examples/data_domain_api_intro.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import datetime\n",
    "\n",
    "from btgym import BTgymEnv, BTgymDataset\n",
    "from btgym import BTgymRandomDataDomain, BTgymSequentialDataDomain\n",
    "from btgym import DataSampleConfig, EnvResetConfig\n",
    "\n",
    "from logbook import WARNING, INFO, DEBUG"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Motivation\n",
    "\n",
    "- BTGym data is basically discrete timeflow of equitype records. For the sake of\n",
    "defining episodic MDP over such data and setting formal problem objective it should\n",
    "be somehow structured. \n",
    "\n",
    "- From the other hand, all train data can laso be seen as big external replay memory. Giving algorithms some degree of control over data sampling parameters can be beneficial for overal performance.\n",
    "\n",
    "- The idea is to pass all desirable sample properties as kwargs to env.rest() method.\n",
    "\n",
    "- This notebook is brief introduction to API realisation of formal definitions introduced in Section 1 (Data) in     this draft: https://github.com/Kismuz/btgym/blob/master/docs/papers/btgym_formalism_draft.pdf\n",
    "\n",
    "- objects described here can be thought as nested data containers with bult-in properties like sampling and         splitting data to train and test subsets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Make data domain - top-level data structure:\n",
    "domain = BTgymRandomDataDomain(\n",
    "    filename='../examples/data/DAT_ASCII_EURUSD_M1_2016.csv',\n",
    "    target_period={'days': 50, 'hours': 0, 'minutes': 0},  # use last 50 days of one year data as 'target domain'\n",
    "                                                           # so we get [360 - holidays gaps - 50] \n",
    "                                                           # days of train data (exclude holidays)\n",
    "    trial_params=dict(\n",
    "        start_weekdays={0, 1, 2, 3, 4, 5, 6}, \n",
    "        sample_duration={'days': 30, 'hours': 0, 'minutes': 0}, # let each trial be 10 days long\n",
    "        start_00=True,  # ajust trial beginning to the beginning of the day\n",
    "        time_gap={'days': 15, 'hours': 0},  # tolerance param \n",
    "        test_period={'days': 6, 'hours': 0, 'minutes': 0},  # from those 10 reserve last 2 days for trial test data\n",
    "    ),\n",
    "    episode_params=dict(\n",
    "        start_weekdays={0, 1, 2, 3, 4, 5, 6},\n",
    "        sample_duration={'days': 0, 'hours': 23, 'minutes': 55},  # make every episode duration be 23:55 \n",
    "        start_00=False, # do not ajust beginning time\n",
    "        time_gap={'days': 0, 'hours': 10},\n",
    "    ),\n",
    "    log_level=INFO, # Set to DEBUG to see more output\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here Domain instanse is defined such as:\n",
    "- Holds one yera of 1min bars data;\n",
    "- Splits data to source and target domains, target domain data gets last 29 days of year period;\n",
    "- Defines each Trial to consist of 8 days of train and following 2 days of test data;\n",
    "- Each Episode lasts maximum 23:55."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sampling parameters\n",
    "\n",
    "- For convinience currently implemented control options are referenced in `btgym.datafeed.base.DataSampleConfig`:\n",
    "```\n",
    "DataSampleConfig = dict(\n",
    "    get_new=True,\n",
    "    sample_type=0,\n",
    "    b_alpha=1,\n",
    "    b_beta=1\n",
    ")\n",
    "```\n",
    "...which simply mirrors base data sample() kwargs:\n",
    "\n",
    "        Args:\n",
    "            get_new (bool):                 sample new (True) or reuse (False) last made sample;\n",
    "            sample_type (int or bool):      0 (train) or 1 (test) - get sample from train or test data subsets\n",
    "                                            respectively.\n",
    "            b_alpha (float):                beta-distribution sampling alpha > 0, valid for train episodes.\n",
    "            b_beta (float):                 beta-distribution sampling beta > 0, valid for train episodes.\n",
    "\n",
    "\n",
    "\n",
    "- environment reset kwargs are referenced in `btgym.datafeed.base.EnvResetConfig`:\n",
    "```\n",
    "EnvResetConfig = dict(\n",
    "    episode_config=DataSampleConfig,\n",
    "    trial_config=DataSampleConfig,\n",
    ")\n",
    "```\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sampling cycle\n",
    "\n",
    "- Reset domain which mean load and describe data and reset counters in case of stateful classes such as BTgymSequentialDataDomain:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2018-07-07 06:09:14.150795] INFO: RndDataDomain_0: Loaded 372678 records from <../examples/data/DAT_ASCII_EURUSD_M1_2016.csv>.\n"
     ]
    }
   ],
   "source": [
    "domain.reset()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Sample trial from source domain:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2018-07-07 06:09:25.068405] INFO: RndDataDomain_0: New sample id: <train_trial_w_0_num_1_at_2016-09-07>.\n"
     ]
    }
   ],
   "source": [
    "trial= domain.sample(\n",
    "    get_new=True,\n",
    "    sample_type=0,\n",
    "    b_alpha=1,\n",
    "    b_beta=1\n",
    ")\n",
    "\n",
    "trial.reset()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Here new Trial object from `Source`(or train) domaini have been requested as uniform sample from entire source data interval\n",
    "\n",
    "- If target period duration is set to 0:0:0, trying to get target sample will rise an exeption.\n",
    "\n",
    "- Need to reset before sampling as well:\n",
    "\n",
    "- Implementatdion detail: \n",
    "    - during real BTgym operation domain instance is held by `btgym_data_server`;\n",
    "\n",
    "    - Different Trial samples are sent to every environment instance, so corresponding `btgym_server` can sample           multiple episodes 'in place'.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Sample episode from trial test interval:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2018-07-07 06:09:28.047474] INFO: Trial_0: New sample id: <test_episode_w_0_num_0_at_2016-10-17 13:18:00>.\n"
     ]
    }
   ],
   "source": [
    "episode = trial.sample(\n",
    "    get_new=True,\n",
    "    sample_type=1,\n",
    "    b_alpha=1,\n",
    "    b_beta=1\n",
    ")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Episodes can be sampled from Trial `train` or `test` subsets, just like Trials from Source/Target domains. Here new Episode from test subset of the Trial is requested. Since it is test subset, alpha and beta params doesnt count - it is always uniform (just a common sense heuristic, one can always change it)     \n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Convert to bt.feed object:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data_feed = episode.to_btfeed()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Now print whole path:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Got instance of: <class 'btgym.datafeed.derivative.BTgymRandomDataDomain'>\n",
      "holding data: ['../examples/data/DAT_ASCII_EURUSD_M1_2016.csv']\n",
      "metadata: {'sample_num': 0, 'type': None}\n",
      "  |\n",
      "sample()\n",
      "  |\n",
      "got instance of: <class 'btgym.datafeed.derivative.BTgymDataTrial'>\n",
      "holding data: train_trial_w_0_num_1_at_2016-09-07\n",
      "metadata: {'sample_num': 1, 'type': 0, 'first_row': 254819, 'last_row': 298019, 'parent_sample_num': 0, 'parent_sample_type': None}\n",
      "  |\n",
      "sample()\n",
      "  |\n",
      "got instance of: <class 'btgym.datafeed.derivative.BTgymEpisode'>\n",
      "holding data: test_episode_w_0_num_0_at_2016-10-17 13:18:00\n",
      "metadata: {'sample_num': 0, 'type': 1, 'first_row': 41030, 'last_row': 42465, 'parent_sample_num': 1, 'parent_sample_type': 0}\n",
      "  |\n",
      "to_btfeed()\n",
      "  |\n",
      "got instance of: <class 'dict'>\n",
      "...holding default data line: {'base_data_line': <backtrader.feeds.pandafeed.PandasDirectData object at 0x10f5a4a58>}, wich value is ready to be fed to bt.Cerebro.\n"
     ]
    }
   ],
   "source": [
    "print('Got instance of: {}\\nholding data: {}\\nmetadata: {}'.\n",
    "      format(type(domain), domain.filename, domain.metadata))\n",
    "\n",
    "print('  |\\nsample()\\n  |')\n",
    "\n",
    "print('got instance of: {}\\nholding data: {}\\nmetadata: {}'.\n",
    "      format(type(trial), trial.filename, trial.metadata))\n",
    "\n",
    "print('  |\\nsample()\\n  |')\n",
    "\n",
    "print('got instance of: {}\\nholding data: {}\\nmetadata: {}'.\n",
    "      format(type(episode), episode.filename, episode.metadata))\n",
    "\n",
    "print('  |\\nto_btfeed()\\n  |')\n",
    "\n",
    "print('got instance of: {}\\n...holding default data line: {}, wi ready to be fed to bt.Cerebro.'.format(type(data_feed), data_feed))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Note:\n",
    "- BTgymDataSet class used in most examples is simply special case where we set Trial=Episode by definition.\n",
    "- This nested data structure is intended mostly for upcoming implementation of meta-learning and guided-policy-search algorithms."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Skewed sampling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Beta-distribution quick recap:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW4AAAEVCAYAAADARw+NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xt8XHWd//HXZ2Yyk0kmadKmTduUtpR7qXhpFVwVU2VR\nUBe3IoKgqw+k4q4s/n6ygpddcFHAVVR2VS6rrgqrRUD8sYAKQi+L3GyRtlBaKKXQ+zX3TDK37++P\nczKdpDPJpE1STvt+Ph7zYOac7/l+P5/vOeczZ85MijnnEBGR4Agd6gBERGR4VLhFRAJGhVtEJGBU\nuEVEAkaFW0QkYFS4RUQC5ogs3Gb2KTN7rOB1p5nNGqG+v2JmP/afzzQzZ2aREep7uh9reCT6G43x\n/XyPHcu4pD8z22hmZ4x020PBzK4xszsOdRyvN2NSuP2DI+mf9C1m9oCZHVXmts1mtnk043POJZxz\nG0YiDufcdc65z4xEXANPKufca36s2ZHof7gGjm9mS8xsRHINmiM59+Eys++Y2Utm1mFma83sk2M0\n7qgUfTOLmtnd/vnpzKx5pMcYylhecX/IOZcApgA7gP8Yw7HHxEhdWcvwBW3uD9WnpkOkC/gQMA74\nO+AmM/urQxvSQXsMuAjYfkhGd86N+gPYCJxR8Pps4MWC1zHgO8BreEX9FiAOVANJIAd0+o+pwNuA\nJ4BWYBvwAyA6yPgTgPuAduBp4FrgsYL1Dji2ILY1QAewBbhikDiuAe4G7vD7/oy/7A6/r5l+3wuB\nrX6sVxSM+zPgGwWvm4HN/vPb/fGS/nhfKugv4reZ6ue1F1gPXFLQ1zXAr4Ff+Lk8D8wrMT9fB/7D\nf16Bd6J9238dB3qA8YXjA98Esv66TuAHBXN5KfCSv39+CFiJcfvm704/xmeANxasvwp42V+3Bvjb\ngnWfAv4EfA/YA3wDOAZ41H+9G/hvoG7AcfhPwCo/x58AjcDv/DH+CNQXtD8NeNzPYyXQ7C8vlfuJ\nwMP+/lgHnDdgX98MPOiPfQZFjrUyz6dy8jyjzDneiHeMrwLa/HaV/rp64H5gF9DiP582AvXgPuCL\nZbYdKv6pwD1+jK8A/+gvfz+QAtL+PlrpL/808ILf1wbgsweZy+a+42IsH2NeuIEq4OfALwrWf8/f\nmeOBGuB/gOv9dc34xayg/Vz/pIrgFZMXgC8MMv4ivCJWDczxT5JShXsb8K6CA/ctg8RxjX9gfBjv\n00uc4oX7V/7Yb/APsL65+BklCvfAeRvQX1/hXgb8CKgE3uT3/Z6C2HrwikMYuB54ssT8vAdY7T//\nK7xi+VTBupUlxl8CfGZAXw7vBK8DpvsxvX+QkzINnIv3hnEF3slX4a//KN6JGQI+hlfwpvjrPgVk\ngMv84yAOHAv8Nd6FwER/fr4/YD6fxCvWTcBOvELwZn8OHwWu9ts24RXGs/3x/9p/PbFY7v7+3YRX\nGCJ+n7uB2QX7ug14h99fJSWONf91K/DOEvNWTp5nlDnHG/EuZqbinX8vAJf66yYAH8E7Z2uAu4Df\nFozzIz/OYo9VJWKP+3kXPSaGc4z487gC+BcgCszCK8bvK9j2jgH9fQDvjc+AdwPd7DvHpw+STyvw\n8SLxHfaFu9NPPo139fkGf53hnZDHFLR/O/CKK1LMSvT/BeDeEuvC/pgnFiy7jtKF+zXgs0DtgH72\ni8M/MJYVWTawcBeO/W/ATwpO5gMq3MBReFd9NQXrrwd+VhDHHwvWzQaSg5xMPXgn6lXAV/wDMoF3\nNf7vA8f3Xy+heOF+Z8HrXwNXDXJSPlnwOkRBMSvS/lngHP/5p4DXhjguPgz8ZcB8Xljw+h7g5oLX\nl+EXJuBK4PYB/f0B+LtiueO9sfzvgPa3su+N4GcUXKwMdqwdwPlVLM/Cwl1yjv22Fw04Pm8pMc6b\ngJaDjPXnwO8p8SlsOMcIcOrAYwD4MvBfA8/FQfr/LXD5QeRzSAr3WN7j/rBzrg7vSuPzwFIzm4x3\nxVAFrDCzVjNrxduxE0t1ZGbHm9n9ZrbdzNrxCnGDv+4r/pegnWZ2i99PBO9qqM+rg8T5EbyrrFfN\nbKmZvX2IvDYNsX5gm1fxrm4O1lRgr3OuY0DfTQWvC++/dQOVxe4FO+eSwHK8K5DTgaV4twje4S9b\nOszYBo6bGKRtfm6cczm8E2EqgJl90syeLTgu5uDv54Hb+u0bzWyRmW3xj4s7BrQH71Zcn2SR132x\nzgA+2je2P/478b6jKWYGcOqA9hcCk0vFy/CPteHkWajkHPuK7i8zqzKzW83sVX+cZUDdgd6fN7Nv\n4+3D85xf9cpUKv4ZwNQBc/4VvE9UpWI4y8yeNLO9fvuzGXzuXpfG/OeAzrmsc+43eFeL78T7OJkE\nTnbO1fmPcc77IhO8K7iBbgbWAsc552rxdpb5/V/nvF8+JJxzl+J9VM/gXaH2mT5IfH92zp0DTMJ7\nN/71IHEMtrzQwLG3+s+78N60+hSe5EP1vRUYb2Y1A/reUkY8xSzFuy3yZuDP/uv34X2fsKzENsM5\n+UrJz42ZhYBpwFYzmwH8J96b/AT/Tf85/P1cYvzr/GVv8I+Liwa0H45NeFfcdQWPaufcDSXG3gQs\nHdA+4Zz7XKl4BznWhjLcPIvOcRnjfBE4ATjVH+f0vm78vm4puEga+Hi+sCMz+zpwFnCmc669vDSH\njH8T3ifzwjmvcc6d7TfvN99mFsP7lPUdoNE/ph4syGf6IPl0mtmFw4x71Ix54TbPOXj39F7w30H/\nE/iemU3y2zSZ2fv8TXYAE8xsXEE3NXhfBnaa2YlA4cnRj/N+uvYb4Br/CmI23jfbxWKLmtmFZjbO\nOZf2x8gNEke5/tkf+2S8e6B3+sufBc42s/H+p48vDNhuB959u2J5bcK7Kr7ezCrN7BTgYryrrwOx\nFPgksMY5l8K/FYB3YuwqsU3J+IZhrpkt8D8JfAHoxbsPXY134u0CMLNP412tDaYG75Zcm5k14X0R\neaDuAD5kZu8zs7A/x81mNs1fPzD3+4HjzewTZlbhP95qZicV63yIY20ow82z1ByXM04SaDWz8cDV\nhSudc5cWXCQNfJxckOuXgY/j3b7ZM3AQ/2d1nzqA+J8GOszsSjOL+/tpjpm91d9uBzDTL/bg3QeP\n4V/MmdlZwJkF+bw2SD4J59x/F8QcM7PKvn794+NALxKGbSwL9/+YWSfeAfpNvHuFfe/KV+L9KuJJ\n/yPZH/He6XHOrcX7cm+D/3FoKt4XFB/H+2b4P9lXCEv5PN7Hv+149xr/a5C2nwA2+nFcivdxt1Qc\n5Vrq5/cI8B3n3EP+8tvxfq2wEXioSB7XA1/zx7uiSL8X4N133grci3c/9Y/DiKvQ43j3uvuurtfg\n3fcudbUNcBNwrnm/zf/3Axz3/+HdH27Bm/sFzrm0c24NcCPer4d24H2x+6ch+vo68Ba8LwEfwHvD\nPiD+G+M5eJ/mduFd3f0T+86Zfrn7t6zOBM7H2x/bgW/hFYpSih5rkP+jsHeV2G64eRad4yG2Afg+\n3jGxG69Q/r6MbYq5Du/T4PqCq9evgPcGhvfdymBvJKWOkSzwQbx776/4cf4Y72eH4H2ZCrDHzJ7x\n99E/4n2yacGrIfcdYE7r8N7UmvC++0ji3boZEza8W00iI8fMrsH7UviiQx3L4er1Psdm9k7gH5xz\nFxzqWIIkUH+0ICKHF+fcY3h/zCLDcET+WyUiIkGmWyUiIgGjK24RkYBR4RYRCRgVbhGRgFHhFhEJ\nGBVuEZGAUeEWEQkYFW4RkYBR4RYRCRgVbhGRgFHhFhEJGBVuEZGAUeEWEQkYFW4RkYBR4RYRCZhR\n+R8pNDQ0uJkzZxZd19XVRXV19WgM+7p2pOYNR27uR2recOTmfjB5r1ixYrdzbmI5bUelcM+cOZPl\ny5cXXbdkyRKam5tHY9jXtSM1bzhycz9S84YjN/eDydvMXi23rW6ViIgEjAq3iEjAqHCLiASMCreI\nSMCocIuIBExZvyoxs41AB5AFMs65eaMZlIiIlDacnwPOd87tHrVIRESkLLpVIiISMOUWbgc8ZGYr\nzGzhaAYkIofW2u3tg65/bU83Xb2ZMYpmcEPFOpp60lle2d11SMY259zQjcyanHNbzGwS8DBwmXNu\n2YA2C4GFAI2NjXMXLVpUtK/Ozk4SicRBBx40R2recOTmHtS8t7QkaaqPl1y/u7OXmsoKYpHS131j\nlftQsY6mTNaxpztFY00sv+xg8p4/f/6Ksr8/dM4N6wFcA1wxWJu5c+e6UhYvXlxy3eHsSM3buSM3\n96Dm/aW7Vg66/pYl6936nR2Dthmr3IeKdTRta026Gx9a12/ZweQNLHdl1uEhb5WYWbWZ1fQ9B84E\nnjugtxQRETlo5fyqpBG418z62v/SOff7UY1KRERKGrJwO+c2AG8cg1hERKQM+jmgiEjAqHCLiASM\nCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3\niEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hI\nwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCU\nXbjNLGxmfzGz+0czIBERGdxwrrgvB14YrUBERKQ8ZRVuM5sGfAD48eiGIyIiQzHn3NCNzO4Grgdq\ngCuccx8s0mYhsBCgsbFx7qJFi4r21dnZSSKROJiYA+lIzRuCmXvOOfZ0pZiYiB1wH31592Ry9KSz\n1MUrRjDC/ra39zC5tnJE+trSkqSpPl5y/e7OXmoqK4hFvOu+3kyOpJ9fKpujsydD1KVIJBLs6Uwx\nrqqCSMjKGruzN4NhVMfC/ZZvb++hsbaSne29NNbu2yd9sZab/472XibVxigWTVcqy9bWJMdNSuTH\nHKzPTNaxpztFY00s3/ZgjvX58+evcM7NK6uxc27QB/BB4Ef+82bg/qG2mTt3ritl8eLFJdcdzo7U\nvJ0LZu6tXSl33QNrDqqPvrxXbWp1v3hi4whEVdqX7lo5Zn3dsmS9W7+zI/969eZW94vHX3HOObd+\nZ4e7Zcn6fO7ffWid29raXfbYD67a6h5du2O/5VfevdKlMln3ld+sKhprufl/7d7VrjedLbpuybqd\nbs7Vv+835mC2tSbdjQ+t69f2YI51YLkborb2Pcq5VfIO4G/MbCOwCHiPmd0x/PcTEREZCUMWbufc\nl51z05xzM4HzgUedcxeNemQiIlKUfsctIhIwkeE0ds4tAZaMSiQiIlIWXXGLiASMCreISMCocIuI\nBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASM\nCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3\niEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMAMWbjNrNLMnjaz\nlWb2vJl9fSwCExGR4iJltOkF3uOc6zSzCuAxM/udc+7JUY5NRESKGLJwO+cc0Om/rPAfbjSDEhGR\n0sq6x21mYTN7FtgJPOyce2p0wwq+pS/uOtQhjKrC/IbKNeccK17dC0BPOsuTG/b0W7+zo4fnt7aV\nPXYmm+Oxl3aXPf5YWL5xL529mRHr77ktbbywrZ3blr3MC9va2daWHHYf67Z35LcrnKNNe7t5eVfn\nfssLFVs+2Dw/tn432Zxj2Qjui77xVm5qpaU7NWT7zt4Myzfu7bdsW1uSdds7Sm6zYVcnr+3pHjKO\nx17and8fffvktb3dPPNqCx096Xyc2dzYXNOad0FdZmOzOuBe4DLn3HMD1i0EFgI0NjbOXbRoUdE+\nOjs7SSQSBxxwUGxpSdJUH8+/PtzyLsxvYK4DdXR0kCTKpJoY2ZxjV2cvk2sr8+uT6SzJVJbx1dGy\nxnbAttYeptZ5fWxpTdJUV3r8A1EszsHs6OhlQlWUSNjyy/r2+XDzA9jblcIMNrckmVYfJxYJUxUN\nl2xfbB+0dKeJRUJURcP91nf2ZsjlHLXxiqLbbWnxin2x5U31cXZ39lJTWUEs4l339aSzvLSzkzlN\n43hlVxe18QiVpEkkEuzs6KW+KkpFwbwMpj2ZxkJGTSzSb7z2ZIajJ1azrTXJ1IJ93ddmS0uSxtpK\n9nSnaKyJ5dd3p7L0ZnLUV1UAsLW1hyl1lViR8Tp7M7y2t5vZU2q9vv3jaktLEjOjMhrKz800f242\ntyQ5cXItOzt6iEVCRF2K2pqasnIdaP78+Succ/PKauycG9YD+BfgisHazJ0715WyePHikusOJ1+6\na2W/14db3oX5Dcx1oIf++Kj73sPrnHPOtXT1uuseWNNv/apNre4XT2wse+yedMZ97d7V+ddX3j34\n+AeitSu1X5yDufGhdW5ba7Lfsr59Ptz8nHPuF4+/4n711KtuxpX3u1899ap7asOeQdsX2wd3Ld+U\n365w/eK1O9wDq7aW3O5Ld60sudw5525Zst6t39mRX756c6ubceX9rjeddR+9+XF3y5L1+dy/+9A6\nt7W1e4hs93lw1Vb36Nod/ca7del6d+7Nf3KpTNZ95Terisb0pbtWum2tSXfjQ+v6rX/6lT3uruWb\n8q+/du9q15vOFh1vybqdbs7Vv8+v6zuuvnTXSve1e1e7X/r7o2+f9O2fLS3d7sq7V7pbl653v3/4\nkbJzHQhY7sqsw+X8qmSif6WNmcWBvwbWHtBbioiIHLRyflUyBfi5mYXx7on/2jl3/+iGJSIipZTz\nq5JVwJvHIBYRESmD/nJSRCRgVLhFRAJGhVtEJGBUuEVEAkaFW0QkYFS4RUQCRoVbRCRgVLhFRAJG\nhVtEJGBUuEVEAkaFW0QkYFS4RUQCRoVbRCRgVLhFRAJGhVtEJGBUuEVEAkaFW0QkYFS4RUQCRoVb\nRCRgVLhFRAJGhVtEJGBUuEVEAkaFW0QkYFS4RUQCRoVbRCRgVLhFRAJGhVtEJGBUuEVEAkaFW0Qk\nYFS4RUQCRoVbRCRgVLhFRAJGhVtEJGCGLNxmdpSZLTazNWb2vJldPhaBiYhIcZEy2mSALzrnnjGz\nGmCFmT3snFszyrGJiEgRQ15xO+e2Oeee8Z93AC8ATaMd2OtFKpM7qLY96SyZbPl9HKxUJodzbr9Y\ncjlHuiCOdNZrV6qPgXrSWXrS2f2WZ3OOTG7/fnoz2X7j9bXNFmlbatzCmJ3zng9nfwymb56Girkw\nrlzO5fdn1o/NOYdzbr99PDDPdMF46Wwu30df28HmpW+bwriLzUWx/Z7O5rw4cfnt0lmXX54rsk3h\ntv3idP3jdM6bj5wryMMVPx76xkhlcvlHn550Nt9vys+zL+aBsWX9fdCTzu63//ri7XukC/pKZXLk\n/GO1J5Pdt/8AnBdDKpMDt6+985fnCvZbKc5BOjv4PhxJVurkLdrYbCawDJjjnGsfsG4hsBCgsbFx\n7qJFi4r20dnZSSKROMBwx96WliRN9fFht+17/uKODibVVhLJ9o5J3ltakkytj7OttYepdZX55b2Z\nHB09GRoSUQB2tPcyIRElErL9+2hN0lTXP+c1W9sJhYwTJ9fkx2mqj9OVyvLK7i7mTK3t137Dri5q\n4xEaEjE6OjpIEiVeESaTy9GbyTG5dl9syXSWZMp7FM51Outo6U4xqSZGNufY1dlLNuv2y69YvEPZ\n2ppkSl2cwuw37ummKhruN97k2sp8rsl0lvU7O5lWHycWCdOeTDOxNkYuB2u3t3PS5FoiYa/HrlSW\ndE83dbU1+e3mNI3DgO1tPcQqQlSEQyRiETp6M7icozZekY9lb1cKM9jckmRafZxdHSmOb0ywpTXJ\npJpKWrpTpDO5fsebNy9Jpvpz0dKdZldHL7XxCKlMjinj4rR0p9jblSLnHJNqYrQnM8QioX7zvqUl\nCUA4bMQiISLhEDWxCGu3dzB1XCWpbI6aygqikRDPbWljWn08H+dmf9uZtSFqamrY2dFLfVWUne09\nNNXH2dKSxAEhIx/nc1vbObqhmupomDXb2jlqfBXJVJbuVJaKkBGrCNGezHD0xGqe29KWj3NO0zi2\n+vtmS0uSdM7R0ZOmqd7br73pHLXxCl7e1cmE6ijjq6O8tLMTgFkTE3T0pIlGQhjk4waYmPD2vwNa\nulPEK8JURSPs6erNt5nmz9fmliQnTq5lW1uStmSaGbUhamtqhnUs9pk/f/4K59y8ctqWc6sEADNL\nAPcAXxhYtAGcc7cBtwHMmzfPNTc3F+1nyZIllFr3enTl3au4sPmUYbfte37Ntxdz+RnHkWhbPyZ5\nX3n3Ks5958l84/4X+HjznPzy9Ts7WLx2F+eePguA7z60juZ5M2gsKKB9rrpn/5wvu/oPJCojPHF+\nc36cC5tP4ckNe/g/S5/mxY8392v/g5sf58yTGzn39GN4+JHFvOyaeOO0Ona09/DK7i7Obz4p33b1\n5jae3dzKc5vb+o27tTXJr5dv4rzm42ntTnHzkpdpTaX3y69YvEP56r2r+ei7TqYivO9D5yd/+jRv\nmV7Hec3H09ad5kdL1nN+80n5XFdvbuNzjzzGDQtOpGligr+s2c5Zpx1HdyrDJdc/ypMffDuTx3nz\n+dSGPWxb+wzNzc357V46/3QqwiG++cAajmlIUD+ukuYTJrF43U56Ulma3zAlH8vtT2ykIhzixmWr\nuWHBiSxas4FHP9bMVfes4vJ5x7Ho6U1sb+vpd7x99J1z+Nf7n+fjzW8A4O4Vm1m0Zj1nzG7ktbZu\nrn7vbBY9vYk7V2+iO5XhH+Yfyx8372BWQ6Lf/F159yoAxsUqmNVQTZ0f51XXPcK/fGg22/Z28+bZ\njUwfX8Wnv/o7blhwYj7OG5etBuAnZ1bR3NzM9x5+kXfNO4rvP/wSFzafwpV3ryKdyxGvCOfjvOSr\nD3LHxW/i1FkT+Mdr/sBNF5zCy5taeWZXK1NqKzlmQjUPb97BLxecxqe/+rt8nC9+7HT++bfP5fvd\n3t7D0hd3cf2CEzHg5V2dnHn8ZP7+0Se48NRpXDB7Opc+8hgAdy58E6vW7mRmQzUG+bgBLnnXdNqT\nGTI5xz2rNzN7Si1zZ9Rz++pX821uWHAi+Ns9/oHTuPeBNTy4eju3nhEfk/O8rF+VmFkFXtH+b+fc\nb0Y3JBERGUw5vyox4CfAC865745+SCIiMphyrrjfAXwCeI+ZPes/zh7luEREpIQh73E75x4D9v8G\nS0REDgn95aSISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhw\ni4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuI\nBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjAqHCLiASM\nCreISMCocIuIBMyQhdvMfmpmO83subEISEREBlfOFffPgPePchwiIlKmIQu3c24ZsHcMYhERkTKY\nc27oRmYzgfudc3MGabMQWAjQ2Ng4d9GiRUXbdXZ2kkgkiq5LZ3OkMo7qWLjo+pbuNACRkBGPhomE\nLL+8vqpiyDwAejM5sjlHVbT4GANtaUlSG6+gJ5NlYiLWL46BY25pSdJUH6elO013b4am+jgv7uhg\nUm0lkWwvsXgVO9p7mVYfZ1dHL72ZHE11ccygvSdDdTSMmdHRk2ZcfF/fOzp6SUTDpLIuP2ZLd5ps\nzhGLhKgIh8g6Ryabo7MnQzwapiedY2pdZb5tbyZLJBSiIRElm3Ns2N3FlHGVmBmZbI7KijC7OnpJ\nVEbo6s3QVBfP5xky2LQ3iRmMi1dQWRGmN52ltqqCPZ0pOnrSHDW+irqCmDfs6iKH45iGBK3t7ezp\nDdFQE2NPZ4rqWJjJtZX5/ZZMZ9na2kM255g5oYpoJMTW1h4yOS+uSTUxdnemyORydPdmaayNsaW1\nh6pomPqqivxzgKa6ODnn2NbW48UeMqaOq+y3z1q60yRTGabUxdnd0Us8GiYRi7BxTzdV0X3jtXSn\nOLqhmu1tPSRiEbrTWfZ09lIVjdCQiLKzo5fKijDOOdqSaRprK5lUE6OjN8OezhQ1oTQT6seRTGdZ\nv7OT+qooAMl0loZElIpwiEQsQkdvht0dvYyvjlJZESabc/Sks95xlMpQFY3Qm85y1IQqdrT1EI2E\nqKwIk87kqIpFqK+qYMOurv3mZUdHL5msozYeoa07TTwaprIizN6uFDnnmFQTY1dHinDImFQTozoW\nZkd7L92pLABm0JCIEgmHcA427e0mHg2Tc46j6qvo6s2wpTXJtPo4m1v2/Rdgek2IbldBOGT0pLP0\nZnJMGVfJ7s4UkZDRncrm43xlTzdH1ccJhcwbw5+DSDhERciIVYTY1ZFi+vgqNuzuzB9jc5rG8dqe\nbsJ+fznnSGdz3jnYlc7n+NrebmKRMJPHVfLqni4AaiorSGdzhMyrId2pTL5fb5+Cw5HK5AiHjEgo\nRG8mm29TFY3ktxsXryCZypLK5phRG6K2pqas2jLQ/PnzVzjn5pXTNnJAIxThnLsNuA1g3rx5rrm5\nuWi7JUuWUGrdyk2tPLe1jQ+cOqPo+ivvXgXAjIYqzp4zhZkN1fnl3zr7lLLifOLlPWxvT3L2m6eV\n1f7Ku1cxs7KaJet2cudn355fFgoZ33j/HML+m0ff8gubT+Gqe1bhHFzYfArXfHsxl59xHIm29Yw/\n9k1c8oM/8fJ17+a8W59gxastrLrmHdRWVvDdh9Zx4bwZRMMhbln2Ml9uPinf71k3/S8Xnjqd57a0\ncYOf55V3r2JrW5J5M8ZzyrRxbG/vYeOeLlqzaVKdORKxCB9vnpNv+/KuTt538mTOPX0WW1uTXHzD\no9yw4HgAXtndxfxZk/jsbU/yidOmkM7muLD5lPx811RG+PHqV/xocjSfMIHGmkqOrqzmhsVrgQif\nOG0K15798GwRAAAJN0lEQVS17339Bzc/znI/v3vuf4hvPAPXLziBa5eu5rOnz+D85pPy+2315jY+\n94PHAOOuS9/EW2eO5w3X/IGOHscXzjia85qP57O3L2fmhGoeenEHC0+fxXWLVwM5rv3wcfzb0ufJ\n5HIArLz6HXT1Zrj4hkcB741m5dXN+Vy+dbaXVyRsfPRdJ3PBbU9y5smNLDz9GD7506d5y/Q6zms+\nnktvX8Hvn9/Oo198Kw8+up4qC/PMq62s2eadvNcvOIFvLvVi8ER49/H1/PxDb+PWpS9z/eK1XHGK\n8ZG/bfbye+SxgrbGDQtOpH5cJc0nTGLxup1c9l9/ZuHp03jvrElsbknSncpw7f8+z75TNMyXzzqa\nHz6xnvaeDJe/92i2t/VAt5fTv35nCZcUzss5x3L9sjUcVV/NGbMbue3ZDYDj8vcezZ2rN9GdyvAP\n84/lW4+tBeCi0yZz/hunc8l/ePuhzw0LTqRuXCXPvtbKTateysf/yBffxg8Xr+c3q7dww4ITuXHZ\n6vx/Ab53egXffzbE+W+dzk1/eomaygoue8/RPPzKDo4aX8Vv/rIlH+dNy9Zw7Tkn0JZMc8s6Lz8w\nTj++gSm1lRwzoZpvPbaWGxYcx42LV+dje/Fjp3Pzj5/i6Y17/ZgNCHH9ghO5dtlqf55P4Malq/1c\njufGJX3bu355Fi+FXn+Dt4n4fYWAELeeES9Z30aSflUiIhIwKtwiIgFTzs8BfwU8AZxgZpvN7OLR\nD0tEREoZ8h63c+6CsQhERETKo1slIiIBo8ItIhIwKtwiIgGjwi0iEjAq3CIiAaPCLSISMCrcIiIB\no8ItIhIwKtwiIgGjwi0iEjAq3CIiAaPCLSISMCrcIiIBo8ItIhIwKtwiIgGjwi0iEjAq3CIiAaPC\nLSISMCrcIiIBo8ItIhIwKtwiIgGjwi0iEjAq3CIiAaPCLSISMCrcIiIBo8ItIhIwKtwiIgGjwi0i\nEjAq3CIiAaPCLSISMCrcIiIBo8ItIhIwZRVuM3u/ma0zs/VmdtVoByUiIqUNWbjNLAz8EDgLmA1c\nYGazRzswEREprpwr7rcB651zG5xzKWARcM7ohiUiIqVEymjTBGwqeL0ZOHU0golHw4yvipYOpD4O\nQEMiRjQS2m95OapjYeripccoNmZDIsqsidX9loWsdHxT6/bFc1xjjTdeG8QrwsxpqgVgVkM1vZks\nIfM6mlhbSSRkhELG5NrKfv0eOynB+Opov36b6uNEIyEm1cZIVEaoz0TpSWepjkZIZ3PEo+F+bR2O\nCQkv70jYmNNUS50/14213nZzmmqZUldJJuv65VMV3Rc3wMwJ1YyvjtKQiOWXT6nrH/OsidX0+PnF\nImHmNFVTXxVlTlMtjX5+ff3Ho6F8P/EKL+6TptTSncowqcZrO318FY21lRw3KZHvB2B8VZSTm8aR\nzeUACIcsnx9AIhbpN1bf87C/A2dNrKYhEfPzqto33oQq5jTVEo2EmFYfJ1YR5thJGUL+YVcYw755\nqQLIz0u8on2//PrUVUWpqfRiq4lF8vNSHYtQV1VBZUV4v20aEjFOmlJLVyrDpNpYPgeA4xoHzEt1\njNlTxzG5NkZjbWV++aTaGCdMrqEnne2//8bFqawoEWcswqTaWL910XCIafVV+eOo8L/efujlhMm1\n+e2qohEaEjGObqhm8rj4fnHWVUWpCIfy+fXNZ0Milo+zbsCcm8HRDdV0pzP9Yi6ch7oSz0dLyNKj\n2n8fc84N3sDsXOD9zrnP+K8/AZzqnPv8gHYLgYX+yxOAdSW6bAB2H0zQAXWk5g1Hbu5Hat5w5OZ+\nMHnPcM5NLKdhOVfcW4CjCl5P85f145y7DbhtqM7MbLlzbl45wR1OjtS84cjN/UjNG47c3Mcq73Lu\ncf8ZOM7MjjazKHA+cN/ohiUiIqUMecXtnMuY2eeBPwBh4KfOuedHPTIRESmqnFslOOceBB4coTGH\nvJ1ymDpS84YjN/cjNW84cnMfk7yH/HJSREReX/Qn7yIiATNqhXuoP5M3s5iZ3emvf8rMZo5WLGOp\njLz/r5mtMbNVZvaImc04FHGOhnL/aQQz+4iZOTM7LH51UE7eZnaev9+fN7NfjnWMo6GMY326mS02\ns7/4x/vZhyLOkWZmPzWznWb2XIn1Zmb/7s/LKjN7y4gH4Zwb8Qfel5gvA7OAKLASmD2gzd8Dt/jP\nzwfuHI1YxvJRZt7zgSr/+ecOh7zLzd1vVwMsA54E5h3quMdonx8H/AWo919POtRxj1HetwGf85/P\nBjYe6rhHKPfTgbcAz5VYfzbwO8CA04CnRjqG0briLufP5M8Bfu4/vxt4r5kV+XvEQBkyb+fcYudc\nt//ySbzfxR8Oyv2nEa4FvgX0jGVwo6icvC8BfuicawFwzu0c4xhHQzl5O6DvTxXHAVvHML5R45xb\nBuwdpMk5wC+c50mgzsymjGQMo1W4i/2ZfFOpNs65DNAGTBileMZKOXkXuhjvnflwMGTu/kfGo5xz\nD4xlYKOsnH1+PHC8mf3JzJ40s/ePWXSjp5y8rwEuMrPNeL9Ku2xsQjvkhlsHhq2snwPKyDOzi4B5\nwLsPdSxjwcxCwHeBTx3iUA6FCN7tkma8T1jLzOwNzrnWQxrV6LsA+Jlz7kYzeztwu5nNcc7lDnVg\nQTdaV9zl/Jl8vo2ZRfA+Su0ZpXjGSln/PICZnQF8Ffgb51zvGMU22obKvQaYAywxs4149/7uOwy+\noCxnn28G7nPOpZ1zrwAv4hXyICsn74uBXwM4554AKvH+LY/DXVl14GCMVuEu58/k7wP+zn9+LvCo\n8+/sB9iQeZvZm4Fb8Yr24XCvs8+guTvn2pxzDc65mc65mXj39//GObf80IQ7Yso51n+Ld7WNmTXg\n3TrZMJZBjoJy8n4NeC+AmZ2EV7h3jWmUh8Z9wCf9X5ecBrQ557aN6Aij+M3r2XhXFi8DX/WX/Sve\nyQreTrwLWA88Dcw61N8Wj1HefwR2AM/6j/sOdcxjlfuAtks4DH5VUuY+N7zbRGuA1cD5hzrmMcp7\nNvAnvF+cPAuceahjHqG8fwVsA9J4n6YuBi4FLi3Y3z/052X1aBzn+stJEZGA0V9OiogEjAq3iEjA\nqHCLiASMCreISMCocIuIBIwKt4hIwKhwi4gEjAq3iEjA/H9X7UP/Ac8TOAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1129f3f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "alpha, beta, size = 2, 1, 500\n",
    "\n",
    "y = np.random.beta(alpha, beta, size)\n",
    "\n",
    "plt.figure(0)\n",
    "plt.suptitle(\n",
    "    'Beta-distribution with parameters: alpha={}, beta={}'.format(alpha, beta), fontsize=12\n",
    ")\n",
    "plt.grid(True)\n",
    "fig = plt.hist(y, histtype='stepfilled', bins=500)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### What it looks like for an episode: \n",
    "- here we plot a histogram of first record time-dates for set of sampled test and train episodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABB0AAAIdCAYAAABx1QWkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11264b748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Suppose Trial is ready, if no - use cells above to get one\n",
    "\n",
    "trial.set_logger(level=13) # make it qiuet\n",
    "\n",
    "train_num_samples = 600\n",
    "test_num_samples = 300\n",
    "\n",
    "# Beta distribution paramters:\n",
    "alpha = 3 # give priority to recent samples \n",
    "beta = 1\n",
    "\n",
    "train_start = np.zeros(train_num_samples)\n",
    "test_start = np.zeros(test_num_samples)\n",
    "\n",
    "# Sample train episode and get time of first record as some realtive number:\n",
    "for i in range(train_num_samples):\n",
    "    train_start[i] = trial.sample(True, 0, alpha, beta).data[0:1].index[0].value / 1e18\n",
    "    \n",
    "# Sample test episode and get time of first record as number:\n",
    "for i in range(test_num_samples):\n",
    "    test_start[i] = trial.sample(True, 1, alpha, beta).data[0:1].index[0].value / 1e18\n",
    "    \n",
    "\n",
    "fig = plt.figure(1, figsize=(18,8))\n",
    "plt.suptitle('Begining times distribution for episodes within single trial', fontsize=16)\n",
    "plt.xlabel('Realtive trial time', fontsize=14)\n",
    "plt.ylabel('Number of episodes', fontsize=14)\n",
    "fig.text(0.25, 0.8, 'Train interval, alpha={}, beta={}'.format(alpha, beta), fontsize=14)\n",
    "fig.text(0.75, 0.8, 'Test interval, uniform', fontsize=14)\n",
    "fig.text(0.25, 0.5, 'Note holidays empty gaps', fontsize=12)\n",
    "\n",
    "plt.grid(True)\n",
    "fig = plt.hist(train_start, histtype='stepfilled', bins=train_num_samples)\n",
    "fig = plt.hist(test_start, histtype='stepfilled', bins=test_num_samples)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Running environment\n",
    "\n",
    "TODO"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}