{
 "metadata": {
  "name": "Subtable totals test"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Set up data\n",
      "N = 10000\n",
      "data = pd.DataFrame({\n",
      "  'Weight': 200*numpy.random.rand(N),\n",
      "  'Length':  40*numpy.random.rand(N),\n",
      "  'Count': numpy.random.randint(0, 100, N),\n",
      "  'Year':  numpy.random.randint(1995, 2012, N),\n",
      "  'Width':  100*numpy.random.rand(N),\n",
      "  'Height':  30*numpy.random.rand(N),\n",
      "  'City': numpy.random.randint(0, 3, N),\n",
      "  'Cars': numpy.random.randint(0, 6, N),\n",
      "})\n",
      "city = 'Bangalore Chennai Delhi Mumbai'.split()\n",
      "cars = 'Mercedes Premier Jaguar BMW Maruti Audi Porsche MM'.split()\n",
      "data['City'] = data['City'].apply(lambda v: city[v])\n",
      "data['Cars'] = data['Cars'].apply(lambda v: cars[v])\n",
      "data['Weight Series'] = data['Weight']"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 53
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Arguments\n",
      "calc = {\n",
      "    'Density': lambda v: v['Weight'] / (v['Length'] * v['Width'] * v['Height']),\n",
      "    'Height %': lambda v: v['Height'],\n",
      "    'Aspect ratio': lambda v: 1,\n",
      "}\n",
      "\n",
      "format = {\n",
      "    'City' : lambda x: '{:s}'.format(str(x)),\n",
      "    'Count': lambda x: '{:,.0f}'.format(float(x)),\n",
      "    'Year' : lambda x: '{:d}'.format(int(x)),\n",
      "    'Weight' : lambda x: '{:,.1f} kg'.format(x),\n",
      "    'Length' : lambda x: '{:,.1f} cm'.format(x),\n",
      "    'Width' : lambda x: '{:,.1f} cm'.format(x),\n",
      "    'Height' : lambda x: '{:,.1f} cm'.format(x),\n",
      "    'Height %': lambda x, data: '{:.0f}%'.format(100. * x / data['Height'].max()),\n",
      "    'Aspect ratio': lambda x, row: '{:,.1f}'.format(row['Length'] / row['Width']),\n",
      "    'Weight Series': lambda x: '{:.1f} kg'.format(x) if numpy.isscalar(x) else SVG('sparkline.svg', width=100, height=30, data=x[:20]) ,\n",
      "}\n",
      "\n",
      "headers = {\n",
      "    'Count': '#',\n",
      "    'Weight': 'Weight (kg)',\n",
      "}\n",
      "\n",
      "agg = {\n",
      "    'City': lambda v: v.iget(0) + ', ...',\n",
      "    'Count': numpy.sum,\n",
      "    'Cars': 'count', # lambda v: v.iget(0) + ', ...',\n",
      "    'Year': numpy.mean,\n",
      "    'Weight': numpy.mean,\n",
      "    'Length': numpy.mean,\n",
      "    'Width': numpy.mean,\n",
      "    'Height': numpy.mean,\n",
      "    'Weight Series': lambda v: tuple(v),\n",
      "}\n",
      "\n",
      "kwargs = {'data':data, 'format':format, 'agg': agg, 'headers': headers, 'classes':True}"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 54
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Specific configuration\n",
      "groupby = ['City', 'Cars']"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 55
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "groups = data.groupby(groupby)\n",
      "by = {k: v for k, v in agg.iteritems() if k not in groupby}"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 56
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "total = len(groupby) - 1\n",
      "subgroupby = groupby[:total]\n",
      "subgroups = data.groupby(subgroupby)\n",
      "subby = {k: v for k, v in agg.iteritems() if k not in subgroupby}\n",
      "subdata = subgroups.sum() if not subby else subgroups.agg(subby)\n",
      "\n",
      "data = groups.sum() if not by else groups.agg(by)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 57
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "if calc is not None:\n",
      "    for new_column, fn in calc.iteritems():\n",
      "        data[new_column] = fn(data)\n",
      "        if subdata is not None:\n",
      "            subdata[new_column] = fn(subdata)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 61
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "indices = {False: [], None: groupby, True: data.index.names}.get(None, None)\n",
      "index_col = {col: i for i, col in enumerate(data.index.names)}\n",
      "index_pos = {col: index_col[col] for col in indices}"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 64
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}