      <td>0.082065</td>\n",
       "      <td>-0.014144</td>\n",
       "      <td>-0.003994</td>\n",
       "      <td>-0.003530</td>\n",
       "      <td>-0.005299</td>\n",
       "      <td>-0.004885</td>\n",
       "      <td>-0.001799</td>\n",
       "      <td>-0.004669</td>\n",
       "      <td>-0.012595</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000282</td>\n",
       "      <td>0.000167</td>\n",
       "      <td>0.000062</td>\n",
       "      <td>0.000103</td>\n",
       "      <td>0.000147</td>\n",
       "      <td>0.000128</td>\n",
       "      <td>0.000111</td>\n",
       "      <td>0.000106</td>\n",
       "      <td>0.000140</td>\n",
       "      <td>0.000109</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 109184 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          0         1         2         3         4         5         6  \\\n",
       "0 -0.005660 -0.011786 -0.014118 -0.006284 -0.003553 -0.005048 -0.005099   \n",
       "1 -0.005618 -0.011514 -0.014186 -0.006267 -0.003553 -0.003380 -0.005049   \n",
       "2 -0.005381  0.027035 -0.008809 -0.004444 -0.003553 -0.004703 -0.002357   \n",
       "3 -0.005660 -0.006885 -0.013874 -0.005545 -0.003553 -0.005309 -0.005071   \n",
       "4 -0.005660  0.082065 -0.014144 -0.003994 -0.003530 -0.005299 -0.004885   \n",
       "\n",
       "          7         8         9    ...       109174    109175    109176  \\\n",
       "0 -0.006175 -0.004669 -0.013202    ...     0.000000  0.000000 -0.000000   \n",
       "1 -0.006169 -0.004669 -0.013199    ...     0.000000  0.000000  0.000000   \n",
       "2  0.034120 -0.004669 -0.011317    ...    -0.000033 -0.000045  0.000073   \n",
       "3 -0.004592 -0.004641 -0.003102    ...     0.000117  0.000162  0.000260   \n",
       "4 -0.001799 -0.004669 -0.012595    ...     0.000282  0.000167  0.000062   \n",
       "\n",
       "     109177    109178    109179    109180    109181    109182    109183  \n",
       "0  0.000000  0.000000  0.000000  0.000000 -0.000000 -0.000000 -0.000000  \n",
       "1  0.000000 -0.000000 -0.000000  0.000000 -0.000000 -0.000000 -0.000000  \n",
       "2  0.000007  0.000005  0.000017  0.000100  0.000085 -0.000007  0.000024  \n",
       "3  0.000236  0.000272  0.000201  0.000194  0.000229  0.000218  0.000239  \n",
       "4  0.000103  0.000147  0.000128  0.000111  0.000106  0.000140  0.000109  \n",
       "\n",
       "[5 rows x 109184 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "fv_info = pd.read_csv('fisher_vectors.csv')\n",
    "fv_info.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4 Feature space visualization\n",
    "We use t-SNE to visualize the feature space for the iDT features extracted from the iCetanaPrivate dataset. We use the labels of the videos to color each fisher encoded representation in this feature space for further analysis.\n",
    "\n",
    "First, defining the mapping between labels and colors, as well as the relationship between labels and dynamic environment issues."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def colormap(c):\n",
    "    if c==0:\n",
    "        color='r'\n",
    "    elif c==1:\n",
    "        color='g'\n",
    "    elif c==2:\n",
    "        color='b'\n",
    "    elif c==3:\n",
    "        color='c'\n",
    "    elif c==4:\n",
    "        color='m'\n",
    "    elif c==5:\n",
    "        color='k'\n",
    "    return color\n",
    "\n",
    "def labelmap(c):\n",
    "    if c==0:\n",
    "        label='normal video'\n",
    "    elif c==1:\n",
    "        label='illumination'\n",
    "    elif c==2:\n",
    "        label='noise video'\n",
    "    elif c==3:\n",
    "        label='rain'\n",
    "    elif c==4:\n",
    "        label='camera shake'\n",
    "    elif c==5:\n",
    "        label='tree wave'\n",
    "    return label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# power and L2 normalization for fisher vector\n",
    "def power_normalize(xx, alpha=0.5):\n",
    "    \"\"\"Computes a alpha-power normalization for the matrix xx.\"\"\"\n",
    "    return np.sign(xx) * np.abs(xx) ** alpha\n",
    " \n",
    "def L2_normalize(xx):\n",
    "    \"\"\"L2-normalizes each row of the data xx.\"\"\"\n",
    "    Zx = np.sum(xx * xx, 1)\n",
    "    xx_norm = np.divide(xx, np.sqrt(Zx[:, np.newaxis]))\n",
    "    xx_norm[np.isnan(xx_norm)] = 0\n",
    "    return xx_norm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "fv_info=power_normalize(fv_info, alpha=0.5)\n",
    "fv_info=L2_normalize(fv_info)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.1 t-SNE with PCA applied\n",
    "\n",
    "#### 4.1.1 PCA for feature dimensionality reduction\n",
    "In this part, we keep the feature importance for up to 98% when using the PCA."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(270, 226)\n",
      "Saving the features after applying PCA into a csv file...\n",
      "Done!\n"
     ]
    }
   ],
   "source": [
    "from sklearn.decomposition import PCA\n",
    "import time\n",
    "from sklearn.manifold import TSNE\n",
    "\n",
    "time_start = time.time()\n",
    "\n",
    "pca = PCA(n_components=0.98)\n",
    "pca_result = pca.fit_transform(fv_info.values)\n",
    "\n",
    "pca_features=pca_result\n",
    "\n",
    "print(pca_features.shape)\n",
    "\n",
    "print(\"Saving the features after applying PCA into a csv file...\")\n",
    "pca= pd.DataFrame(pca_features)\n",
    "pca.to_csv(\"iCetana_FV_PCA_features_DT.csv\", index=False)\n",
    "print(\"Done!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[t-SNE] Computing 46 nearest neighbors...\n",
      "[t-SNE] Indexed 270 samples in 0.001s...\n",
      "[t-SNE] Computed neighbors for 270 samples in 0.028s...\n",
      "[t-SNE] Computed conditional probabilities for sample 270 / 270\n",
      "[t-SNE] Mean sigma: 0.401712\n",
      "[t-SNE] KL divergence after 250 iterations with early exaggeration: 72.762047\n",
      "[t-SNE] Error after 2250 iterations: 0.566349\n",
      "t-SNE done! Time elapsed: 8.701471090316772 seconds\n"
     ]
    }
   ],
   "source": [
    "pca_tsne = TSNE(n_components=2, verbose=1, perplexity=15, n_iter=10000)\n",
    "pca_tsne_results = pca_tsne.fit_transform(pca_features)\n",
    "\n",
    "print('t-SNE done! Time elapsed: {} seconds'.format(time.time()-time_start))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_label = pd.read_csv('all_labels.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.1.2 t-SNE space visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = (15,8)\n",
    "x_tsne=pca_tsne_results[:, 0]\n",
    "y_tsne=pca_tsne_results[:, 1]\n",
    "\n",
    "labels=df_label.values\n",
    "#print(labels)\n",
    "\n",
    "x=np.c_[x_tsne]\n",
    "y=np.c_[y_tsne]\n",
    "\n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "plt.title('t-SNE dimensions (PCA) colored by dynamic environment', fontsize=20)\n",
    "\n",
    "seen=set()\n",
    "for i, j, c in zip(x, y, labels):\n",
    "    c=int(c)\n",
    "    plt.scatter(i, j, c=colormap(c), marker=\"o\", cmap='jet', label=labelmap(c) if c not in seen else None)\n",
    "    seen.add(c)\n",
    "    #plt.annotate(c, (i,j))\n",
    "plt.legend()    \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def colormap_h(c):\n",
    "    if c==0:\n",
    "        color='r'\n",
    "    else:\n",
    "        color='g'\n",
    "\n",
    "    return color\n",
    "\n",
    "def labelmap_h(c):\n",
    "    if c==0:\n",
    "        label='no human motion'\n",
    "    else:\n",
    "        label='human motion detected'\n",
    "    return label\n",
    "\n",
    "df_label = pd.read_csv('human_labels.csv')\n",
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = (15,8)\n",
    "x_tsne=pca_tsne_results[:, 0]\n",
    "y_tsne=pca_tsne_results[:, 1]\n",
    "\n",
    "labels=df_label.values\n",
    "#print(labels)\n",
    "\n",
    "x=np.c_[x_tsne]\n",
    "y=np.c_[y_tsne]\n",
    "\n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "plt.title('t-SNE dimensions (PCA) colored by dynamic environment', fontsize=20)\n",
    "\n",
    "seen=set()\n",
    "for i, j, c in zip(x, y, labels):\n",
    "    c=int(c)\n",
    "    plt.scatter(i, j, c=colormap_h(c), marker=\"o\", cmap='jet', label=labelmap_h(c) if c not in seen else None)\n",
    "    seen.add(c)\n",
    "    #plt.annotate(c, (i,j))\n",
    "plt.legend()    \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def colormap_all(c):\n",
    "    if c==0:\n",
    "        color='silver'\n",
    "        # color='w'\n",
    "    elif c==1:\n",
    "        color='black'\n",
    "        # color='w'\n",
    "    elif c==10:\n",
    "        color='fuchsia'\n",
    "    elif c==11:\n",
    "        color='red'\n",
    "    elif c==20:\n",
    "        color='paleturquoise'\n",
    "        # color='w'\n",
    "    elif c==21:\n",
    "        color='darkcyan'\n",
    "        # color='w'\n",
    "    elif c==30:\n",
    "        color='hotpink'\n",
    "    elif c==31:\n",
    "        color='pink'\n",
    "    elif c==40:\n",
    "        color='cornflowerblue'\n",
    "    elif c==41:\n",
    "        color='b'\n",
    "    elif c==50:\n",
    "        color='greenyellow'\n",
    "    elif c==51:\n",
    "        color='limegreen'\n",
    "    return color\n",
    "\n",
    "def labelmap_all(c):\n",
    "    if c==0:\n",
    "        label='normal video_N'\n",
    "    elif c==1:\n",
    "        label='normal video_Y'\n",
    "    elif c==10:\n",
    "        label='illumination_N'\n",
    "    elif c==11:\n",
    "        label='illumination_Y'\n",
    "    elif c==20:\n",
    "        label='noise video_N'\n",
    "    elif c==21:\n",
    "        label='noise video_Y'\n",
    "    elif c==30:\n",
    "        label='rain_N'\n",
    "    elif c==31:\n",
    "        label='rain_Y'\n",
    "    elif c==40:\n",
    "        label='camera shake_N'\n",
    "    elif c==41:\n",
    "        label='camera shake_Y'\n",
    "    elif c==50:\n",
    "        label='tree wave_N'\n",
    "    elif c==51:\n",
    "        label='tree wave_Y'\n",
    "    return label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_h_e_label = pd.read_csv('human_env_labels.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1oAAAHZCAYAAACfGUrBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xl8VNXd+PHPmYCBhCUGK7W2kGBFQkgIkGiQLT4IUgUsIsKPFEwRUBahrlBjC7SNSsUiVqlCla2xQkEUa3lKFUFQIxCIwYAGbRIUeeICAUJIJOT8/jgzQzKZmcyWyfZ9v155Tebc5Zy7zMz93rNcpbVGCCGEEEIIIUTgWBq6AEIIIYQQQgjR3EigJYQQQgghhBABJoGWEEIIIYQQQgSYBFpCCCGEEEIIEWASaAkhhBBCCCFEgEmgJYQQQgghhBABJoGWEEIIIYQQQgSYBFpCBJlSSiuldjikLbSmpzRMqYJPKVWolCps6HL4Sil1tVJqs1Lq/6zHrqSByvFDpdQapdSXSqkL1rJENERZmiKl1HCl1PtKqZPWffdaEPJcbc0rqr7z8pZS6gZr2cY5pDfpz2tL0xJ/UzyhlBpr3S9DG7osomVo1dAFEM2X9SKiAFijtU7zcR3jgLuAvsClwGmgGNgP/EdrvcZJfgB7gGTt5IncSikNHNNa/9jFsu5Ea60Lvd8S0ZwopUKA14CfAuuAL4HyBirOamA48HfgM0AHqyxKqdXAnTTRz4X1c/86UAKswny/fNKARWpQSikLsBT4CNjYwMVpcpRSzwH3AD/UWn/T0OURTr2KuX74k1Kqj9a6qqELJJo3CbREo6WUWgFMA84Bb2ICoXCgGzAKSAHWuFj8WmA88IqX2Z4CnnYzvb5qLZ7FlPVoPa2/MWrKdxSjgZ7ASq319IYqhFLqEmAY8JbWOrWhytGE3Qi0AR7QWr8cxHx/DTwBHAtinp6YAPQGUp3dpBKuKaUUcCuwW4KsxktrrZVSi4H1mPM9mJ970QJJoCUaJaXUAEyQ9SXQX2v9pcP01phAy5mjwA+Bx5RSr2qtv/ci6xKt9ULvS+wfrfW3wLfBzrchaa0/b+gy+OFH1tevGrQU5jy3NIJyNFUNchy11seB48HM00OzMLV6mxu6IE3QtcCVwJKGLoiok60WeyYSaIl6Jn20RL1QSi3kYlO8O61tom1/aR6sYoD1dZNjkAWgtT6vtf6Pi2W/AP6CqXW417uSB4ZS6hKl1G+UUp8rpSqUUgVKqT8opUJdzO+0Pb2tP5dSqrNS6iWlVLFS6qy1T8kg6zzhSqknlVJF1rzyHPtXOKzz/yml3rH2SSlXSh1WSj3qrGzV8r9MKbVCKXW8Wh6/dDK/UkrdaS3fN9b1f6GU+rdSarzDvE77fCilQpVS85VSuUqpMqXUaaXULqXUHU7mjbKWcbX1/1eUUt9a892nlBrpZJlLlFJzlFL7rfugzFqW15VSN7rab9X3CbDT+nZBtfN6YbV5OiqlHldKfWoty0nrPqi1fqVUim15pdS1Sqk3lVInVB19eKz7rsj6tvpnbLXDfN4c758rpf6mlMq3nmelSqls6/6yOMyrMc0GAQqq5V9YvYzOjrF1Wl3n/A+VUn9VSh1Tpu9ZWrV5wpRSv1ZK5VQr5wdKqf/nan855JFiLf8ia9I71cqf4kfZByml3lCmv1yFMv33spRSCxzmc9lHSyl1h1LqXaXUKaXUOaXUQeu2Ojtehda/MGW+A45a8/1MKTVPKaU82R/WdfUArge2aK3PuZnPo7yqn9cu1lNr/yql0qzLpCmlhinzuS9V5rtklbL2PVRK9VFK/dN6Tpcqpba42Jf9lFLLlFIfWT9T5UqpI0qpp5RSlzqZv3r+N1jPwzPKfAe9qZSKcbMLb7O+2vv4+Xp8vDkHPKGUul0ptUeZ77oTynxPXulkPn8+r37/RinzvfmQUmq79TP0vfXYb1FKJbsol1e/UQBa6wrMcRpgPe+FqDdSoyXqyw4gApiLae9fvYN5jgfLf2d97e5j/r/DXASmK6VWaa1P+Lger1l/PDdgmpF8jmkWeAkwBYjzYZURwHvAGUw/nEhMk4d/K6X6Ay9Y0/4JtAb+H7BeKfWF1jrLoWwvWsvxJaategmQDPweGKqUGqa1rnSR//eYfhttgNuBl5RSVdX7yQEZmGZRBdZ9cAq4AkgCxmGaa7ikTFO4fwNDMH1lngPCrPmtV0olaK0fcbJoV0y/vP9i+kxFYpqOvq6UulFr/U61eVdb99HHwFpM09QfAQOBEcBb7sqIuTiPwpxfOzHnOrZX68Xge5imhXsxTVEvA+4AtimlZmitX3Cy3v6YfbcbeMm6jLva2Ket5XD8jNk/Xz4c7yeAKuBDTLO2jsD/AMswx3CSw374Oaap2TIuNqsNRPPaSCALKLWWuwrTN9O2f7cDfTB9LV7C3DS8CXhZKRWrtX60jvUXWsufgjnX1ljTqPbqFaXUCEwT59PAFsz+iwRiMHfOF7le2r6OxzDnwLeYO+2lwM+Ax4CbrMfrvMNirYFtmHN4K1CJOS5PYD6rdeZrZbsJsNvNPIHKqy6jgZGY77TnMQFgGhCtlJoPvA3sAl7EfKeOAq5SSsU59LmZBozBfE7fAkIw/X3vB36mlLpOa33GSf4jMd/fW6359wRuBpKUUj2tLRAcjQH2O+mr6NU+8/EccGcmZn9uweyH6zDfjb2t36cVXqzLlUD9RsVgfkPexXyWTgJdrOX/mVJqlNb6f93k78lvlM17mHPqRlpwv0wRBFpr+ZO/evnDXARqYLUPy16JuWDTmB+IicDVgPIgv93W9w9a3//JYT4NfOli2RJgoYu/ezws+0Truj4A2lRLj8QEXhrY4bDMQmt6ipOyasyPvaVa+iRr+gngDYd8BlmnbXZYV5o1/VWgrYv857rI/69ASLX0npgLhkMO83+HuagPc7JfLnN4XwgUOqT92prfv4BW1dIvt86vgeudHDcNLHBY1022dVVL64i5aN9XfXuqTe/k4TFOsa57oZNpL1invVD9fMWcv6eACiDKybo0cHegPmM+Hu+rnKzHgglENHCdw7TV1vQoF+WrdYy9OOfXVj8HnOT5sEN6G+B/rcc3wcP957QMvpQd2GRN6+3BuV9rv2ECbY216XO19FaYz7gGHnFSRtvnpW219Msx32UlQGsP98Ur1nX1c3MsPc4LN58RV/u32jlbCQxxOAf/w8XvvFSH5V60TrvVIb0rzj/nd1nnn+cm/6EO0x53dt5Zp8Vapz3q5z7z+hzw4Nw+DcQ5THvZOu2OAH5eA/Eb1RGHz4o1/ceY5r2HnUzz6jeq2vTe1uU2eLI/5U/+fP1r8ALIX/P9w49Ay7r8DVwcRc32dxpzMfULxx9QagdaoZialQqgW7X5NK4DLXd/OR6W23ZBcIOTaWl4H2idBdo7pIcA563TuznJpwAocEg7YF0mwsn8IZg7qHtc5N/ByTI7rdPbV0v7zpp3qAf7qZDaF1pHMBfKPZzMb7s4esnJcSt0PB+s04uAb6u972Cd/z3cBO0elD0FJxeRmLu1ZzF3diOdLPd763K/dbKuA4H8jPlyvN3k09ex3Nb01dRPoFUBXO5kmU6Yi6e9LtZpu3j6o4fb5bQMvpSdi4FWdw/yrbXfgJXWtOlO5u8OXAD+66SMGvipk2VswXEvD/fF+9b5f+TmWHqcl6vPiLv9y8Xvx3VO5p9snfauk2lDcHKzxc22KsxNj+0u8v+bk2WirdM2Opn2qHVarJ/7zOtzwINz+w9Opt1gnbbEn3Pemh6w36g6tucZ67q6uMjfo9+oatM6W6dleVoG+ZM/X/6k6aBoENbmP79yMulprXUJgNb6HaVUd0x/rSGYpkIDMDUVN2H6pYzULpo+aK0rlFKPYO7ePYFpulWXIq11lLfb46AvJlhw1gRnhw/ry9cOzVu01heUUsVAuNb6v06WOYZpIgKYfhWYi9BvgV+56BpQgWm64eiI1vq0k/QvrK8RmMACIBPTLy5PKfUPzA/dB1rrU642rloZ22OGSz+mtXbWlGO79bWPk2k5WusLLsrY3/ZGa31aKfUGpqlRjlJqE6YJ0oda67K6yuiBHpimju9p581Vt2Muypxtw54A5A/4fryVUp2AhzDNpLphRvmsrla/jnpSqLX+2kl6EuYCzlXfn9bWV3d9aepLJqafzodKqfXAO5jzoFYfUxf6Wl+3O07QWucrpb7ENJ2LsH1HWp3SWn/mZH22z2etvkgudLK+nnQzT6Dyqss+J2m2AUuynUyzjd744+qJygyadDemGVtPTI1J9b6Grs5nZ/m728bbMN+TeU6mebPPfD0H3PF2W3zh92+UjTIDYc3FfG9fjml2X92V1B6d15vfKBvb9/NlTpYTImAk0BINJQJY4CR9NdX6eGjT3n6X9c/W/2kY5m7gjcAM3A/H/gpwHzBOKZWsHfos1ZOOwAntvB39//mwPldBSmUd06p/vi/F3MX9Ac73uzuuftBtfXtCqqXdh2keOQWYb/2rVEr9CzOEtrMLDpuO1ldXo7HZ0p09jNddGR0H/RkPzMM08bT1jyhXSm0EHtRaF7spY1382QZfzg1XvD7e1psfezF37vdgmu6dwOxDW39Lnzrj+8DVvrAFA0nWP1faBbY4ddNav6rM4CsPYM7/uwGUUtnAr7XrwXtsPDl3uljnq36+e/P5dMc2AEabav87ClRedXH2vVbpwbTWDunrMX2n/osZae7/MDcYwNzoc3U+19pOrXWl9YZFjW1USnXF3DhZ7Om6HMpcfX2+ngPuOJsvGMfLlo+nv1EopcZg+liVY1qGfI6prariYn9KZ8fMl/OyrfXV5cAvQgSCBFqiQWjTYdjjEbGqLacxAwo8immT/T+4CbS01lop9SCmZmUJZsCD+nYKiFRKtXYSbP0wCPk7Y/uxO6C17ut2Tj9Ya5WWAcuUUpdj9vcEzEAYsdaBClx1vraV0dU+usJhPl/LeA5rvzul1E+AwZgmQ7/ANMUb5Mfq/dkG7Ue+rsrhzfGeigmyFmmHRxxYO7TP9aEcVdS+I23jLNi0cbUvbNu1VGt9vw/l8YbXZddavwm8qZQKx9ytH4m5GfRPZR6OeshNftXPHWePPgjI+e+GrQaxE+5rtTxlG5TC1XVGR+pvW1BKJWKCrLeAm6t/FyszgubDAcpqjPU1EEPiN/Q54OvnNVB+jxnQIlFrfbj6BKXUC5hAK1BsN22c1ZwLETAyvLuoT7amXIG6a1adrRlAncGa1vpdzN3MAUqpsfVQFkf7MZ8tZ0FdShDyr0VrXQrkYYKdyCDl+bXW+lWt9R2YpjBXAb3czH8Gc3FxpVLqaiez3GB93R/AMn6htc7ENEU9Agy0Np/z1adAGZCgnAwfTT1sgzM+Hu+fWl83OZnm6gKnrs/4SaCztQmXo0QPy1XdHszFoD/BsKd8LrvW+qzWers1GHwMc/H6szryO2B9TXGcoJT6KaZZXIEXTca8lWt9DdRw17Zg7SeOE6zbU98X7rbzeYuTG17XcrFGw19jMM0aA9H0t6HPgUB/Xr31U8zgFY5BlqvfU3/YznNPRkEWwmcSaIn6dBJr51VvF1RKjVBK3ebsC18p1Y6L/bve9XCVD2OaETzhbVl8sMr6mqGUamNLtF7w1jXsdH36E+aC7yVrM7EalFKXKqV8ru1S5vlXQ5Wq9Tyd1pgRF8EEIe68hAmen1RK2S/elVKXAb+pNo+vZfyBUqpWvwBMX6T2mHPEmwdc16DNw7EzMU3XfueQ91XAHEwH8XW+5uEFb493ofU1xWG+PpjRIJ2xPYbB1Wd8D6ZGo8bzbJR5JtYAZwu4Y+23lQkkKvOculq1JUqpq5RS0d6u2wmvym49951dvHe2vnpy7gM8qpT6QbX1hmBq4y2Y0fXqyw7rq9PnFfngE8zgRbdaa7cBsO6jZwKUhzuF1teU6onWsjwXiAysx2kA8Jq1tYW/GvocCOjn1QeFwNVKKduDxG3dBRZg+tgFku08fyfA6xWiBmk6KOqN1rpUKfUhMEgplQnkY+6Ab9Fa57pfmh7AUuCkUmoXprahEnNH7xbM3dAPMc+o8qQs+UqpFZhnirgT4aKTvc1qXfs5KY7+jukHNBr4WCn1OqbvwO2YPjBXeVLmQNNav6SU6ofZB58rpf6N6VQciWkyNhgTJN7jYxZtMc10Cq3HvQjT32MYZnCCLY53Kp1YgrnzfyvwkbVvVxim6eHlmNHk3D3npy5XAllKqcOYWqUvMCMRjsQ013nGsVO3D+ZjalxmK6WSMD/ktudotQdma60L/MyjTj4c77WYgTCeVkrdgPnMXY3ZN69izmlHb1uXWWnt41YKlGitbZ/LP2Mu2v6ilBqK2d+9Mc9F+qd13d6abS3X74BJSqndmGds/QhzniVhntPj7z72tuxPAVFKqR2YC8bvgX6Y5s1FmP6iLmmt31dK/RFzU+hj6/48i/k89MIMrvOkn9vkznZMX5ebCMANIa31eaXUMswNkgNKqc2Ya45hmBqgr9wtHwB7MaOL3qaUeh+z/zpj9uenAcr/Vkxt7qsBWFdjOAfq4/PqjaWYYeIPWAcqOo8J8HpihogfFcC8hmPO91oDjwgRSBJoifo2CfPlOQJz8aMwz1mqK9D6G+Zu6DDMF/1gTC1BCaaq/x/AX601CJ5aiOmH08HNPB1xP3jADup4oKm1X9g4zAV3GubC8DjmovZ3mI6+DUJrPUsptRVzcX0jJmA9gbkAfxKz3311FjPIxA2YH+afY5p4fo7pp1JnTZTW+nul1DDMA0UnYkYwrMQ8kPdXWuu/+1E+MMduAeYu9w2YAOgE5sJrPnVcDHtCa33C2qfp15jRyO7HdLjeAzyptd7mbx5elMXj4621/kopNQhT6zsQc8H9CSZQewsngZbW+t9KqQcwD4a9D1ODVoT1BojW+pBS6kZM87lRmGO5CzOi2G34cOFmHTlyCDAdc46MxQT0xZjg8D5MR3q/+FD2xzDNyBIx+7oKs58fw4ymWme/J631PKXUAcx3xmTMDZrPMYHPU15+33lFa12mlFqNGaUyxoObIp5YgKnJm4Y5Xv+H+YwtBNz1V/ObddS70cAfMKNozsGMdPdXa1og8h+DabmxMwDrAhr8HAj459XL/F9QSlVgWqzcifne3IUJ/sYSoEBLmdGMk4FlARptVgiXVGBqu4UQQgjRlCmlojDB9Qtaa18GP2kxlHkcxTfAeq31nQ1dHuE5pdRTmEA2xsXQ80IEjPTREkIIIYRtNNhngOlKqWA9M62puhkz1HggRhsUQaKUugLTwuLPEmSJYJAaLSGEEEIAoJTqgGl++ZbW+r2GLo8QgWRt1j0c02ywvkZvFMJOAi0hhBBCCCGECDCvBsO47LLLdFRUVD0VRQghhBBCCCEat+zs7G+11j+oaz6vAq2oqCj27dvne6mEEEIIIYQQoglTShV5Mp8MhiGEEEIIIYQQASaBlhBCCCGEEEIEmARaQgghhBBCCBFgXvXRcub8+fN8+eWXlJeXB6I8opFo06YNP/7xj2ndunVDF0UIIYQQQogmx+9A68svv6R9+/ZERUWhlApEmUQD01rz3Xff8eWXXxIdHd3QxRFCCCGEEKLJ8bvpYHl5OZ06dZIgqxlRStGpUyeppRRCCCGEEMJHAemjJUFW8yPHVAghhBBCCN/JYBhCCCGEEEIIEWASaAkhhBBCCCFEgEmgFQApKSns27fP4/lvvvlmSkpKaqUvXLiQJUuWBKxcCxcuJCwsjK+//tqe1q5du4CtXwghhBBCCOFc0AOt4uJisrKy2LlzJ1lZWRQXFwe7CDVUVlYGPc9//etfREREBCWvyy67jKeeeiooeQkhhBBCCCGMoAZaxcXF5OfnU1FRAUBFRQX5+fl+BVuFhYXExMQwbdo0YmNjGT58OOfOnQMgJyeH5ORk4uPjGTNmDCdPngRMDdQjjzzCkCFDWLZsGWlpacyYMYMbbriBbt26sXPnTqZMmUJMTAxpaWn2vGbMmEFiYiKxsbEsWLDAbbm2bt3KHXfcYX+/Y8cORo0aBUBUVBTffvstABkZGVxzzTXceOONfPrpp/b5P//8c0aMGEG/fv0YNGgQn3zyCQBFRUUMHTqU+Ph4hg4dytGjR92WY8qUKaxfv54TJ054uEeFEEIIIYQQ/gpqoFVQUEBVVVWNtKqqKgoKCvxa75EjR5g1axZ5eXlERESwadMmACZPnszixYvJzc0lLi6ORYsW2ZcpKSlh586dPPDAAwCcPHmS7du3s3TpUkaNGsV9991HXl4eBw8eJCcnBzBB0b59+8jNzWXnzp3k5ua6LNOwYcPIysri7NmzAKxfv57x48fXmCc7O5tXXnmFAwcO8Oqrr7J37177tOnTp/PnP/+Z7OxslixZwsyZMwGYPXs2kydPJjc3l9TUVObMmeN237Rr144pU6awbNkyT3enEEIIIYQQwk9BDbRsNVmepnsqOjqahIQEAPr160dhYSGnTp2ipKSEIUOGAHDnnXfy7rvv2pdxDHpGjRqFUoq4uDg6d+5MXFwcFouF2NhYCgsLAdiwYQN9+/alT58+5OXlcejQIZdlatWqFSNGjOCNN96gsrKSN998k1tvvbXGPLt27WLMmDGEhYXRoUMHRo8eDUBpaSnvv/8+48aNIyEhgbvvvpvjx48D8MEHHzBx4kQAJk2axO7du+vcP3PmzGHNmjWcPn26znmFEEIIIYQQ/msVzMxCQ0OdBlWhoaF+r9cmJCTE3nTQnfDwcKfrsFgsNdZnsViorKykoKCAJUuWsHfvXi699FLS0tLqfKDv+PHjee6554iMjCQpKYn27dvXmsfZ86qqqqqIiIiw16S548nzriIiIpg4cSLLly+vc14hhBBCCCGE/4JaoxUdHY3FUjNLi8VCdHR0wPPq2LEjl156Kbt27QJg3bp19totX5w+fZrw8HA6duxIcXExW7durXOZlJQU9u/fz8qVK2vVoAEMHjyYzZs3c+7cOc6cOcMbb7wBQIcOHYiOjuYf//gHAFprPvroIwCuv/56XnnlFQAyMzMZOHCgR+W///77eeGFFxpk8A8hhBBCCCFamqAGWp07d6Z79+72GqPQ0FC6d+9O586d6yW/NWvW8NBDDxEfH09OTg6//e1vfV5X79696dOnD7GxsUyZMoUBAwbUuUxISAgjR45k69atjBw5stb0vn37Mn78eBISEhg7diyDBg2yT8vMzOTFF1+kd+/exMbG8vrrrwPwzDPPsGrVKuLj41m3bp3Hfa8uu+wyxowZ43czTSGEEEIIIUTdlNba45kTExO14/OiDh8+TExMTKDLJRoBObZCCCGEEMFRXFxMQUEBFRUVhIaGEh0dXW+VEcI/SqlsrXViXfMFtY+WEEIIIYQQoibbI5Bso3PbHoEESLDVhEmg1QxkZGTY+3PZjBs3jvT09AYqkRBCCCGE8JS7RyA150ArK7+CzVnnOFFaRWQ7C2OS25Lc3b9B8hoTCbSagfT0dAmqhBBCCCGaqPp6BFJjlpVfwbodZ/neOk7bidIq1u0wz59tLsGWBFpCCCGEEEI0oPp6BFJ98rc2anPWOXuQZfN9pUmXQEsIIYQQQjSI5t7kqqWJjo6u0UcL6u8RSIEQiNqoE6VVXqU3RUEd3l0IIYQQQvjHdpFruyC1XeRm5TffZmbNXbAfgeQvd7VRnops5zwMsSiYtvwE89aWNPlzWmq0hBBCCCEaGXc1Vi2hyVVL1LlzZ6eBVWOsvQxEbdSY5LY1asVsqvTFdTX1PltSoxUAKSkpOD5fzJ2bb76ZkpKSWukLFy5kyZIlASnTtm3b6N+/P7bnpF24cIGEhATef//9gKxfCCGEEPWjrhqrltDkShjBqr3Myq9g3toSj2uSXNVGuUp3Jrl7KJNSwu3LWFTtebytJWtsgh5oZWZmEhUVhcViISoqiszMzGAXoYbKysq6Zwqwf/3rX0RERNRrHsOHD6dr1668+OKLAPz5z38mKSmJ66+/vl7zFUIIIYR/6mqWFYiLXNE0BKKJXl18CebGJLflEod2cSEWqDhf5VWzv+TuoSyeHMHKmZH2mixHTfkGQlA/kZmZmUyfPp2ioiK01hQVFTF9+nS/gq3CwkJiYmKYNm0asbGxDB8+nHPnzMmXk5NDcnIy8fHxjBkzhpMnTwKmBuqRRx5hyJAhLFu2jLS0NGbMmMENN9xAt27d2LlzJ1OmTCEmJoa0tDR7XjNmzCAxMZHY2FgWLFjgtlxbt27ljjvusL/fsWMHo0aNAiAqKopvv/0WMM/Auuaaa7jxxhv59NNP7fN//vnnjBgxgn79+jFo0CA++eQTAIqKihg6dCjx8fEMHTqUo0ePuizD0qVLefzxx8nLy+PZZ59l8eLFXuxZIYQQQjSEumqsnF3kXtLKpFfnbS1Fc5OZCVFRYLGY1wa+t++TYNRe+hLMOdZGhbdRoOFsxcXyeVvz1hxvIAS15Onp6ZSVldVIKysr8/sZUEeOHGHWrFnk5eURERHBpk2bAJg8eTKLFy8mNzeXuLg4Fi1aZF+mpKSEnTt38sADDwBw8uRJtm/fztKlSxk1ahT33XcfeXl5HDx4kJycHMAERfv27SM3N5edO3eSm5vrskzDhg0jKyuLs2dN29L169czfvz4GvNkZ2fzyiuvcODAAV599VX27t1rnzZ9+nT+/Oc/k52dzZIlS5g5cyYAs2fPZvLkyeTm5pKamsqcOXNcluGKK67gV7/6Ff379+fRRx8lMjLSm90qhBBCiAZQ1wWn40VuZDsLk1LCa/Rj8baWorkFZZmZMH06FBWB1uZ1+vSmF2wFI/jwNZirXhsV2kpxwaFGytuaN09vIDQlQQ20XNW+uKuV8UR0dDQJCQkA9OvXj8LCQk6dOkVJSQlDhgwB4M477+Tdd9+1L+MY9IwaNQqlFHFxcXTu3Jm4uDgsFguxsbEUFhYCsGHDBvr27UufPn3Iy8vj0KFDLsvUqlUrRowYwRtvvEFlZSVvvvkmt956a415du3axZgxYwgLC6NDhw6MHj0agNLSUt5//33GjRtHQkICd999N8ePHwfggw8+YOLEiQBMmjSJ3bt3u903s2bN4sKFCzVq5oQQQgjbwtWpAAAgAElEQVTReLm74LQFRC++ZW7k3nVjOIsnR9QaLMCbWormOIphejo43NunrMykNyXBCD7cBXOeBuCBqHnz5AZCUxPUUQe7dOlCUVGR03R/VH+YW0hIiL3poDvh4eFO12GxWGqsz2KxUFlZSUFBAUuWLGHv3r1ceumlpKWlUV5e7jaP8ePH89xzzxEZGUlSUhLt27evNY9StXv+VVVVERERYa9Jc8fZ8tVZLJY65xFCCCFE41F9dMHqI80BHj+7yJsL3+Y4iqGre/h+3tsPOlfnQiCPi7PR/y5pBXFdW3l8vkW2szg9t7yteUvuHtpkzzlnglqjlZGRQVhYWI20sLAwMjIyAp5Xx44dufTSS9m1axcA69ats9du+eL06dOEh4fTsWNHiouL2bp1a53LpKSksH//flauXFmrBg1g8ODBbN68mXPnznHmzBneeOMNADp06EB0dDT/+Mc/ANBa89FHHwFw/fXX88orrwCmz9vAgQN93iYhhBBCNE7Vm2XZaqy8qaXypslZcxzF0NU9fD/v7TcIZ+dCoNfvrCbpYFGlx+dbc2z2FwhBrdFKTU0FTF+to0eP0qVLFzIyMuzpgbZmzRruueceysrK6NatG6tWrfJ5Xb1796ZPnz7ExsbSrVs3BgwYUOcyISEhjBw5ktWrV7NmzZpa0/v27cv48eNJSEiga9euDBo0yD4tMzOTGTNm8Ic//IHz588zYcIEevfuzTPPPMOUKVN48skn+cEPfuDXNgkhhBCi6fAmIHJVS+HswjdQtRH1zZvnSWVkmD5Z1ZsPhoWZ9JbE033mrCbJ1jzVkbNzJRg1b02Rsj1nyROJiYna8XlRhw8fJiYmJtDlEo2AHFshhBCi8Zi3tsRlQLR4cu3Hxnh6kW3ro+UYlDWm/jG+lDEz0/TJOnrU1GRlZEA93dtvlPw9rt6eby2JUipba51Y13xBrdESQgghhBC+8aaWCjzv79IUaiN86UeWmtqyAitH/va98/Z8E7VJoNUMZGRk2Ptz2YwbN87vYfOFEEII0XjUZ0DU2AchcNds0lbz0hgDxIbkb9+7phCAN3YSaDUD6enpElQJIYQQLUBjD4jqi6t+ZECtYemh9qh4LVEg+t611PMtUBpXL0chhBBCCCEcOBvVzhlvH5LbnMlIgA1ParSEEEIIIUSj5qwZW3Mclj6QpOlfw5NASwghhBBCNEruRk50NyqeMKTpX8OSQEsIIYQQQjQ6jsOTO/bB8mRUPG+evdUUNffta+qaRch//fXXA1BYWEivXr0A2LFjByNHjgzI+rds2cITTzzh07IlJSUsX77c/v6rr77i9ttvD0i5bKKiohg7dqz9/caNG0lLSwtoHkIIIYQQweRueHIwwdaklHB7DVZkO0uNZ0TZAjXHwTKy8iuCtxEeyMqvYN7aEqYtP8G8tSUel6+pbF9LFvxAKxOIsuYcZX3vp/fff9//lbgxevRo5s+f79OyjoHWj370IzZu3Biootnt27ePvLy8gK9XCCGEEKIheNIHK7l7KIsnR7ByZiSLJ0fUqM2pK1BrDPwJlprC9rV0wQ20MoHpQBGgra/T8TvYateundvpCxcuZMmSJfb3vXr1orCwkMLCQnr06MHUqVPp1asXqampvPXWWwwYMICrr76aPXv2ALB69Wpmz54NQFpaGnPmzOH666+nW7du9qCptLSUoUOH0rdvX+Li4nj99dcBmD9/Pp9//jkJCQk89NBDNWrdysvL+eUvf0lcXBx9+vThnXfesed32223MWLECK6++moefvjhOvfBgw8+yGOPPeblnhNCCCGEaJxc9bXytA9WUxgsw59gqSlsX0sX3EArHShzSCuzpjeQzz77jLlz55Kbm8snn3zCyy+/zO7du1myZInLwOX48ePs3r2bf/7zn/aarjZt2rB582b279/PO++8wwMPPIDWmieeeIKrrrqKnJwcnnzyyRrree655wA4ePAgf//737nzzjspLy8HICcnh/Xr13Pw4EHWr1/PF1984XY77rjjDvbv389nn33m7y4RQgghhGhw/g5P7m+gFgz+BEtNYftauuAeiaNepgdBdHQ0cXFxWCwWYmNjGTp0KEop4uLiKCwsdLrMz3/+cywWCz179qS4uBgArTWPPPII8fHx3HjjjRw7dsw+zZXdu3czadIkAHr06EHXrl3Jz88HYOjQoXTs2JE2bdrQs2dPioqK3K4rJCSEhx56iMcff9zLPSCEEEII0fjU1QerLk3hOVL+BEtjktsS4jBbiKVxbV9LF9xRB7tgmgs6S69HrVq1oqrq4p0BW60RQGjoxQ+rxWKxv7dYLFRWOtTlOllGaw1AZmYm33zzDdnZ2bRu3ZqoqKga+ThjW7auPEJCQlyWpbpJkybx+OOPExsbW+e8QgghhBCNnT/DkzeF50h5MnKiW46Xkq4vLUUDCG6NVgYQ5pAWZk2vR1FRUezfvx+A/fv3U1BQEPA8Tp06xeWXX07r1q1555137DVQ7du358yZM06XGTx4MJmZpoNafn4+R48e5ZprrvG5DK1bt+a+++7j6aef9nkdQgghhBDNhbvBMhoDf2rtNmed44JDYHVBy2AYjUlwa7RSra/pmOaCXTBBVqrLJQJi7NixrF27loSEBJKSkujevXvA80hNTWXUqFEkJiaSkJBAjx49AOjUqRMDBgygV69e/OxnP2PWrFn2ZWbOnMk999xDXFwcrVq1YvXq1TVqsnxx11138Yc//MGvdQghhBBCiODwtdZOBsNo/JS75muOEhMT9b59+2qkHT58mJiYmECXSzQCcmyFEEII0Vi0tIfz1rW989aWOA2qIttZWDw5IphFbXGUUtla68S65pNhSYQQQgghRKPW0h7O68n2NoXBPlq64DYdFH657rrrqKio+YWybt064uLiGqhEQgghhBD1z93zpppjrZYn2xuowT5aWk1hMEmg1YR8+OGHDV0EIYQQQoigc9cfadryE80uQPC0/5U/ozLCxZozW1BnqzmzrVv4R5oOCiGEEEKIRq2u50o1t6aEwXoYsbuaM+E/CbSEEEIIIUSj5qw/kqOmECBk5Vcwb20J05afYN7aEpeBYbD6X8nIhfVLmg4KIYQQQohGzbE/kiuNOUDwppleXf2vAtWvKrKdxeXIhcJ/EmgJIYQQQohGr3p/JHdDmzdW3g7o4ar/VSD7VY1JbltjXSAjFwZS4z0bvXD99dcDUFhYSK9evQDYsWMHI0eODMj6t2zZwhNPPOHTsiUlJSxfvtz+/quvvuL2228PSLkAJk6cyF/+8hf7+w8//JD4+HgqKyvdLCWEEEII0XQ1xaHNA9VML5D9qpK7hzIpJdweoEa2szApJVwGwgiQ4AdamZkQFQUWi3nNzPR7le+//77f63Bn9OjRzJ8/36dlHQOtH/3oR2zcuDFQRWPp0qU8+eSTfPPNN1RVVTF79myWL19Oq1ZSWSmEEEKI5qkpBgiBGuAi0P2qkruHsnhyBCtnRrJ4ckSj3odNTXADrcxMmD4diopAa/M6fbrfwVa7du3cTl+4cCFLliyxv+/VqxeFhYUUFhbSo0cPpk6dSq9evUhNTeWtt95iwIABXH311ezZsweA1atXM3v2bADS0tKYM2cO119/Pd26dbMHTaWlpQwdOpS+ffsSFxfH66+/DsD8+fP5/PPPSUhI4KGHHqpR61ZeXs4vf/lL4uLi6NOnD++88449v9tuu40RI0Zw9dVX8/DDD7vcts6dO/Pggw/y8MMP8/zzzxMfH8/AgQN93JNCCCGEEE1DUwsQAlULF95GOU1vzM0mW6rgVnukp0NZWc20sjKTnpoa1KLYfPbZZ/zjH/9gxYoVJCUl8fLLL7N79262bNnCY489xmuvvVZrmePHj7N7924++eQTRo8eze23306bNm3YvHkzHTp04NtvvyU5OZnRo0fzxBNP8PHHH5OTkwOY5o02zz33HAAHDx7kk08+Yfjw4eTn5wOQk5PDgQMHCA0N5ZprruHee+/lJz/5idNtuOeee1izZg07duxg3759Ad5DQgghhBDCX4F4wHBWfgXlFbpWeoilcTebbKmCG2gdPepdehBER0cTFxcHQGxsLEOHDkUpRVxcXI2gqLqf//znWCwWevbsSXFxMQBaax555BHeffddLBYLx44ds09zZffu3dx7770A9OjRg65du9oDraFDh9KxY0cAevbsSVFRkctAy2KxcPfdd7Nv3z46derk9T4QQgghhBCBG83PFX8fMLw56xwXasdZtGktDxhujIIbaHXpYpoLOkuvR61ataKq6mK71fLycvv/oaEXT0qLxWJ/b7FYXA4oUX0Zrc3ZnpmZyTfffEN2djatW7cmKiqqRj7O2JatK4+QkJA6B7ewWCxYLFJlLIQQQviqvi+yReMWyNH86ourflhnm8dzmpud4F6ZZ2RAWFjNtLAwk16PoqKi2L9/PwD79++noKAg4HmcOnWKyy+/nNatW/POO+9QZA0o27dvz5kzZ5wuM3jwYDKt/dPy8/M5evQo11xzTcDLJoQQQgj3bBfZtgtZ20W2qwfKiuYnkKP51ZdADaghgiO4RyU1FVasgK5dQSnzumJFvffPGjt2LCdOnCAhIYG//OUvdO/ePeB5pKamsm/fPhITE8nMzKRHjx4AdOrUiQEDBtCrVy8eeuihGsvMnDmTCxcuEBcXx/jx41m9enWNmiwhhBBCBEdDX2Rn5Vcwb20J05afYN7aEgnwGkCgR/OrD01xWPuWTLlrvuYoMTFROw62cPjwYWJiYgJdLtEIyLEVQgjRUkxbfsLltJUzI+s1b8cma2Aunhv7cOXNjbuHIC+eHNEAJXJOmrg2PKVUttY6sa755GFLQgghhGjxIttZXF5k1zd3tWlyAR08Y5LbOg14G1ttkb8DaojgkUCrCbnuuuuoqKjZlGDdunX2UROFEEII4ZuGvMhuCk3WWoJADL8uRHUSaDUhH374YUMXQQghhGiWAn2R7U3zroasTRM1SW2RCCQJtIQQQgghCNxFtrfDhDeVJmtNmfRrEg1BAi0hhBBCiADyts+Vt7VpEjR4pyk8H0s0TxJoCSGEEEIEkC99rjytTZOgwXsy2IhoKBJoCSGEEEIEUH32uZKgwXu+BL5SaygCIeiB1hcVFRwqL+dcVRVtLRZ6tmnDT4L8kN7f/va3DB48mBtvvDHg696yZQuHDh1i/vz5taa1a9eO0tLSgOWllOL+++/nqaeeAmDJkiWUlpaycOHCgOUhhBBCiLpVvzAPD4UQC1yodh3v2OfK1wt5GaHQe94GvlJrKAIlqMPZfFFRQU5ZGeeqzMl+rqqKnLIyvqgI7tPPf/e739VLkAUwevRop0FWfQgNDeXVV1/l22+/DUp+QgghRIuQCURhrpKirO/dsF2Y2y7mz1YAGsLbKMBc0Fd/+LDj/LYL+az8uq+HXAUHMkKha2OS23KJQ9WCs8B33toSpi0/waq3z7qsNRTCG0H9VB4qL+eCQ9oFa7qvCgsLiYmJYdq0acTGxjJ8+HDOnTMfhJycHJKTk4mPj2fMmDGcPHkSgLS0NDZu3AjA/Pnz6dmzJ/Hx8Tz44IMAfPPNN4wdO5akpCSSkpJ47733auV73XXXkZeXZ3+fkpJCdnY2q1evZvbs2QAUFBTQv39/kpKS+M1vflNj+SeffJKkpCTi4+NZsGCBPf1Pf/oTvXr1olevXjz99NNut71Vq1ZMnz6dpUuXervbhBBCCOFMJjAdKAK09XU6boMtZ835LmgIbaVYOTOSxZMjatSEuGv+VxdPggZRU3L3UCalhNuD0boC3yrtfD1Sayi8FdRAy1aT5Wm6p44cOcKsWbPIy8sjIiKCTZs2ATB58mQWL15Mbm4ucXFxLFq0qMZyJ06cYPPmzeTl5ZGbm8ujjz4KwNy5c7nvvvvYu3cvmzZtYurUqbXynDBhAhs2bADg+PHjfPXVV/Tr16/GPHPnzmXGjBns3buXH/7wh/b0bdu2ceTIEfbs2UNOTg7Z2dm8++67ZGdns2rVKj788EOysrJYuXIlBw4ccLvts2bNIjMzk1OnTnm/44QQQghRUzpQ5pBWZk13wdvmfP40/6sraBDOJXcPZfHkCI8DX2cCWWtYvQZt3toSj2ozm5qWsI11CWofrbYWi9Ogqq3FvxM3OjqahIQEAPr160dhYSGnTp2ipKSEIUOGAHDnnXcybty4Gst16NCBNm3aMHXqVG655RZGjhwJwFtvvcWhQ4fs850+fZozZ87Qvn17e9odd9zBsGHDWLRoERs2bKi1boD33nvPHvRNmjSJefPmASbQ2rZtG3369AGgtLSUI0eOUFpaypgxYwgPDwfgtttuY9euXfb5nOnQoQOTJ0/mmWeeoW1buZslhBBC+OWol+l43wfI38Ey5KG6geVJgBvIWsOW0AesJWyjJ4Jao9WzTRtCHNJCrOn+CK02mEZISAiVlR7clsA0vduzZw9jx47ltddeY8SIEQBUVVXxwQcfkJOTQ05ODseOHasRZAFceeWVdOrUidzcXNavX8+ECROc5qGUqpWmtebXv/61ff2fffYZd911F1q7qKuuw69+9StefPFFzp4969PyQgghhLDq4mU63jfnk+Z/jYurANeiLk4PZK2hP01Hm4qWsI2eCGqg9ZPQUBLCwuw1WG0tFhLCwupl1MGOHTty6aWXsmvXLgDWrVtnr92yKS0t5dSpU9x88808/fTT5OTkADB8+HCeffZZ+3y2dEcTJkzgj3/8I6dOnSIuLq7W9AEDBvDKK68AkJl5sXH3TTfdxEsvvWQfgfDYsWN8/fXXDB48mNdee42ysjLOnj3L5s2bGTRoUJ3bGhkZyR133MGLL75Y57xCCCGEcCMDCHNIC7Omu+DYnM+iLl5UOmsuZZs/vNrlT+tWtW/MiuBwFfj+cmi406aGznjTTK6pjxzpybY29W0MlKAP7/6T0NCgDee+Zs0a7rnnHsrKyujWrRurVq2qMf3MmTPceuutlJeXo7W2DyrxzDPPMGvWLOLj46msrGTw4ME8//zztdZ/++23M3fu3FoDXdgsW7aMiRMnsmzZMsaOHWtPHz58OIcPH6Z///6AGfb9b3/7G3379iUtLY1rr70WgKlTp7ptNljdAw88UCM4FEIIIYQPUq2v6Zjmgl0wQVaqyyWAi82hvGkudb7aCGFny3WLbFrVGNj2t6/PzfK2mVx9PmctUIqLiykoKKCiooLQ0FCio6Pp3Lmzx9vaFLYxGJQ3zdUSExP1vn37aqQdPnyYmJiYQJdLNAJybIUQQgjPzVtb4vLicvHkCJ/nFY2bt8fSMVgBU4PWWAY1KS4uJj8/n6pq4ypYLBa6d+/On/4d6tG2NvZt9JdSKltrnVjXfEGv0RJCCCGEaI68aS4lTauaD2+Ppb81aPWtoKCgRpAFZvyCgoICTpR2d7qM47Y29m0MFgm0GrnvvvuOoUOH1kp/++236dSpUwOUSAghhBDOeNNcSppWNR++HMvGPHJkRYXz/mUVFRVebWtj3sZgkU9zI9epUyf76ITV/yTIEkIIIRoXb0YTlJEHm4/mdixDXYylEBoa2uy2tb5JjZYQQgghRAB401xKmlY1H83tWEZHRzvto2UGxGhe21rfJNASQgghhHCQlV/h08WkN82lpGlV89GcjmXnzp0BnI46CM1rW+ubBFpCCCGEENV4O1y3EM1N586d7YGV8J300RJCCCGEqGZz1rkaw1LDxQcQCyGEp4IeaGUeOkTUihVYliwhasUKMg8dCnYR+O1vf8tbb71VL+vesmULTzzxhNNp7dq1C0geWmsGDhzI1q1b7WkbNmxgxIgRAVm/EEII0ZLJ0OstR2ZmJlFRUVgsFqKiosjMzPRpPVn5FcxbW8K05SeYt7aErHznI/eJliWoTQczDx1i+rZtlFWa20RFp08zfds2AFJ79gxaOX73u9/V27pHjx7N6NGj6239AEopnn/+ecaNG8cNN9zAhQsXSE9P53//93/rNV8hhBCiJZCh1xuOr33jfJGZmcn06dMpKysDoKioiOnTpwOQmprqVZmlqalwJqjfGOm7d9uDLJuyykrSd+/2eZ2FhYXExMQwbdo0YmNjGT58OOfOmar9nJwckpOTiY+PZ8yYMZw8eRKAtLQ0Nm7cCMD8+fPp2bMn8fHxPPjggwB88803jB07lqSkJJKSknjvvfdq5XvdddeRl5dnf5+SkkJ2djarV69m9uzZgOlE2L9/f5KSkvjNb35TY/knn3ySpKQk4uPjWbBggT39T3/6E7169aJXr148/fTTLre7V69ejBo1isWLF7No0SImT57MVVdd5csuFEIIIUQ1MoR1w7AFLLYg1xaw1FftUHp6uj3IsikrKyM9Pd2r9UhTU+FKUAOto6dPe5XuqSNHjjBr1izy8vKIiIhg06ZNAEyePJnFixeTm5tLXFwcixYtqrHciRMn2Lx5M3l5eeTm5vLoo48CMHfuXO677z727t3Lpk2bmDp1aq08J0yYwIYNGwA4fvw4X331Ff369asxz9y5c5kxYwZ79+7lhz/8oT1927ZtHDlyhD179pCTk0N2djbvvvsu2dnZrFq1ig8//JCsrCxWrlzJgQMHXG73ggULePnll9m6dSsPP/ywbztPCCGEEDUkdw9lUkq4vQYrsp2FSSnhUjtRz4IdsBw9etSrdFekqalwJahNB7t06ECRk6CqS4cOfq03OjqahIQEAPr160dhYSGnTp2ipKSEIUOGAHDnnXcybty4Gst16NCBNm3aMHXqVG655RZGjhwJwFtvvcWhan3HTp8+zZkzZ2jfvr097Y477mDYsGEsWrSIDRs21Fo3wHvvvWcP+iZNmsS8efMAE2ht27aNPn36AFBaWsqRI0coLS1lzJgxhIeHA3Dbbbexa9cu+3yOwsPDGT9+PO3atXP5cDkhhBBCeE+GsA6+YAcsXbp0oaioyGm6N6SpqXAlqGdAxsCBhLWqGduFtWpFxsCBfq23epAREhJCpUPzRFdatWrFnj17GDt2LK+99pp9MImqqio++OADcnJyyMnJ4dixYzWCLIArr7ySTp06kZuby/r165kwYYLTPJRStdK01vz617+2r/+zzz7jrrvuQmvt6SbbWSwWLBb5IAshhBCiaXMVmNRXwJKRkUFYWFiNtLCwMDIyMrxaT1zX2vUW0tRUQJADrdSePVkxfDhdO3RAAV07dGDF8OH1MhBGx44dufTSS9m1axcA69ats9du2ZSWlnLq1Cluvvlmnn76aXJycgAYPnw4zz77rH0+W7qjCRMm8Mc//pFTp04RFxdXa/qAAQN45ZVXAGqMYnPTTTfx0ksvUVpaCsCxY8f4+uuvGTx4MK+99hplZWWcPXuWzZs3M2jQID/2ghBCCCFE0xDsvnGpqamsWLGCrl27opSia9eurFixwuuBMD749Pta6f2vuURqREXwH1ic2rNn0EYYXLNmDffccw9lZWV069aNVatW1Zh+5swZbr31VsrLy9Fas3TpUgCeeeYZZs2aRXx8PJWVlQwePJjnn3++1vpvv/125s6dW2ugC5tly5YxceJEli1bxtixY+3pw4cP5/Dhw/Tv3x8ww77/7W9/o2/fvqSlpXHttdcCMHXqVJfNBoUQQgghmhNbYBKsUQfBBFveBFaOnPUrAzhY5FnrKtG8KW+aqyUmJup9+/bVSDt8+DAxMTGBLpdoBOTYCiGEEEK4Nm35CZfTVs6MDGJJRDAppbK11ol1zSede4QQQgghhPBBsPuViaYl6E0HhXe+++47hg4dWiv97bffplOnTg1QIiGEEEIIAab/WPWHFYMMhCEukkCrkevUqZPLwTiEEEI0cXuOwZZP4cQ5iGwLo6+Ba69s6FIJITzUEP3KRNMhgZYQQgjREPYcg5cPwvcXzPsT58x7kGBLtCiZmZCeDkePQpcuYBtd3THNjzEr6pU8c024IoGWEEII0RC2fHoxyLL5/oJJl0Cr0crKr5DaiwDKzITp06GszLwvKoJf/hKUgu+/v5g2fbr5v7EGW19UVHCovJxzVVW0tVjo2aYNPwl1fV54O79omqSnnhBCCNEQTpzzLl00uKz8CtbtOMuJ0ioATpRWsW7HWbLyKxq4ZE1XevrFIMvm/PmLQZZNWZmZtzH6oqKCnLIyzlWZ8+JcVRU5ZWV8UeH8vPB2ftF0SY2WEEIIESje9LmKbOs8qIqUTvSNlbNnJn1fadKlVss3R4/Wz7zBdKi8HIe6aS5Y053VUnk7v2i6gl+jtecYPLodZr5pXvccC1rWU6dO5dChQ14vt3DhQsLCwvj666/tae3atQtk0YQQQjR1tj5XtuDJ1ufK1e/c6GvgkpCaaZeEmHTRKNlqsjxNF3Xr0qV+5g0mW81UfaWLpiu4gZa3P0I+0FpT5eJE/etf/0rPnj19Wu9ll13GU0895U/RhBBCNDfVbx6u/ch1nytnrr0SJsZdrMGKbGveS/+sRkuemRR4GRkQFlYzrXVruOSSmmlhYRcHyWhs2lqcH/9ApYumK7hH1F3HXz8UFhYSExPDzJkz6du3L3fddReJiYnExsayYMEC+3wpKSns27cPMDVS6enp9O7dm+TkZIqLi93mMWXKFNavX8+JE66fAC6EEKIFcbx5WKWdz+euz9W1V8If/geW32JeJchq1MYkt+USh04X8swk/6SmwooV0LWrGQCja1dYtQpeeqlm2ooVjXcgjJ5t2uBQN02INT0Q84umK7iBVj12/P3000+ZPHkyBw4c4KmnnmLfvn3k5uayc+dOcnNza81/9uxZkpOT+eijjxg8eDArV650u/527doxZcoUli1b5ndZhRBCNAPObh46I32umo3k7qFMSgm312BFtrMwKSVc+mf5KTUVCguhqsq8pqY6T2usfhIaSkJYmL1Gqq3FQkJYmMv+Vt7OL5qu4A6GUY8df7t27UpycjIAGzZsYMWKFe9zUYUAACAASURBVFRWVnL8+HEOHTpEfHx8jfkvueQSRo4cCUC/fv34z3/+U2cec+bMISEhgQceeMDv8gohhGjiPLlJKH2umh15ZpJw5iehoV4FSt7OL5qm4NZo1WPH3/DwcAAKCgpYsmQJb7/9Nrm5udxyyy2Ul5fXmr9169YopQAICQmhsrKy1jyOIiIimDhxIsuXL/e7vEIIIZo4VzcJLeridOlzJUTTlwlEYa6ao6zvhfBAcGu0bD82ng5964PTp08THh5Ox44dKS4uZuvWraSkpARs/ffffz9JSUkeBWZCCCGasdHXmD5a1ZsPXhIiwZUQzUkmMB2wPeuryPoeoBE3ZxSNQ/Cfo3XtlfX6A9S7d2/69OlDbGws3bp1Y8CAAQFd/2WXXcaYMWNYunRpQNcrhBCiiQnCzUMhRANL52KQZVNmTZdAS9RBae1ilCQnEhMTtW3UPpvDhw8TExMT6HKJRkCOrRBCCCFaNAvg7FJZAfLYqxZLKZWttU6sa77g12gJIYQQQjRyxcXFFBQUUFFRQWhoKNHR0XTu3LmhiyWCrQumuaCzdCHqIE9GqyYjI4OEhIQafxmN9el4QgghhKgXxcXF5OfnU1FRAUBFRQX5+fl1PnNTNEMZgMMDlQmzpgtRB6nRqiY9PZ309PSGLoYQQgghgikT0+fmKNAFTv/yNFUpNduFVVVVUVBQILVaLY2tH1a184MMpH+W8IgEWkIIIYRouZyMKtft8W6cP3+eb4Z9U2NWWw2XaJqy8ivYnHWOE6VVRLazMCa5LcndQ12m26UigZXwiQRaQgghhGi5nIwqF1IRQre/dqsVaIXKA2abrKz8CtbtOMv31qfznCitYt2Os3x2/DwffPp9rXRAHkwt/CZ9tIQQQgjRch11nhz6dc2LbIvFQnR0dBAKJOrD5qxz9mDK5vtK2HXoe6fpm7POBa9wotmSQKuBFBYW0qtXL6+WSUtLY+PGjX7lm5aWxpVXXmlv/vDtt98SFRXl1zqFEEKIJsvF6HFVV1bZa7BCQ0Pp3r279M9qwk6UOh+LvcrFU45czS+EN4LedLDOdrBNSGVlJa1aNb3WlyEhIbz00kvMmDGjoYsihBBCNKwMavbRAgiDkCdCSE5ObqBCiUCLbGdxGjxZlPNgK7Kd1EUI/wX1LLK1j7Wd6LZ2sFn5/nUuXbt2LfHx8fTu3ZtJkyYB8MYbb3DdddfRp08fbrzxRvuQrAsXLuTOO+9k+PDhREVF8eqrr/Lwww8TFxfHiBEjOH/+PADZ2dkMGTKEfv36cdNNN3H8+HEAUlJSeOSRRxgyZAjLli1zmU91eXl5XHvttSQkJBAfH8+RI0cAuHDhAtOmTSM2Npbhw4dz7pyppl65ciVJSUn07t2bsWPHUlbm+Ehy+M1vfkNaWhpVVVUuy+rKr371K5YuXUplZaXb+YQQQohmLxVYAXTFPIS2q/W9DH7QrIxJbsslDvfGL2kFg3pe4jR9THLb4BVONFtBDbRctY/1px1sXl4eGRkZbN++nY8++ohly5YBMHDgQLKysjhw4AATJkzgj3/8o32Zzz//nDfffJPXX3+dX/ziF9xwww0cPHiQtm3b8uabb3L+/HnuvfdeNm7cSHZ2NlOmTKkx7HtJSQk7d+7kgQcecJuPzfPPP8/cuXPJyclh3759/PjHPwbgyJEjzJo1i7y8PCIiIti0aRMAt912G3v37uWjjz4iJiaGF198scb6Hn74Yb7++mtWrVrFhQsX3JbVmS5dujBw4EDWrVvn204XQgghmpNUoBCosr5KkNXsJHcPZVJKuL2mKrKdhUkp4fxiSDun6U21tZVoXILa7s1Ve1d/2sFu376d22+/ncsuuwyAyMhIAL788kvGjx/P8ePH+f7772t0YP3Zz35G69atiYuL48KFC4wYMQKAuLg4CgsL+fTTT/n4448ZNmwYYGqerrjiCvvy48ePt//vLh+b/v37k5GRwZdffsltt93G1VdfDUB0dDQJCQkA9OvXj8LCQgA+/vhjHn30UUpKSigtLeWmm26yr+v3v/891113HStWrACos6yuPPLII4wePZpbbrmlznmFEEIIIZq65O6hTgMoV+lC+CuoNVqu2rv60w5Wa41Sqlb6vffey+zZszl48CAvvPAC5eXl9mm2zq0Wi4XWrVvbl7dYLFRWVqK1JjY2lpycHHJycjh48CDbtm2zLx8eHu5RPjYTJ05ky5YttG3blptuuont27fXKAeYflO2pnxpaWk8++yzHDx4kAULFtRYZ1JSEtnZ2Zw4ccK+/e7K6spPf/pTEhIS2LBhQ53zCiGEEEIIIbwT1EDLVftYf9rBDh06lA0bNvDdd98B2AOQU6dOceWVVwKwZs0ar9Z5zTXX8M033/DBBx8AcP78efLy8pzO60k+//3vf+nWrRtz5sxh9OjR5Obmus3/zJkzXHHFFZw/f57MzMwa00aMGMH8+fO55ZZbOHPmjFdldZSens6SJUs8mlcIIYQQoiXIyq9g3toSpi0/wby1JX6PJSBarqAGWq7ax/pTXRsbG0t6ejpDhgyhd+/e3H///YAZ9GLcuHEMGjTI3qzQU5dccgkbN25k3rx59O7dm4SEBN5//32n83qSz/r16+nVqxcJCQl88sknTJ482W3+tuaBw4YNo0ePHrWmjxs3jmnTpjF69GguXLjgcVkdxcbG0rdvX4/mFUIIIYRo7upr4DbRMimtXTxAwInExES9b9++GmmHDx8mJiYm0OUSjYAcWyGEEEK0JPPWljgdOyCynYXFkyMaoESiMVJKZWutE+uar+k9BEoIIYQQwkfFxcUUFBRQUVFBaGgo0dHR8iBiYVcfA7eJlksCrWZq1qxZ/P/27j/KzbLO//8raUumQbpItSmo+bHI7NLS0soUplLo8tFWh7qiuD1F8t2F7tF4dmllRXE5hkUO2+xxhfUcBV2NuMpKzoEVgbMurVvQtdKuoxY6WDrALG4mI9qmpUgHdtr0R+7vH5kMk5ncmfy4k9z3nefjHM4w12SSqzOZmfuV6329r127dpWM3Xjjjdq4cWObZgQAQHtls1kNDQ0pny9cNOdyOQ0NDUkSYQuSzA825gBj1IOg5VJf/epX2z0FAABsJZ1OT4Ssonw+r3Q6TdCCpEKDtu/+5P9Kzn3lAGPUi6AFAAA6Qi5XvqGB2Tg6T7FB2yP9R/XK63md9SavPtw7l3O2UBeCFgAA6Ag+n69sqJp8riXAAcawCgWnAACgI0QiEXm9pZc+Xq9XkUikTTMC4GasaAEAgI5Q3IdF10EArcCKVpsMDw/rggsuqOlzrr/+ej300EN1P2YymdSGDRsm3h8dHdW5556rdDpd930CAOAkgUBAvb29Wr16tXp7ewlZAJqm5UErlZLCYcnrLbxNpVo9A+ucPHly5hvZyMc//nG99NJLeuKJJyRJt912m/7yL/+SkgkAAADAYi0NWqmUFItJmYxkGIW3sVjjYetf//VftXTpUl144YX68z//c0nSD37wA11yySVavny53vve9yqbzUqSbr/9dl133XVau3atwuGwHn74YX32s5/VkiVL9P73v18nTpyQJD311FNavXq1LrroIr3vfe/T/v37JUl/8id/os997nNavXq1vvzlL5s+zmT79u3TxRdfrGXLlmnp0qX6n//5H0nSqVOn9PGPf1yLFy/W2rVrdfToUUnSN7/5Ta1YsUIXXnihPvKRj2hsbGzaff7d3/2drr/+euXzedO5TuXxePTP//zP+pu/+Rvt3r1bP/rRj3TzzTc39sUHAAAAMJ1hGFX/d9FFFxlTDQ4OThszEwoZRiFilf4XClV9F9M8++yzRnd3t3Ho0CHDMAzj8OHDhmEYxiuvvGLk83nDMAzjm9/8pnHTTTcZhmEYn//8541LL73UOH78uDEwMGDMnTvX2Lp1q2EYhvGhD33IeOSRR4zjx48bK1euNA4ePGgYhmE88MADxsaNGw3DMIzVq1cbf/VXfzXx+GaPM9mmTZuM+++/3zAMw8jlcsbY2JiRTqeNWbNmGXv27DEMwzDWr19vfPe73zUMwzBefvnlic+Nx+PGV77yFcMwDOO6664zvve97xk333yzEYvFjHw+X3GuZm666SZj3rx5xo9//OOKt6vlewsAAAB0Akm7jSqyU0ubYYyM1DZejR//+Mf6sz/7M73lLW+RJJ111lmSpJdeekkbNmzQ/v37dfz48ZLyuL6+Ps2ZM0dLlizRqVOn9P73v1+StGTJEg0PD+uFF17Qs88+qzVr1kgqrDydffbZE58/eZ9TpccpWrlypRKJhF566SVdffXVOu+88yQVuh8tW7ZMknTRRRdpeHhYkvTss8/q1ltv1auvvqrXX39d73vf+ybu6+///u91ySWXKJlMStKMcy3nhhtu0LZt23TFFVfM9OUFAAAAUIeWlg4Gg7WNV8MwDHk8nmnjmzdv1qZNm7R371594xvf0LFjxyY+Vjwvw+v1as6cOROf7/V6dfLkSRmGocWLF2tgYEADAwPau3evtm/fPvH5p59+elWPU3Tttdfq3//93zV37ly9733v049//OOSeUjSrFmzJvZ8XX/99brnnnu0d+9eff7zny+5zxUrVuipp57SK6+8MvHvrzTXcrxe77T2tgAAAACs09Kr7URC8vtLx/z+wni93vOe9+jf/u3fdPjwYUmaCCBHjhzR2972NknSfffdV9N9/tEf/ZEOHTqkn/3sZ5KkEydOaN++fWVvW83j/O///q/+8A//UJ/85Cf1wQ9+UL/61a8qPv5rr72ms88+WydOnFBqyga297///brlllu0bt06vfbaazXNFQAAAEBrtDRoRaNSMimFQpLHU3ibTBbG67V48WLF43GtXr1aF154oW666SZJhaYX69ev12WXXTZRVlit0047TQ899JD+9m//VhdeeKGWLVum//7v/y5722oe58EHH9QFF1ygZcuW6fnnn9df/MVfVHz8YnngmjVr9Md//MfTPr5+/Xp9/OMf1wc/+EGdOnWq6rkCAAAAaA1PYT9XdXp6eozdu3eXjD333HM6//zzrZ4XbIDvLQDADbLZLIcUA7CMx+N5yjCMnplu19JmGAAAAM1gFqay2ayGhoaUz+clSblcTkNDQ5JkGrYIZgCsQNByqRtuuEG7du0qGbvxxhu1cePGNs0IAIDmqBSm0un0xHhRPp9XOp0uG57qCWYAUA5By6W++tWvtnsKADCzlKS4pBFJQUkJSQ3s20VnqhSmcrlc2c8xG681mAGAGYIWAKA9UpJiksbG38+Mvy8RtlCTSmHK5/OV/fjkI1aqvS8AqAWHKQGAw6VSKYXDYXm9XoXD4WnHQthWXG+ErKKx8XGgBmahqbi/aurZkV6vV5FIpOb7AoBaELQAwMFSqZRisZgymYwMw1Amk1EsFnNG2BqpcRwwUSlMBQIBdXd3TwQln8+n7u5u0zLAWoMZAJihdBAAHCwej2tsrHRZaGxsTPF4XNFGDilshaAK5YLlxoEaFEOTWafAQCBQ9f6qme4LAKrl+BWtV199VV/72tfaPY2m+M53viOv16tf/epXE2MXXHCBhoeH2zcpAC1RbTngyEj55R+zcVtJSPJPGfOPjwM1CgQC6u3t1erVq9Xb29tQMLLyvgB0rpYHrRFt1Tb16WEt1zb1aURbG7q/SkHr1KlTDd23Hbz97W9XIsFVB9BJaikHDAbLL/+YjdtKVFJSUkiSZ/xtUjTCAAC4QkuD1oi2ao/u0FHtl2ToqPZrj+5oKGzdcsst+vWvf61ly5bp5ptv1k9+8hNdccUVuvbaa7VkyRJJ0v3336+LL75Yy5Yt0yc+8YmJALZ9+3atXLlS73rXu7R+/Xq9/vrrJfd98OBBXXTRRZKkZ555Rh6PZ+JV4nPPPVdjY2P6wQ9+oEsuuUTLly/Xe9/7XmWzWeXzeYXDYb366qsT9/XOd75T2WxWhw4d0kc+8hGtWLFCK1asmHbW1VQf+MAHtG/fPr3wwgt1f40AOEulcsCpEomE/P7SZSG/3++cF2iikoYl5cffErIAAC7R0qC1T3frlI6VjJ3SMe3T3XXf5xe+8AWde+65GhgY0J133ilJ+sUvfqFEIqHBwUE999xzevDBB7Vr1y4NDAxo1qxZSqVSevnll7VlyxY98cQTevrpp9XT06MvfelLJfe9YMECHTt2TKOjo3ryySfV09OjJ598UplMRgsWLJDf79eqVavU39+vPXv26JprrtEXv/hFeb1eXXXVVXrkkUckST//+c8VDocVCAR044036lOf+pR++ctf6vvf/74+9rGPVfz3eb1effazn9U//MM/1P01AuAsZmV/mUxm2qpWNBpVMplUKBSSx+NRKBRSMpm0//4sAABcrqXNMI7qQE3j9br44osnugP96Ec/0lNPPaUVK1YUHuvoUS1YsED9/f0aHBzUpZdeKkk6fvy4Vq5cOe2+3v3ud2vXrl366U9/qs997nP64Q9/KMMwdNlll0mSXnrpJW3YsEH79+/X8ePHJx53w4YNuuOOO7Rx40Y98MAD2rBhgyTpiSee0ODg4MT9j46O6rXXXtMZZ5xh+u+59tprlUgklE6nLfjqALC7YDCoTKZclwgpFiscNDU5SEWjUYIVAAA209IVrblaWNN4vU4//fSJ/zcMQ9ddd50GBgY0MDCgF154QbfffrsMw9CaNWsmxgcHB/Wtb31r2n1ddtllE6tYV111lZ555hnt3LlTl19+uSRp8+bN2rRpk/bu3atvfOMbOnassGK3cuVKvfjiizp06JAeffRRXX311ZIKp8v/7Gc/m3jc3/72txVDliTNnj1bn/70p/WP//iPVn2JANhYuXLAIrMSQgAAYC8tDVqLtVmz1FUyNktdWqzNdd/nGWecoddee8304+95z3v00EMP6eDBg5KkV155RZlMRr29vdq1a5defPFFSYWLl6GhoWmff/nll+v+++/XeeedJ6/Xq7POOktbt26dWAk7cuSI3va2t0mS7rvvvonP83g8+vCHP6ybbrpJ559/vubPny9JWrt2re65556J2w0MDFT177z++uv1xBNP6NChQ1XdHoBzFcsBzTiio2CNHHvoMgAAJloatIK6Ust1m+bqbEkezdXZWq7bFNSVdd/n/Pnzdemll+qCCy7QzTffPO3jixYt0pYtW7R27VotXbpUa9as0f79+/XWt75V3/nOd/TRj35US5cuVW9vr55//vlpnx8OhyVpYgVr1apVOvPMM/XmN79ZknT77bdr/fr1uuyyy/SWt7yl5HM3bNig+++/f6JsUJK+8pWvaPfu3Vq6dKkWLVqkr3/961X9O0877TR98pOfnAiMANwtGo0qFAqV/ZgjOgqaKBeoHH3oMgAAJjyGYVR9456eHmP37t0lY88995zOP/98q+cFG+B7C7RXMYBM7kDo9/sd2+zC7N8zd+5cHT58eNrtQ6FQY+cGplJSPC6NjEjBoJRISA78unWybDbLwcEAbMfj8TxlGEbPTLdz/IHFAOBWbusoaNa2vlzIkhoskUylpFhMymQkwyi8jcUK43CEbDaroaEh5XI5SVIul9PQ0JCy2WybZwYA1SFo2cC3v/1tLVu2rOS/G264od3TAmAD0WhUw8PDyufzGh4edmzIkmoPTg2VSMbj0pRQp7GxwriLuWmvWzqdVj6fLxnL5/N04AXgGC1t747yNm7cqI0bN7Z7GgDQVGZt6+fPn6+jR49OKyls6NBls1DnwkYiRVNLM4t73SQ5MqAXV7KqHQcAu2FFCwDQEuXa1vv9fn35y1+2vkTSbDXMwY1EZmJWmunU4wB8Pl9N4wBgN6xoAQBaohic4vG4RkZGFAwGlUgkJsYtXXVJJAp7siYHD7+/MO5SZqWZTj0OIBKJaGhoqKR80Ov1KhKJtHFWAFA9ghYAoGWi0WhrytiKj9FBXQfNSjOdehxAsbsgXQcBOBVBCwDgTtGoq4PVVIlEomz7/Ib2urVZIBAgWAFwLMfv0Xr11Vf1ta99rd3TsNy+ffvU3d2to0ePToytW7dODzzwQBtnBQCwq1YdB5AaHFQ4mZT3rrsUTiaVGhy09P4BwC1aHrS2Hd6mdXvXqefpHq3bu07bDm9r6P4qBa1Tp041dN/ttHjxYl199dUTr0Q++uijOnHihK655po2zwwAYFfNPg4gNTio2PbtyoyOypCUGR1VbPt2whYAlNHSoLXt8DZtGdmiAycOyJChAycOaMvIlobC1i233KJf//rXWrZsmW6++Wb95Cc/0RVXXKFrr71WS5YskSTdf//9uvjii7Vs2TJ94hOfmAhg27dv18qVK/Wud71L69ev1+uvv15y3wcPHtRFF10kSXrmmWfk8XgmNhWfe+65Ghsb0w9+8ANdcsklWr58ud773vcqm80qn88rHA7r1Vdfnbivd77zncpmszp06JA+8pGPaMWKFVqxYoV27dpl+m+77bbb9L3vfU8DAwO65ZZb9NWvfrXurxMAWMVNZzWhNvGdOzV28mTJ2NjJk4rv3NmmGQGAfbU0aN3zu3t0zDhWMnbMOKZ7fndP3ff5hS98Qeeee64GBgZ05513SpJ+8YtfKJFIaHBwUM8995wefPBB7dq1SwMDA5o1a5ZSqZRefvllbdmyRU888YSefvpp9fT06Etf+lLJfS9YsEDHjh3T6OionnzySfX09OjJJ59UJpPRggUL5Pf7tWrVKvX392vPnj265ppr9MUvflFer1dXXXWVHnnkEUnSz3/+c4XDYQUCAd1444361Kc+pV/+8pf6/ve/r4997GOm/za/36+77rpLl19+ua655hqdd955dX+dAMAKxbOaMpmMDMOYOKuJsNUZRkZHaxoHgE7W0mYY2RPZmsbrdfHFF0+0f/3Rj36kp556SitWrJAkHT16VAsWLFB/f78GBwd16aWXSpKOHz+ulStXTruvd7/73dq1a5d++tOf6nOf+5x++MMfyjAMXXbZZZKkl156SRs2bND+/ft1/PjxicfdsGGD7rjjDm3cuFEPPPCANmzYIEl64oknNDipxGJ0dFSvvfaazjjjjLL/lj/90z/VmWeeqb/+67+26KsDAPWrdFaTEw/FRW2C8+YpUyZUBefNa8NsmiObzdLpEIAlWrqiFZhT/heV2Xi9Tj/99In/NwxD1113nQYGBjQwMKAXXnhBt99+uwzD0Jo1aybGBwcH9a1vfWvafV122WUTq1hXXXWVnnnmGe3cuVOXX365JGnz5s3atGmT9u7dq2984xs6dqywYrdy5Uq9+OKLOnTokB599FFdffXVkqR8Pq+f/exnE4/729/+1jRkFXm9Xnm9ju9bAsAF3HZWk9tZXeaZWLVK/tmlr9H6Z89WYtWqhu7XLrLZrIaGhpTL5SRJuVxOQ0NDymatfUG4E2SzWfX392vHjh3q7+/na4iO1NKr903nbFKXp6tkrMvTpU3nbKr7Ps844wy99tprph9/z3veo4ceekgHDx6UJL3yyivKZDLq7e3Vrl279OKLL0oqvCI7NDQ07fMvv/xy3X///TrvvPPk9Xp11llnaevWrRMrYUeOHNHb3vY2SdJ999038Xkej0cf/vCHddNNN+n888/X/PnzJUlr167VPfe8USo5MDBQ978dAFrN7EwmwzDYr2UzzSjzjC5apOTatQrNmyePpNC8eUquXavookVlH99pe/nS6XTJAclS4QXSdDrdphk5E4EVKGhp0Oqb36dbg7dq4ZyF8sijhXMW6tbgreqb31f3fc6fP1+XXnqpLrjgAt18883TPr5o0SJt2bJFa9eu1dKlS7VmzRrt379fb33rW/Wd73xHH/3oR7V06VL19vbq+eefn/b54XBYkiZWsFatWqUzzzxTb37zmyVJt99+u9avX6/LLrtMb3nLW0o+d8OGDbr//vsnygYl6Stf+Yp2796tpUuXatGiRfr6179e978dAFotkUjI7/eX/Rj7teylUplnI6KLFmk4FlP+M5/RcCxmGrKcuJevGAwk6fHHH9eGDRt0xRVX6EMf+pDt524nBFagwGMYRtU37unpMXbv3l0y9txzz+n888+3el6wAb63AMpJpVKKx+PKZDJlPx4KhTQ8PNzaSWEar9ercn/jPR7PtItgq4XD4bLPD7s/N/r7+5XL5fT444/rrrvuKglefr+/KeeSudGOHTtMP7Z69eoWzgRoDo/H85RhGD0z3Y6NPwCAmhTPavJ4PGU/zn4tezAr8zQbN1NPCaBT9/JFIhF5vV7de++9JSFLsmY1sFP4fL5pY4+//LKu+dWvOOgaHYWgZQPf/va3tWzZspL/brjhhnZPCwAqsupCHtWrJfRceeWV08Kw3+9XIpGo6fHqKQF06nMjEAiou7t7Yl/3VHYPinZRDKxFj7/8su7KZHQgl+Oga3QUS4JWLeWHmG7jxo0TXQiL/7X7cGK+pwBmUm6/Vq0X8qheLaEnlUrpvvvuK/ld7vF4dN1119VU+lbvPi8nPzcCgYBjg6JdFANrcWXrW7/7nXJTylU56BqdoOGg1dXVpcOHD3Nh7iKGYejw4cPq6uqa+cYAOlY0GlUymVQoFJLH41EoFGIPSxPVEnrK3dYwDG3durWmx6y3BNDpzw0nB0W7CAQC6u3t1erVq5WdUoZZxEHXcLuGm2GcOHFCL7300sT5UXCHrq4uvf3tb9ecOXPaPRUAgGprbmFVIwynNrWwQrHpy8jIiILBoBKJhGOCot2Ek8myB12H5s3TcCzWhhkBjam2GcbsmW4wkzlz5igSiTR6NwAA2FsqJcXj0siIFAxKiYTUwgvvYDBYNvSUK2er5baVJBIJxWKxktWxTlnZiUajBCuLJFatUmz7do2dPDkx5qaDrgEzNMMAAGAmqZQUi0mZjGQYhbexWGG8RWopZ7Oq9M3pJYCwh1oOugbcpOHSQQAAXC8cLoSrqUIhqYUldLWUs9mx9M2OcwKAWlVbee3fowAAIABJREFUOkjQAgCXSQ0OKr5zp0ZGRxWcN0+JVat45bgek0sFzf5WejxSkw//dYti18SpZYiskHWolKS4pBFJQUkJSTwN4BAELQDoQKnBwbJ7ISjTqVGxVHBK575pWryi5WSd3FgDU6QkxSRN/vHyS0qKsAVHqDZosUcLAFwkvnNnSciSOK+mLvH4zCHL7y80xEBV6m0VDxeKqzRkafz9ysezAY5D0AIAFzE7l4bzampU6eLf4ymsZCWTLe066HQcAowJZj9eZG64DEELAFwkOG9eTeMwYXbxHwoV9mQNDxOyalSuE+Jpp52m119/XV6vV+FwWKkWdnFEG5llazI3XIagBQAukli1Sv7ZpUckcl5NHRKJQmngZJQKNmRqq/j58+fLMAwdPnxYhmEok8koFosRtjpBQoU9WZP5x8cBFyFoAYCLcF6NRaLRQmlgKESpoIWi0aiGh4eVz+f1pje9SSdOnCj5+NjYmOJxNuq4XlSFxhchqfCLSjTCgCvRdRAAALSc1+tVuWsQj8ejPC3zAdhYtV0HZ890AwBws2w2q3Q6rVwuJ5/Pp0gkokAg0O5pAa4XDAbLtnunOQYAt6B0EEDHymazGhoaUi6XkyTlcjkNDQ0pm822eWaA+5VrjuH3+5VgHxwAlyBoAQ6RzWbV39+vHTt2qL+/nzBggXQ6Pa1EKZ/PK51Ot2lGaIVUKqVwOEynuzab2hwjFAopmUwqyj44AC7BHi3AAYorL5NDgdfrVXd3N2VuDdixY8fE/z/++OO69957dfDgQS1YsED/9E//xAWfC6VSKcViMY1NOozY7/dzgQ8AqFq1e7RY0QIcgJWX5vD5fJIKIeuuu+5SNpuVYRjKZrO0mXapeDxeErIkOt0BAJqDoAU4QHEPUbXjqE4kEpHX69W999477WvJxbc7jYyM1DTeMqmUFA5LXm/hLSEf5aQkhVW4eguPvw/AtghagAMUV16qHUd1AoGAuru7dfDgwbIfb/vFNxo3JcBsOuussjdra6e7VEqKxaRMRjKMwttYjLCFUilJMUkZScb425gIW4CNEbQAByiuvEzm9XoViUTaNCP3CAQCphfZtJl2uDIB5kuvvabr58wpuVnbO93F49KUckaNjRXGgaK4pClPE42NjwOwJYIW4ADFlZfiCpbP56MRhoVoM+1SZQLM7OPHdfe8efbqdGe2csqKKiYzezrwNAFsi6AFOEQgEFBvb69Wr16t3t5eQpaFOr3NtGvbnZsElTe98oqGh4eVz+c1PDzc/u+z2copK6qdbep+rPJVrxJPE8C2CFoAoELYstXFd4sU251nMhkZhqFMJuPsjouT92R5Tf7E2S3AJBLSlBVVeTzSlVe2Zz5ov3L7sUYlnTbldn5JLLwDtkXQAoA2sMsqkqvanU/dk3Xq1PTb+P2FYGMn0ah03XWFcFVkGNJ999EQo1OV2491QtIZkkKSPONvk+MfC4tOhIANEbQAl8hms+rv79eOHTvU39+vbDbb7inBRCtWkaoNcrZtd16Pck0lJGnWrEKICYWkZLIQbOxm69ZCuJqMhhidy+zH7xVJw5Ly42+lpnYitMsLQoBTeYypv9gr6OnpMXbv3t3E6QCoRzab1dDQUMmhxl6vl4YZNhUOh5XJZKaNh0IhDQ8PN3z/xSA3eaXK7/eX3XfW7Lm0lNc7PaxIhZA15cBv23Hy3GG9sAqhaaqQ3ghYtdyuDrX8HgE6jcfjecowjJ6ZbseKFuBQk1ewnn/++ZKQJUn5fF7pdLpNs0MlzVhFmvzK83XXXVd1OaCrOi46uamEk+duF2469Dmhwv6rycrtx2piJ0JXlRUDbULQAhyouIKVy+Uq3m6mj6M9rD63a2op4qlye5NUPsi5quNiuaYSdtyTVY6T524Hbjv0OarC/qup+7GiKu1GaHYVZ0E+d1VZMdAmBC3AgdLp9LQVrHKK527BXqxeRSr3ynM5ZkHONR0Xo9HCHqxQyP57sqZy8txr0LQ9P2489Dmq0v1YxZA1eU9WuddULOpEyEHuQOMIWoADVbNS5fV6FYlEWjAb1MrqVaRqXmF2bDlgraJRaXi4sK9peLjuoNKWJgAWzd2uamoCU2sZYKcc+lyuG6EkzdL0la8GuaqsGGgTmmEADtTf318xbPl8PkUikYlGGNlsVul0WrlcbtrH4HxmDS1mzZqlfD6vYDCoRCLh3JWqFrNFE4BUqrAaMzJS2KeVSDg+eFXdeKVYBjh5hcrvr7zCFw4XygWn33khtLqFV4WVrKk8Kqx8NWjq34qnn35aX/ziFzUyMsLvEWCSapthELQAB6qlyyAdCZsnlUopHo+3/SLEFsHARdreibGeoOEAXq9X5a45PB5PaSl0PaHJpV+zacJqWpdB/lYA1aPrIOBigUBA3d3dE3uwfD6f6R/Dcvu56EjYuFachVUtVzW0sIG2NwFw434j1bDnp54ywA7Z41Z1N8I68LcCsB4rWoDL7dixw/Rjq1evbuFM3KXtqx5omrZ/b5t5plYbSxKrXnntlDLAeqVU2Ks1okJ3wYQs2ZPF3wqgeqxoAZBk3nmQjoSNafuqB5qm7U0AmnWmVhNboFfTPKTqlVda3VdWrhuhBfhbAViPoAW4XCQSkddb+qNOR8LGVVsG1ZbudWhI20sxmxU0mlSSWEsZbVVHCXRKGaDN8LcCsB6lg0AHoOug9aopg2pHkwq7NOhAg5pR4tekksS2l1qiJpX+HvC3AqgOXQcBoMlmCjWtvgCl+2AHqiWQNWnvk1k3QUmm42gPOgsC1mCPFgA02UxlULXs47KixDAej5eELEkaGxtT3OHd6mCi1j1XTSpJNCuj9Xg8lMraDJ0FgdYiaAFAk9Syj8uKVvEd2aAjlSqs1Hi9hbeddGFf656rJu19SiQS8ng808YNwyDk24zZQfdm4wAaQ+kgADRJtaV8VpUYdtxemU45pNZMM9vA16hc0CqOT11BQfv09/eXDVU+n0+9vb1tmBHgTJQOAkCbVdu9zqqVqLa3JW8Ws1Urlx7sW7VmtYGvQygUKjtutqqL9qCzINBaBC0AqFZKUliF35zh8fdnUE0762pLDKt5rLa2JW+GSvuQzIKom0slJ7PReVOuDfkuEwgE1N3dPXE2ls/noxEG0ESUDgJANVKSYpImL6D4JSXV8IGhdAusoFKnPKkpXfQcpRlt4OueCkcLAOgMtHcHACuFJZW5pldI0nDjd89FqolK+5C++93O3qMF1IhzsgBrsEcLAKxkVo1mUZVaNSWGHanSPqQmddErUW9Xw07uhghbKp6hVWyGkcvlNDQ0pGw22+aZAe5F0AKAaphtl2Kvf3PNtA8pGi2UCebzhbdWh6xazqlq9POAJuIMLaD1CFoAUI2ECnuyJvOPj6N5WrFqZaberoZWd0Nkdaxpstms+vv7tWPHDvX397t6dYcztIDWm93uCQCAIxSv6+MqlAsGVQhZVPg1XzTanj1X9XY1tLIb4tSzwoqrYxL70BpULKUrrvIUS+kkuXLfks/nMz1DC0BzsKIFANWKqtD4Ij/+torr3FQqpXA4LK/Xq3A4rBSrEc5R7zlVVp5v1elnhTVRp5XScYYW0HoELQBokmLb9kwmI8MwlMlkFIvFCFtOUec5VTuvvFL/N2Xs/8bHa9bpZ4U1UaeV0nGGFtB6tHcHgCYJh8PKlDnnKRQKabhTznlyujrOqQqHw3p3JqN/UKHCdETS5yT9dz3f90rniPEcakh/f79pKV1vb28bZgTAKThHCwDazOv1qtzvWI/HM61kCe5h6fd96h4tibPCLDJ1j5ZU+N6xygNgJtUGLZphAMAkVh7oGQwGy65oBevZqwPHsPT7XgxTNa6qYWbFn2sO8AXQLOzRAoBxVh/omUgk5J+yx8fv9ysxwx4fOJvl3/cmnRVGo5ZC2Ort7dXq1avV29tLyAJgKYIWAIyzugtZNBpVMplUKBSSx+NRKBRSMplUlNUIV3PC951GLQDQfOzRAoBxO3bsMP3Y6tWrWzgToLlo1AIA9at2jxYrWgAwzuzgTg70RD3sXJo3YtIe3mwcAFA7ghYAjONAT1jF7qV5Zo05aNQCANYhaAHAOA70hFXi8bjGJrdklzQ2NqZ4PN6mGZWiUQsANB/t3QFgkkAgQLBCw+xemldszBGPxzUyMqJgMKhEImGrhh0A4HQ0wwAAwGI0mwAA96IZBgAAbUJpXvtks1n19/drx44d6u/vr/scPABoFEELAACLOeEsLTey+tDxGaUkhVW4mgqPvw8A4ygdBAAArtDf3z8Rsibz+Xzq7e219sFSkmKSJvc88UtKSiJPA65G6SAAoLlSKSkclrzewlubtC5H5yoXsiqNNySu0pCl8fft0VgSgA0QtAAAtUulpFhMymQkwyi8jcUIW7L3QcVuN3t2+WbKZuNT/SaX038eOaJHf/97/eeRI/pNpYBm1kDSHo0lAdgAQQsAULt4XJpyTpTGxgrjHczuBxW7ndl2iGq2Sfwml9PA2JiO5vOSpKP5vAbGxszDltnZzpz5DGAcQQsAUDuz86Bsck5Uu9j9oGK3O3XqVE3jkw0eO6aptzo1Pl5WQoU9WZP5x8cBQAQtAEA9giYv25uNdwi7H1Tsdj6fr6bxyYorWdWOK6pC44uQJM/4WxphAJiEoAUAmJAaHFQ4mZT3rrsUTiaVGhwsf8NEQppyTpT8/sJ4BwuaBE2zcVgrEonI6y29tPF6vYpEIjN+7lxv+Usis3FJhVA1LCk//paQBWASghYAQFIhZMW2b1dmdFSGpMzoqGLbt5cPW9GolExKoZDk8RTeJpOF8Q6w7fA2rdu7Tj1P92jd3nXadnibJA4qbrdAIKDu7u6JFSyfz6fu7m4FAoEZP3dRV5dmTRmbNT4OAPXgHC0AcKnf5HIaPHZMR/N5zfV6tairS++oUEL1lnvu0eEy+1FC8+ZpOBZr5lQdZdvhbdoyskXHjDe+Vl2eLt0avFV98/uUSqUUj8c1MjKiYDCoRCLBQcUOUevPDIDOVO05WgQtAHChYge1yZv7Z0la5veXvXBMDQ7q/9u6tex9eSTlP/OZpszTidbtXacDJw5MG184Z6EeW/JYG2YEqxG4AFTCgcUA0MFq7aAW37nT9L6C8+ZZNzEXyJ7I1jSOJklJCqtwJRMef98CNbd5BwATBC0AcKFaO6iNjI6a3ldi1SpL5uQWgTnl9/uYjaMJUpJikjJSYUPh+PsWhK2a27wDgAmCFgC4UK0d1MxWreZ3dSm6aJFl83KDTedsUpentEFCl6dLm87Z1KYZdaC4pCnnZWtsfLxBNbd5b6cyq3rZbFb9/f3asWOH+vv7lc2y0gq0C0ELAFyo1g5qiVWr5J89u2TMP3u2vvz//l9zJuhgffP7dGvwVi2cs1AeebRwzsKJRhhoEbNjySw4rqyuNu/tUGZVL//xvH7/1d8rN17mmMvlNDQ0RNgC2mT2zDcBADhNceN+tRv6i6tW8Z07NTI6quC8eUqsWsVqlom++X22CVbZbFbpdFq5XE4+n0+RSKSqduaOFlQhYJQbb9Cirq6yjWRs1+a9zKqe96hX4WRY2fe8Eazy+bzS6bT7nxOADRG0AMCl3uHz1dQpLbpoEcHKYbLZrIaGhpQfL2srrmBIcveFdUKF1ZzJQcM/Pt6gWl+kaBuT1TvfwenzzNHIA2gLghYAAA6VTqcnQlZRR6xgFI8li6sQOIIqhCyLjiur9UWKtjBZ1cstmB6qfHb/twAuZbOCYwAAUC2zlQrXrmBMbv4QVyFc5SUNy5qQ1aSW8U2RUGEVb5L83LyGY8MlY16vV5FIpGXTAvAGghYAAA5ltlLhyhWMJrZ0b8n9Wy0qKSkppMKp4iHJ+02v3nzDmye+/z6fT93d3e5e3QRszGMYRtU37unpMXbv3t3E6QAAgGpN3aMlFVYwXHlxHVb5BhghFVa07H7/AFzD4/E8ZRhGz0y3Y48WAAAOVQxTHdF1sIkt3Vty/yZ+k8vZv/EGgLoQtAAArjairdqnu3VUBzRXC7VYmxXUle2elmUCgUBzg1UqJcXj0siIFAxKiYQUtajrRC2a2NK9Jfdfxm9yuZJW8kfzeQ2MFVopErYA5yNoAQBca0RbtUd36JSOSZKOar/26A5JclXYappUSorFpPGLf2Uyhfel1oetJrZ0b8n9j5u8guVRYTvYZKdUaC1P0AKcj2YYAADX2qe7J0JW0Skd0z7d3aYZOUw8/kbIKhobK4y3WpnmD0rKspbuTb9/vbGCdXR8T53ZLvmjU1r2A3AmVrQAAK51VAdqGscUIyYblDIjheYRFp5dVZVokx+vyfc/eOzYRJlgJXO9vA4OuAE/yQAA15qrhTWNS1IqlVI4HJbX61U4HFYqZdf+3i0QNNugFLR/+3MbqmalapakRV1dzZ8MgKYjaAEAXGuxNmuWSi9aZ6lLi7W57O1TqZRisZgymYwMw1Amk1EsFuvcsJVISP4pp+JO3rg0psLBwaiK2UqVZ9LHl/n97M8CXIJztAAArlZL18FwOKxMZnrruVAopOHh4SbP1KaKXQczIyq04JtSL+iRxJaiqkztMigVVrAIV4CzVHuOFkHLBtzeehgAGpXNZltyVpTX61W5v4sej6fkUGC7a0pH9rA40NcCnJsFOB8HFjsErYcBoLJsNquhoaGJoJPL5TQ0NCRJVYetaoNaMBgsu6IVNN2rZD9N68jeovbnbvcOn49gBXQI9mi1mVtaD287vE3r9q5Tz9M9Wrd3nbYd3tbuKQFwiXQ6PW01KZ/PK51OV/X5xaCWy+UkvRHUstnstNsmEgn5p+xJ8vv9SiSckyaa1pG9Be3PAcBNCFptZtfWwyPaqm3q08Narm3q04i2mt522+Ft2jKyRQdOHJAhQwdOHNCWkS2ELQCWKAakasenqiWoRaNRJZNJhUIheTwehUIhJZNJRVt9OG8DzDqym43XJKpCmWB+/K1zviwA0HKuLx208/6nEW2VRx4ZZY4srNR6uNlqLWe853f36JhRuip3zDime353j/rm9zV/wgBczefzlQ1VvirLr2oNatFo1LbBqpq/acFgoVxwKgdVPwKAK7h6RasYGI5qvyRjIjBUWp1p9dyMMq2aKrUeboVayxmzJ6aX31QaB4BaRCIReae0xfZ6vYpEIlV9vlkgqzao2UW1f9PKdWT3+wvjAIDWcXXQsvP+p3JzkySPvFqu29q66lZrOWNgTvnN6GbjAFCLQCCg7u7uiWDk8/nU3d1dVSOMbDarkydPThuvJajZRbV/06JRKZmUQiHJ4ym8TSYt6DoIAKiJq4OWXfc/VZqDIaPtpY1ztVDPHvbo7r1eJZ726u69Xj172GNazrjpnE3q8pQeCNrl6dKmcza1YrqoJKVCS2bv+NsOPXMVzhcIBNTb26vVq1ert7e36pA1NDSkU6dOlYzPnj276qBmJ7X8TYtGpeFhKZ8vvHV9yEqlpHBY8noLbzv1gGkAtuLqPVpztXC8xGL6eCWt2NdV79xa4eXDl2vryEM6Mb51bPSEtHXEo4gul+ZPv31xH9Y9v7tH2RNZBeYEtOmcTezPareUSlsxZ8bfl2y3gd3OeynhXOWaYEjSrFmzHBeyJHv/3WirpvWzB4DGuHpFa7E2a5ZKV1pm2v/Uqn1d9cytVR743c6JkFV0wiiMm+mb36fHljym3e/arceWPEbIsoO4Ss+70fj7jbZ4tpid91LC2RrtVmg3bf27YefVcbN+9jfeyCoXgLZyddAK6kot122aq7MleTRXZ8+4/6lV+7rqmVurWNncgvO12sislbMVLZ4tZOe9lHA2tzTBKGrb343i6nhGkqE3VsftklvM+tYfPlxY3TKMN1a5CFsAWsjVpYNS4Q9TLX+EWrmvq9zcth3e1vYSvMCcgA6cmP7vrbW5RfF8rWLr9+L5WpJY8WqFoAoXROXGbcTOeykdJaXCauWICt/jhGxXItpqkUhEQ0NDJeWDTmyCMVmtf9MsUWl13A7PMbN+9lMVT22mnLBu/UM5PdJ/VK+8ntdZb/Lqw71z1dvtzBcugFZw9YpWPcxq3VtRA2+Xg3+tam5R6XwttEBC0pQWz/KPj9tIO3/mXMPuKw5NkM1m1d/frx07dqi/v1/Z7PQV90a6FbZSLQfEt4XZKnhG9ignLNfP3owlpzZ3pv6hnL77k//TK68XXrh45fW8vvuT/1P/kDNLcYFWcP2KVq0Wa3PJYb1S62rg7XLwr1XNLThfq82KL9rafJWjnT9zrmH3FYcqZbNZpdNp5XI5+Xw+RSKRsqGo2E2wuFKVy+U0NDQkSdNuHwgEbBesJqv1gPi2MFsd90wab2ezneIKVTxeCFLBoPT664XSwak4tbluj/Qf1fEpJyUcP1kYZ1ULKI+gNUXxD1s7OqDZKZj0ze9rONxZVYKIBkRl+wvtdv7MuYZD9uNVUkt4KtdNMJ/PK51O2zpUlVNpj6JtfgYSKu1gKhVC1pSmSW0N99FoaUlgKiX9ZUw6PmnSp3FqcyOKK1nVjgMgaJXVlhp4uS+YbDpnU8keLYnztVBeu37mXMMh+/EqqSU8uamboCP2KJZbHTfbEmWbcB8dD4KTJm3YcEnfQf5grqEjRz1lxwGUxx4tG3Hbwb998/t0a/BWzfPOmxjzeSkvACznkP14ldQSnprZTbCavV9WcswexaikYUn58bchk9vZJdzHJZ2YMukTUdsdb+EkyxYc0CxP6Yshszx5LVtgoxcFAJvp6KBltw3IxWCycM5CeeTRwjkLdWvwVsd36DtuHJ/4/yOnjrSlwQfgalFJSRUufj3jb5Ny1Iv3tYSnSCQir7f0z5cV3QSL5YvFcFcsX2xm2LLzmYoV2T3cu6Cc1m7e8aaX1XvOSzp9znFJhk6fc1y957ykd7zp5XZPDbAt15cOjmhr2b0fM21ANvu8ZrNib5Sd2KXBB+A65dq5D7dzQo2ppRV7sZSwmsYZtahUvtiMx5McvEfR7s12XFBOazc+n0+RM48ocuaRaeMAynN10KoUpmY6JNX2XaAcwk4NPgDXKLZzL+7zb2fHN4vUGp6a0U2wUvlitY066uHYPYp2brZTroGHnVbcHMiN59IBzebqoFUpTFXagOyILlAO4bYGH+g8djhEfBqXtHOfqt2t2H0+n2nYckuXw45h9xU3B2rWSjLgZq4OWpXC1Fwt1FHtn/axwrgDukA5BJ0H4WTFQ8SLz9/iIeKS2hu22H/SFGav2E8NWUVO7HLYUey84uZQ7X4xBHAaVzfDqNTNqdIGZMd0gXIAtzb4QGeotMewLVKSwpp+flER+08aEggE1N3dPbHnxOfzlbw/FXtTUKtUKqVwOCyv16twOKxUKtXuKQFoIlevaC3W5pK9VtIbYWqmDchmn4faua3BBzqHrfYYTt2XNRX7Tyxh9oo9e1PQqFQqpVgsprGxwg9xJpNRLFbYXBmNsvQGuJGrV7SCulLLdZvm6mxJHs3V2Vqu2ybCVFBXqk/bdLX2qE/bSsYrfR7QkVIpKRyWvN7C2w54JdZsL2Fb9hiW25dV5MB27q1g1eqB2UoXJVSoRTwenwhZRWNjY4rHOdwLcCuPYVR/ondPT4+xe/fuJk4HgC2lUlIsJk2+SPD7pWRSasIrse06XmGqqXu0pMIew7aUv3pVvmTQo8J5rCgxdfVAkvx+v5LJJKsHaAuv16ty11wej6f8PsBUSorHpZERKRiUEomm/L4FUDuPx/OUYRg9M97OLUHLLhdmaB+eA00UDkuZMofShELS8LClDzX1WAapULrbrlVl23QdDKv8uUAhOfr8rGYJh8PKlHnOhkIhDVv8nAVmks1mdeGFF5Y9/Lrsc7LFL24BqE1HBa1WX5hxQW8tKy5k7XZx7jper1Tud4XHI5l0ZKvXNvWZdAQ9W33aVtt92SUkWaHcHi2/KBk0UfPqAdBE/f39+o//+A/dddddJd0qu7q6dO+9905fZW3hi1sAaldt0HLFHq2ZDh+uZERbtU19eljLtU19GtHWGW+/R3eMXwgaE4cZz/R5KK9YmnXgxAEZMibaZ287XNsFdSPPAVQhaNLOzmy8AVYdr2DVc8s2oiqEqpAK5YLsy6ooaPLcNBsHmimXy2nNmjX6zGc+o0AgII/Ho0AgoE9/+tPlS1lHTM5qMBsHYEuuCFr1XpjVE5q4oLeWVe2zOfusyRKJQtnKZH5/YdxiVh2vYLvW7FaIqlAmmB9/S8gylUgk5J/ynPX7/Uo04TkLzKTYSGXNmjV68MEH9V//9V968MEH9YEPfKD8J7TwxS0AzeOKoFXvhVk9oYkLemtZ1T6bs8+aLBot7A0IhQrlgqFQ0/YKVDrjrha2as2OlotGo7rzzju1cOFCeTweLVy4UHfeeSeNMNAWkUhEXm/pJVfFIwJa+OIWgOZxxTlalc7LqqSe0DRXC032j3BBX4/AnIAOnJj+9a61fXa9zwHUIBo1DVZW7lssd8bdQq3SPt2t3YpXff9WPbfgTNlsVkuWLNEDDzwwMeb1epXNZmnLjpYrPufS6bRyuZx8Pp8ikYj5c7H4u5aug4CjuSJozXT4sJl6QlOnXNC3qonApnM2lW2fvemcTTXdT73PATRuaiOSYgmupIbCVvFz671/q55bqMyuzYHS6fS0phf5fF7pdJqghbYwOwzbVIUXtwA4gyuCllR6YVatekJTJ1zQTz07qNhEQJLlYat4f1aEunqeA2hcpRJcK74f9d6/lc8tlNeMkG2VyZ3dqhkHAMBqrmjvXslMr7ba9dXYdlq3d13ZkquFcxbqsSWPNe1xK62iuapNt8s8rOUyO0n3au2x/f2jfla24rdaf39/2VDl8/nU29vbhhkBANyi2vburlnRKqeaV1vdvApSbzhpRxOBSqtokixfYSO4WafZ+xbZF2lfdm4OFIlENDQ0VFI+WLH5AAAAFnNF10EzndyKvZEzhMyaBTSziUClVtxWt+l23fllpXM8AAANTUlEQVRKbWZVl8B23T/qZ+dun4FAQN3d3RNttX0+n7q7u9mfBQBoGVevaDXj1dZWlho2supSKZzMdB/taCJQzypavStsjXxtMF2z9y06cV+k21dM3/g9OH2lcZa6tFCrxssK2/v9qrn5AAAAFnJ10Kql5KiaANXKjd+NNqRopPzPrImAVNi/1YyLx5lacVvZppvzlazX7BJcJ5X4trKZTDtM/T042VydrYVapRH9wJYNMgAAaCVXlw5WW3JUvHAohDJj4sJgRFtLbtfKUsRGy+UaLf/rm9+nx5Y8pt3v2j3RAKOZ5XabztmkLk/p96q4ilbpY/VoR2mkXaUGBxVOJuW96y6Fk0mlBgctf4wRbdU29elhLdc29U37ubLDY1k5R6tLXe2m3O9B6Y0GGAe0s2NLtgEAmMzVQSuoK7Vct2muzpbk0VydreW6bdqrqtUGqFZu/G501cXqcNLsi8e++X26NXirFs5ZKI88WjhnoW4N3qq++X0VP1YPq782TpUaHFRs+3ZlRkdlSMqMjiq2fbulYavaFzHMPreW8FPvYzUyx3LcvmI60+9BOzfIAACglVxdOihVV3JU7YVBK7ufzVRKZ2aiBHL+AX1Af6Adv/Pp5ROjpqV+1e4lacXFYzFU1fqxeh5H4nyl+M6dGjt5smRs7ORJxXfuVHTRIkseo94zsOop0633saw+B6zen12nmOn3IF0iAQAocPWKVrWq7ZzVyu5n9ay6TH1l/o/mv6q/WjKmh991hx5b8ljZkFVtOaDbyu2mlkZ2WsiSpJHR0ZrG61Hv6kY9Zbr1PpbVKzBuXzGd6fcgXSKdL5VKKRwOy+v1KhwOK5VKtXtKAOBIBC1Vf2FQbSmiFeopl6v14rSWckC3Xzx2ouC8eTWN16Pe9t/1hJ96H8vqFuVWl7razUy/B1v5exLWS6VSisViymQyMgxDmUxGsViMsAUAdfAYhlH1jXt6eozdu3c3cTrt08q27c3ysJZLKvf99Ohq7Zk22vN0j4wyt/fIo93vmv59dnvL6k5T3KM1uXzQP3u2kmvXWlY6WK5D3Sx1zXjhXWgNXq78rNBwwcrHqvfzALdIpVKKx+MaGRmR1+vVqVOnpt0mFAppeHi49ZMDABvyeDxPGYbRM9PtXL9Hq1pOah9tpta9EbXuJbFynxTarxim4jt3amR0VMF585RYtcqykCXVfwbWQq1SWt8rGZup/Kzex3LiOV2AVYorWGNjY5JUNmRJ0sjISCunBQCuwIqWi9T6yvzU836kQjmgVWVOblglROuZndMU0XotV7xNswLcKRwOK5PJzHg7VrQA4A2saHWgWl+Zb2b3vVYe7gx3MTun6YB2tmE2gLtVs1Ll9/uVSCRaMBsAcBeClsvUWgLZrHJAq1tmwx2qWeXkHCagdYLBYNkVrVmzZimfzysYDCqRSCgajbZhdgDgbHQddJlaD3ltFi6WMVW1BwPP1AXQLs9xwA0SiYT8fn/JmN/v13333ad8Pq/h4WFCFgDUiaDlItVeyLaC1S2z4XzVHj9Q6bgFq57jhDWgIBqNKplMKhQKyePxKBQKKZlMEq4AwAIELRep55DXZuHQUkxV7SpnpXOYrHiO2+kFCcAOotGohoeHWcECAIuxR8tF7FSuR8tsTFXL8QNmew2teI43e/9gLd026cwJAIB7EbRcpNZztJrNDWeTwTqLtbns8QO1rHJa8Rxv5gsStXTbpDMnAADuRumgi1CuBzurVBJYLSue483cP1hLaaOdSn0BAID1WNFyEcr10GyNlro1uso503O8mvlZsbJmppbVMjuV+gIAAOsRtFyGcj00i11K3cye49XOr5kvSNRS2mi3Ul8AAGAtSgcBVMXupW61zC+oK9Wnbbpae9SnbZYFxVpKGyn1BQDA3VjRAlAVq0rdmtVpzw6leLWslpndVpK2qY/yXwAAHI6gBaAqVpS6NbP80C6leLWU7069rV3KMwEAQOMoHQRQFStK3ZpZfuiGUjy7l2cCAIDqsaIFoCpWNJFoZnmf07puliuhtEP5IwAAsAZBC0DVGu1qaVV5n9k+L6d03TQrETxNf6DjenXa7elECACA81A6CKBlrCjvK4aUQmAzJkLKiLZaPNvajGirtqlPD2u5tqmv4nzMSgQNGY4vfwQAAAUELQAtE9SVWq7bNFdnS/Jors7Wct1W0ypUtfuYagk+jao1/JmVAp7QaMNfHwAAYA+UDgJoqUbL+6rZx9Tq7n2Vwl+5x6tUQumU8ke7ymazSqfTyuVy8vl8ikQiCgQC7Z4WAKADsaIFwFHM9itNHm91975am1i4oUOiHWWzWQ0NDSmXy0mScrmchoaGlM1m2zwzAEAnImgBcJRqQkqru/dVE/4ms6KEEtOl02nl8/mSsXw+r3Q6bdljpFIphcNheb1ehcNhpVKpsmMAAFA6CKBtzLoHVlJNG/dWH168WJtLShWlmVeoKBG0XnElq9rxWqVSKcViMY2NjUmSMpmMNm7cKI/Ho+PHj0+MxWIxSVI0GrXkcQEAzuQxDKPqG/f09Bi7d+9u4nQAdIqp+6ikQjixYmWnmfdd6TGdcoaXW/X395cNVT6fT729vQ3ffzgcViaTqeq2oVBIw8PDDT8mAMB+PB7PU4Zh9Mx0O1a0ALRFrQ0katGOw4tZoWq/SCSioaGhkvJBr9erSCRiyf2PjIw05bYAAHciaAFoi2bvoyL4dJ5id8FmdR0MBoNVr2gFg0FLHhMA4FwELQBt0ep9VOgMgUCgae3cE4lEyR4tSZozZ07JHi1J8vv9SiQSTZkDAMA56DoIoC2sanHeyoOJ0dmi0aiSyaRCoZA8Ho9CoZC+/e1v61/+5V9KxpLJJI0wAAA0wwDQPo02kGhH0wsAANDZaIYBwPYa3UfVzIYaAAAAjaB0EIBjtfpgYgAAgGoRtAA4llnjDBpqAACAdiNoAXAsqxpqAAAAWI09WgAcqx0HEwMAAFSDoAXA0TiYGAAA2BGlgwAAAABgMYIWAAAAAFiMoAUAAAAAFiNoAQAAAIDFCFoAAAAAYDGCFgAAAABYjKAFAAAAABYjaAEAAACAxQhaAAAAAGAxghYAAAAAWIygBQAAAAAWI2gBAAAAgMUIWgAAAABgMYIWAAAAAFiMoAUAAAAAFiNoAQAAAIDFCFoAAAAAYDGCFgAAAABYjKAFAAAAABYjaAEAAACAxQhaAAAAAGAxghYAAAAAWIygBQAAAAAWI2gBAAAAgMUIWgAAAABgMYIWAAAAAFiMoAUAAAAAFpvd7gkAAEqNaKv26W4d1QHN1UIt1mYFdWW7pwUAAGpA0AIAGxnRVu3RHTqlY5Kko9qvPbpDkghbAAA4CKWDAGAj+3T3RMgqOqVj2qe72zQjAABQD4IWANjIUR2oaRwAANgTQQsAbGSuFtY0DgAA7ImgBQA2slibNUtdJWOz1KXF2tymGQEAgHrQDAMAbKTY8IKugwAAOBtBCwBsJqgrCVYAADgcQQsAmqzac7E4PwsAAPcgaAFAE1V7LhbnZwEA4C40wwCAJqr2XKxGz8/6TS6n/zxyRI/+/vf6zyNH9JtcrrGJAwCAhrCiBQBNVO25WI2cn/WbXE4DY2M6VfycfF4DY2OSpHf4fNVPFgAAWIYVLQBoomrPxWrk/KzBY8cmQlbRqfFxAADQHgQtAGiias/FWqzN8mpOyZhXc6o6P+toPl/TOAAAaD5KBwGgiWo5F8uQUfF9M3O93rKhaq6X19IAAGgXghYANFk152Lt090ydLJkzNBJ7dPdM37uoq6ukj1akjRrfBwAALQHQQsAbKCRZhjFhheDx47paD6vuV6vFnV10QgDAIA2ImgBgA3M1UId1f6y49V4h89HsAIAwEYo4AcAG6i2aQYAAHAGVrQAwAZqaZoBAADsj6AFADZRTdMMAADgDJQOAgAAAIDFCFoAAAAAYDGCFgAAAABYjKAFAAAAABajGQYAONyIttKtEAAAmyFoAYCDjWir9ugOndIxSdJR7dce3SFJhC0AANqIoAUADjN5BcsjjwzlSz5+Sse0T3cTtAAAaCOCFgA4yNQVLENG2dsd1YFWTgsAAExBMwwAcJB9unsiZFUyVwtbMBsAAGCGoAUADlLNStUsdWmxNrdgNgAAwAxBCwAcxGylyiOvJI/m6mwt123szwIAoM3YowUADrJYm0v2aEmFFSzCFQAA9kLQAgAHKYYpzs0CAMDeCFoA4DBBXUmwAgDA5tijBQAAAAAWI2gBAAAAgMUIWgAAAABgMYIWAAAAAFiMoAUAAAAAFiNoAQAAAIDFCFoAAAAAYDGCFgAAAABYjKAFAAAAABYjaAEAAACAxQhaAAAAAGAxghYAAAAAWIygBQAAAAAWI2gBAAAAgMUIWgAAAABgMYIWAAAAAFiMoAUAAAAAFvMYhlH9jT2eQ5IyzZsOAAAAANhayDCMt850o5qCFgAAAABgZpQOAgAAAIDFCFoAAAAAYDGCFgAAAABYjKAFAAAAABYjaAEAAACAxQhaAAAAAGAxghYAAAAAWIygBQAAAAAWI2gBAAAAgMX+f2EqU4hItJaSAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = (15,8)\n",
    "x_tsne=pca_tsne_results[:, 0]\n",
    "y_tsne=pca_tsne_results[:, 1]\n",
    "\n",
    "labels=df_h_e_label.values\n",
    "#print(labels)\n",
    "\n",
    "x=np.c_[x_tsne]\n",
    "y=np.c_[y_tsne]\n",
    "\n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "plt.title('t-SNE dimensions for feature fusion (human/no human)', fontsize=20)\n",
    "\n",
    "seen=set()\n",
    "for i, j, c in zip(x, y, labels):\n",
    "    c=int(c)\n",
    "    plt.scatter(i, j, c=colormap_all(c), marker=\"o\", cmap='jet', label=labelmap_all(c) if c not in seen else None)\n",
    "    seen.add(c)\n",
    "    # plt.annotate(h, (i,j))\n",
    "plt.legend()    \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The end"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_h_e_label = pd.read_csv('human_env_labels.csv')\n",
    "all_filename=pd.read_csv('all_filename.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"864\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"864\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib notebook\n",
    "%matplotlib notebook\n",
    "from matplotlib import pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = (12, 8)\n",
    "x_tsne=pca_tsne_results[:, 0]\n",
    "y_tsne=pca_tsne_results[:, 1]\n",
    "\n",
    "labels=df_h_e_label.values\n",
    "files=all_filename.values\n",
    "#print(labels)\n",
    "\n",
    "x=np.c_[x_tsne]\n",
    "y=np.c_[y_tsne]\n",
    "\n",
    "# plt.xticks([])\n",
    "# plt.yticks([])\n",
    "# plt.title('t-SNE dimensions for feature fusion (human/no human)', fontsize=20)\n",
    "\n",
    "fig,ax = plt.subplots()\n",
    "\n",
    "# seen=set()\n",
    "# for i, j, c in zip(x, y, labels):\n",
    "#     c=int(c)\n",
    "#     sc=plt.scatter(i, j, c=colormap_all(c), marker=\"o\", cmap='jet', label=labelmap_all(c) if c not in seen else None)\n",
    "#     seen.add(c)\n",
    "#     # plt.annotate(h, (i,j))\n",
    "\n",
    "sc=plt.scatter(x, y, c=labels, marker=\"o\", cmap='jet')\n",
    "    \n",
    "annot = ax.annotate(\"\", xy=(0,0), xytext=(20,20),textcoords=\"offset points\",\n",
    "                    bbox=dict(boxstyle=\"round\", fc=\"w\"),\n",
    "                    arrowprops=dict(arrowstyle=\"->\"))\n",
    "annot.set_visible(False)\n",
    "\n",
    "def update_annot(ind):\n",
    "\n",
    "    pos = sc.get_offsets()[ind[\"ind\"][0]]\n",
    "    annot.xy = pos\n",
    "    text=int(files[ind[\"ind\"]])\n",
    "#     text = \"{}, {}\".format(\" \".join(list(map(str,ind[\"ind\"]))), \n",
    "#                            \" \".join([names[n] for n in ind[\"ind\"]]))\n",
    "    annot.set_text(text)\n",
    "    #annot.get_bbox_patch().set_facecolor(cmap(norm(c[ind[\"ind\"][0]])))\n",
    "    # annot.get_bbox_patch().set_alpha(0.4)\n",
    "\n",
    "\n",
    "def hover(event):\n",
    "    vis = annot.get_visible()\n",
    "    if event.inaxes == ax:\n",
    "        cont, ind = sc.contains(event)\n",
    "        if cont:\n",
    "            update_annot(ind)\n",
    "            annot.set_visible(True)\n",
    "            fig.canvas.draw_idle()\n",
    "        else:\n",
    "            if vis:\n",
    "                annot.set_visible(False)\n",
    "                fig.canvas.draw_idle()\n",
    "\n",
    "fig.canvas.mpl_connect(\"motion_notify_event\", hover)\n",
    "\n",
    "# plt.legend() \n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "plt.show()"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}