ava.ipynb

{
 "metadata": {
  "name": ""
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "First, we load the AVA dataset, as provided on the website."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import os\n",
      "import pandas\n",
      "import numpy as np\n",
      "\n",
      "AVA_PATH = 'data_static/AVA_dataset'\n",
      "\n",
      "def load_ava_df():\n",
      "    \"\"\"\n",
      "    Load the whole AVA dataset as a DataFrame, with columns\n",
      "        image_id: int, ratings: list of ints,\n",
      "        semantic_tag_X_id:int , semantic_tag_X_name:string (for X in [1, 2]),\n",
      "        challenge_id: int, challenge_name: string.\n",
      "\n",
      "    Ex. Get ratings for all images with tag of 'Macro':\n",
      "\n",
      "    >>> ind = (df['semantic_tag_1_name'] == 'Macro') | \\\n",
      "              (df['semantic_tag_2_name'] == 'Macro')\n",
      "    >>> X = np.vstack(df[ind]['ratings'])\n",
      "    >>> X.shape\n",
      "    (19171, 10)\n",
      "    >>> X.dtype\n",
      "    dtype('int64')\n",
      "\n",
      "    Returns\n",
      "    -------\n",
      "    df: pandas.DataFrame\n",
      "    \"\"\"\n",
      "    def load_ids_and_names(filename, column_name):\n",
      "        with open(filename, 'r') as f:\n",
      "            lines = f.readlines()\n",
      "        # example of an (id, name) line: \"37 Diptych / Triptych\"\n",
      "        data = [(int(line.split()[0]), ' '.join(line.split()[1:])) for line in lines]\n",
      "        ids, names = zip(*data)\n",
      "        df = pandas.DataFrame(\n",
      "            data=list(names), index=list(ids),\n",
      "            columns=[column_name], dtype=str)\n",
      "        return df\n",
      "\n",
      "    # Load the tag and challenge id-name mapping.\n",
      "    tags_df = load_ids_and_names(AVA_PATH + '/tags.txt', 'semantic_tag_name')\n",
      "    challenges_df = load_ids_and_names(AVA_PATH + '/challenges.txt', 'challenge_name')\n",
      "\n",
      "    # Load the main data.\n",
      "    X = np.genfromtxt(AVA_PATH + '/AVA.txt', dtype=int).T\n",
      "    image_id = X[1]\n",
      "    ratings = X[2:12].T\n",
      "    mean_rating = (\n",
      "        np.arange(1, 11.) * ratings / ratings.sum(1)[:, np.newaxis]).sum(1)\n",
      "    df = pandas.DataFrame({\n",
      "        'ratings': [row for row in ratings], 'mean_rating': mean_rating,\n",
      "        'semantic_tag_1_id': X[12], 'semantic_tag_2_id': X[13],\n",
      "        'challenge_id': X[14]\n",
      "    }, index=image_id)\n",
      "\n",
      "    # Store the names of the tags and challenges along with the ids.\n",
      "    df['semantic_tag_1_name'] = df.join(tags_df, on='semantic_tag_1_id', how='left')['semantic_tag_name']\n",
      "    df['semantic_tag_2_name'] = df.join(tags_df, on='semantic_tag_2_id', how='left')['semantic_tag_name']\n",
      "    df = df.join(challenges_df, on='challenge_id', how='left')\n",
      "\n",
      "    return df\n",
      "\n",
      "df = load_ava_df()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print(df.iloc[:10])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "        challenge_id  mean_rating                               ratings  \\\n",
        "953619          1396     5.637097    [0, 1, 5, 17, 38, 36, 15, 6, 5, 1]   \n",
        "953958          1396     4.698413  [10, 7, 15, 26, 26, 21, 10, 8, 1, 2]   \n",
        "954184          1396     5.674603     [0, 0, 4, 8, 41, 56, 10, 3, 4, 0]   \n",
        "954113          1396     5.773438     [0, 1, 4, 6, 48, 37, 23, 5, 2, 2]   \n",
        "953980          1396     5.209302     [0, 3, 6, 15, 57, 39, 6, 1, 1, 1]   \n",
        "954175          1396     5.600000    [0, 0, 5, 13, 40, 53, 14, 1, 3, 1]   \n",
        "953349          1396     6.101562    [1, 1, 1, 7, 27, 46, 28, 13, 4, 0]   \n",
        "953645          1396     6.007874     [0, 0, 0, 8, 33, 51, 27, 3, 3, 2]   \n",
        "953897          1396     6.523438    [0, 0, 0, 5, 19, 46, 29, 22, 5, 2]   \n",
        "953841          1396     5.984496     [0, 0, 3, 8, 37, 44, 22, 9, 4, 2]   \n",
        "\n",
        "        semantic_tag_1_id  semantic_tag_2_id semantic_tag_1_name  \\\n",
        "953619                  1                 22            Abstract   \n",
        "953958                  1                 21            Abstract   \n",
        "954184                  0                  0                 NaN   \n",
        "954113                 15                 21              Nature   \n",
        "953980                 22                 38               Macro   \n",
        "954175                 15                 65              Nature   \n",
        "953349                 16                 21              Candid   \n",
        "953645                  0                  0                 NaN   \n",
        "953897                  7                 14                 Sky   \n",
        "953841                 14                 53           Landscape   \n",
        "\n",
        "             semantic_tag_2_name challenge_name  \n",
        "953619                     Macro     100_Meters  \n",
        "953958           Black and White     100_Meters  \n",
        "954184                       NaN     100_Meters  \n",
        "954113           Black and White     100_Meters  \n",
        "953980                    Floral     100_Meters  \n",
        "954175              Insects, etc     100_Meters  \n",
        "953349           Black and White     100_Meters  \n",
        "953645                       NaN     100_Meters  \n",
        "953897                 Landscape     100_Meters  \n",
        "953841  High Dynamic Range (HDR)     100_Meters  \n"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "The tag frequency plot agrees with the paper."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "semantic_tags = set(df['semantic_tag_1_name']).union(set(df['semantic_tag_2_name']))\n",
      "tag_frequencies = [df[(df['semantic_tag_1_name'] == tag) | (df['semantic_tag_2_name'] == tag)].shape[0] for tag in semantic_tags]\n",
      "tag_df = pandas.DataFrame(tag_frequencies, index=semantic_tags, columns=['frequency'])\n",
      "tag_df = tag_df.sort(column='frequency', ascending=False)\n",
      "tag_df.iloc[:30].plot(figsize=(8, 4), kind='bar')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "/Users/karayev/anaconda/lib/python2.7/site-packages/pandas/core/frame.py:3095: FutureWarning: column is deprecated, use columns\n",
        "  warnings.warn(\"column is deprecated, use columns\", FutureWarning)\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 3,
       "text": [
        "<matplotlib.axes.AxesSubplot at 0x11b6bf610>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFGCAYAAABDmhfrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noVAoWsb/Ax9G239f/v+BAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x11b6bf350>"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "But the delta stuff doesn't. Here's the histogram of mean ratings."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "_ = hist(df.mean_rating, bins=20)\n",
      "df.mean_rating.mean()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 4,
       "text": [
        "5.3833261234315826"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAD9CAYAAACx+XApAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x111891590>"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "And here is the plot of number of images at least delta away from 5, vs delta."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "deltas = [0, .5, 1, 1.5, 2]\n",
      "num_images = [df[(df['mean_rating'] <= 5 - delta) | (df['mean_rating'] >= 5 + delta)].shape[0] for delta in deltas]\n",
      "plot(deltas, num_images, 's--k')\n",
      "xlabel('delta')\n",
      "ylabel('# images in dataset')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 5,
       "text": [
        "<matplotlib.text.Text at 0x11b784750>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAZwAAAEKCAYAAAAmfuNnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n46hSalsXAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x11b73ec90>"
       ]
      }
     ],
     "prompt_number": 5
    }
   ],
   "metadata": {}
  }
 ]
}