Meetup13-Lab-EndToEndMLPipelime.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib notebook\n",
    "%matplotlib inline\n",
    "\n",
    "import os\n",
    "os.environ['KERAS_BACKEND'] = \"tensorflow\" # you can use \"cntk\" or \"tensorflow\"\n",
    "import time\n",
    "import datetime as dt\n",
    "import numpy as np\n",
    "from numpy import newaxis\n",
    "from numpy import array\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas\n",
    "from urllib.request import urlretrieve\n",
    "\n",
    "import keras\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "from keras.layers.recurrent import LSTM\n",
    "from keras.models import Sequential\n",
    "from keras import backend as K\n",
    "\n",
    "import talib as ta\n",
    "\n",
    "from azure.storage.blob import BlockBlobService\n",
    "from azureml.core import Workspace, Run\n",
    "import azureml.core"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['/job:localhost/replica:0/task:0/device:GPU:0']"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check what GPUs is Keras seeing\n",
    "K.tensorflow_backend._get_available_gpus()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SDK version: 0.1.0.1217125\n"
     ]
    }
   ],
   "source": [
    "# Check Azure ML core SDK version number\n",
    "print(\"SDK version:\", azureml.core.VERSION)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initialize Workspace"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found the config file in: /home/sshuser/notebooks/Meetups-Data-AI-DFW/aml_config/config.json\n",
      "Workspace name: meetup_aml_workspace\n",
      "Azure region: eastus2\n",
      "Subscription id: b1395605-1fe9-4af4-b3ff-82a4725a3791\n",
      "Resource group: meetup_aml_rg\n"
     ]
    }
   ],
   "source": [
    "ws = Workspace.from_config()\n",
    "print('Workspace name: ' + ws.name, \n",
    "      'Azure region: ' + ws.location, \n",
    "      'Subscription id: ' + ws.subscription_id, \n",
    "      'Resource group: ' + ws.resource_group, sep = '\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create Project and Associate with Run History\n",
    "**Project** is a local folder that contains files for the Azure ML experiments. It is associated with a **run history**, a cloud container of run metrics and output artifacts from your experiments. You can either attach a local folder as a new project, or load a local folder as a project if it has been attached before."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# choose a name for the run history container in the workspace\n",
    "run_history_name = 'train-on-local-nohyperopt'\n",
    "\n",
    "# project folder\n",
    "project_folder = './projects/'+run_history_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "run = Run.start_logging(workspace = ws, history_name = run_history_name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define the TICKER time series to work with"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "TICKER = \"SPY\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "run.log('ticker', TICKER)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create File and Dir paths"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODELS_DIR = \"./models/\"\n",
    "DATA_DIR = \"./data/\"\n",
    "TEST_DATA_PATH = DATA_DIR + TICKER + '-test_score_df.pkl'\n",
    "\n",
    "MIN_MAX_DICT = MODELS_DIR + TICKER +'-min_max.pkl'\n",
    "\n",
    "# File name of the model architecture\n",
    "LSTM_MODEL = MODELS_DIR + TICKER +'-modellstm.json'\n",
    "\n",
    "# File name of the last epoch weights\n",
    "MODEL_WEIGHTS = MODELS_DIR + TICKER +'-modellstm.h5'  \n",
    "\n",
    "# File name of the epoch with the lowest loss\n",
    "MODEL_WEIGHTS_BEST = MODELS_DIR + TICKER +'-modellstm-best.h5'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Ticker Historical Time Series Data from On-Prem SQL to  Blob Storage"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### you need to follow this tutorial to connect your on-premises SQL Server to Azure Data Factory\n",
    "https://docs.microsoft.com/en-us/azure/data-factory/quickstart-create-data-factory-copy-data-tool\n",
    "\n",
    "Then, once the data is on Azure Blob Storage, you can continue below"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Connection String\n",
    "block_blob_service = BlockBlobService(account_name='meetup13',account_key='1dgfh6AeuDiMkxatAofkKKKbNeT9cbABK1TqWMtBky1PpXuRe9T0RUK+/IRN5C/410hqEdQLrAaQEzEEhXPu3g==')\n",
    "\n",
    "container_name ='spy'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "List blobs in the container\n",
      "\t Downloading Blob name: SPY.csv\n"
     ]
    }
   ],
   "source": [
    "# List the blobs in the container\n",
    "print(\"\\nList blobs in the container\")\n",
    "generator = block_blob_service.list_blobs(container_name)\n",
    "for blob in generator:\n",
    "    print(\"\\t Downloading Blob name: \" + blob.name)\n",
    "    block_blob_service.get_blob_to_path(container_name, blob.name, DATA_DIR + blob.name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "filepath = DATA_DIR + TICKER +\".csv\"\n",
    "data = pandas.read_csv(filepath,parse_dates=True, names=['Date','Open','High','Low','Close','Volume','Dividend','Split','Adj_Open','Adj_High','Adj_Low','Adj_Close','Adj_Volume'],\n",
    "                       index_col='Date')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.index = pandas.to_datetime(data.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "      <th>Dividend</th>\n",
       "      <th>Split</th>\n",
       "      <th>Adj_Open</th>\n",
       "      <th>Adj_High</th>\n",
       "      <th>Adj_Low</th>\n",
       "      <th>Adj_Close</th>\n",
       "      <th>Adj_Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2000-01-03</th>\n",
       "      <td>148.250000</td>\n",
       "      <td>148.250000</td>\n",
       "      <td>143.875000</td>\n",
       "      <td>145.4375</td>\n",
       "      <td>8164300.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>104.914399</td>\n",
       "      <td>104.914399</td>\n",
       "      <td>101.818275</td>\n",
       "      <td>102.924034</td>\n",
       "      <td>8164300.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-04</th>\n",
       "      <td>143.531204</td>\n",
       "      <td>144.062500</td>\n",
       "      <td>139.640594</td>\n",
       "      <td>139.7500</td>\n",
       "      <td>8089800.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>101.574975</td>\n",
       "      <td>101.950966</td>\n",
       "      <td>98.821646</td>\n",
       "      <td>98.899071</td>\n",
       "      <td>8089800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-05</th>\n",
       "      <td>139.937500</td>\n",
       "      <td>141.531204</td>\n",
       "      <td>137.250000</td>\n",
       "      <td>140.0000</td>\n",
       "      <td>12177900.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>99.031762</td>\n",
       "      <td>100.159604</td>\n",
       "      <td>97.129857</td>\n",
       "      <td>99.075993</td>\n",
       "      <td>12177900.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-06</th>\n",
       "      <td>139.625000</td>\n",
       "      <td>141.500000</td>\n",
       "      <td>137.750000</td>\n",
       "      <td>137.7500</td>\n",
       "      <td>6227200.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>98.810611</td>\n",
       "      <td>100.137521</td>\n",
       "      <td>97.483700</td>\n",
       "      <td>97.483700</td>\n",
       "      <td>6227200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-07</th>\n",
       "      <td>140.312500</td>\n",
       "      <td>145.750000</td>\n",
       "      <td>140.062500</td>\n",
       "      <td>145.7500</td>\n",
       "      <td>8066500.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>99.297145</td>\n",
       "      <td>103.145185</td>\n",
       "      <td>99.120223</td>\n",
       "      <td>103.145185</td>\n",
       "      <td>8066500.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Open        High         Low     Close      Volume  \\\n",
       "Date                                                                   \n",
       "2000-01-03  148.250000  148.250000  143.875000  145.4375   8164300.0   \n",
       "2000-01-04  143.531204  144.062500  139.640594  139.7500   8089800.0   \n",
       "2000-01-05  139.937500  141.531204  137.250000  140.0000  12177900.0   \n",
       "2000-01-06  139.625000  141.500000  137.750000  137.7500   6227200.0   \n",
       "2000-01-07  140.312500  145.750000  140.062500  145.7500   8066500.0   \n",
       "\n",
       "            Dividend  Split    Adj_Open    Adj_High     Adj_Low   Adj_Close  \\\n",
       "Date                                                                          \n",
       "2000-01-03       0.0    1.0  104.914399  104.914399  101.818275  102.924034   \n",
       "2000-01-04       0.0    1.0  101.574975  101.950966   98.821646   98.899071   \n",
       "2000-01-05       0.0    1.0   99.031762  100.159604   97.129857   99.075993   \n",
       "2000-01-06       0.0    1.0   98.810611  100.137521   97.483700   97.483700   \n",
       "2000-01-07       0.0    1.0   99.297145  103.145185   99.120223  103.145185   \n",
       "\n",
       "            Adj_Volume  \n",
       "Date                    \n",
       "2000-01-03   8164300.0  \n",
       "2000-01-04   8089800.0  \n",
       "2000-01-05  12177900.0  \n",
       "2000-01-06   6227200.0  \n",
       "2000-01-07   8066500.0  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "      <th>Dividend</th>\n",
       "      <th>Split</th>\n",
       "      <th>Adj_Open</th>\n",
       "      <th>Adj_High</th>\n",
       "      <th>Adj_Low</th>\n",
       "      <th>Adj_Close</th>\n",
       "      <th>Adj_Volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4.653000e+03</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4653.0</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4653.000000</td>\n",
       "      <td>4.653000e+03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>147.158041</td>\n",
       "      <td>148.018492</td>\n",
       "      <td>146.211340</td>\n",
       "      <td>147.156245</td>\n",
       "      <td>1.149495e+08</td>\n",
       "      <td>0.011062</td>\n",
       "      <td>1.0</td>\n",
       "      <td>126.160803</td>\n",
       "      <td>126.879130</td>\n",
       "      <td>125.370934</td>\n",
       "      <td>126.162028</td>\n",
       "      <td>1.149495e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>47.029426</td>\n",
       "      <td>47.032172</td>\n",
       "      <td>47.011564</td>\n",
       "      <td>47.037064</td>\n",
       "      <td>1.014869e+08</td>\n",
       "      <td>0.093500</td>\n",
       "      <td>0.0</td>\n",
       "      <td>53.636267</td>\n",
       "      <td>53.711409</td>\n",
       "      <td>53.545052</td>\n",
       "      <td>53.647581</td>\n",
       "      <td>1.014869e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>67.949997</td>\n",
       "      <td>70.000000</td>\n",
       "      <td>67.099998</td>\n",
       "      <td>68.110001</td>\n",
       "      <td>1.436600e+06</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>55.969691</td>\n",
       "      <td>57.658257</td>\n",
       "      <td>55.269556</td>\n",
       "      <td>56.101484</td>\n",
       "      <td>1.436600e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>114.230003</td>\n",
       "      <td>114.940002</td>\n",
       "      <td>113.400002</td>\n",
       "      <td>114.199997</td>\n",
       "      <td>4.670420e+07</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>88.991344</td>\n",
       "      <td>89.596576</td>\n",
       "      <td>88.340048</td>\n",
       "      <td>89.073893</td>\n",
       "      <td>4.670420e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>132.839996</td>\n",
       "      <td>133.600006</td>\n",
       "      <td>131.779999</td>\n",
       "      <td>132.770004</td>\n",
       "      <td>8.627080e+07</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>105.568756</td>\n",
       "      <td>106.338640</td>\n",
       "      <td>104.621110</td>\n",
       "      <td>105.435721</td>\n",
       "      <td>8.627080e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>177.029999</td>\n",
       "      <td>177.529999</td>\n",
       "      <td>176.369995</td>\n",
       "      <td>177.289993</td>\n",
       "      <td>1.570000e+08</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>161.182193</td>\n",
       "      <td>161.655405</td>\n",
       "      <td>160.499680</td>\n",
       "      <td>161.336894</td>\n",
       "      <td>1.570000e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>285.930000</td>\n",
       "      <td>286.628500</td>\n",
       "      <td>284.500000</td>\n",
       "      <td>286.580000</td>\n",
       "      <td>8.710263e+08</td>\n",
       "      <td>1.351330</td>\n",
       "      <td>1.0</td>\n",
       "      <td>283.516145</td>\n",
       "      <td>284.208748</td>\n",
       "      <td>282.098217</td>\n",
       "      <td>284.160657</td>\n",
       "      <td>8.710263e+08</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Open         High          Low        Close        Volume  \\\n",
       "count  4653.000000  4653.000000  4653.000000  4653.000000  4.653000e+03   \n",
       "mean    147.158041   148.018492   146.211340   147.156245  1.149495e+08   \n",
       "std      47.029426    47.032172    47.011564    47.037064  1.014869e+08   \n",
       "min      67.949997    70.000000    67.099998    68.110001  1.436600e+06   \n",
       "25%     114.230003   114.940002   113.400002   114.199997  4.670420e+07   \n",
       "50%     132.839996   133.600006   131.779999   132.770004  8.627080e+07   \n",
       "75%     177.029999   177.529999   176.369995   177.289993  1.570000e+08   \n",
       "max     285.930000   286.628500   284.500000   286.580000  8.710263e+08   \n",
       "\n",
       "          Dividend   Split     Adj_Open     Adj_High      Adj_Low  \\\n",
       "count  4653.000000  4653.0  4653.000000  4653.000000  4653.000000   \n",
       "mean      0.011062     1.0   126.160803   126.879130   125.370934   \n",
       "std       0.093500     0.0    53.636267    53.711409    53.545052   \n",
       "min       0.000000     1.0    55.969691    57.658257    55.269556   \n",
       "25%       0.000000     1.0    88.991344    89.596576    88.340048   \n",
       "50%       0.000000     1.0   105.568756   106.338640   104.621110   \n",
       "75%       0.000000     1.0   161.182193   161.655405   160.499680   \n",
       "max       1.351330     1.0   283.516145   284.208748   282.098217   \n",
       "\n",
       "         Adj_Close    Adj_Volume  \n",
       "count  4653.000000  4.653000e+03  \n",
       "mean    126.162028  1.149495e+08  \n",
       "std      53.647581  1.014869e+08  \n",
       "min      56.101484  1.436600e+06  \n",
       "25%      89.073893  4.670420e+07  \n",
       "50%     105.435721  8.627080e+07  \n",
       "75%     161.336894  1.570000e+08  \n",
       "max     284.160657  8.710263e+08  "
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "use_features = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Technical Analysis\n",
    "SMA_FAST = 3\n",
    "SMA_SLOW = 10\n",
    "RSI_PERIOD = 5\n",
    "EMA_FAST = 6\n",
    "EMA_SLOW = 13"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "if use_features:\n",
    "    run.log('use_features', \"yes\")\n",
    "    \n",
    "    run.log('SMA_FAST', SMA_FAST)\n",
    "    run.log('SMA_SLOW', SMA_SLOW)\n",
    "    data['sma_diff'] = data['Close'].rolling(window=SMA_SLOW).mean() - data['Close'].rolling(window=SMA_FAST).mean()\n",
    "    \n",
    "    run.log('EMA_FAST', EMA_FAST)\n",
    "    run.log('EMA_SLOW', EMA_SLOW)\n",
    "    data['ema_diff'] = ta.EMA(data.Close.values,EMA_SLOW) - ta.EMA(data.Close.values,EMA_FAST)\n",
    "    \n",
    "    run.log('RSI_PERIOD', RSI_PERIOD)\n",
    "    data['rsi'] = ta.RSI(data.Close.values, RSI_PERIOD)\n",
    "\n",
    "else:\n",
    "    run.log('use_features', \"no\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "      <th>Dividend</th>\n",
       "      <th>Split</th>\n",
       "      <th>Adj_Open</th>\n",
       "      <th>Adj_High</th>\n",
       "      <th>Adj_Low</th>\n",
       "      <th>Adj_Close</th>\n",
       "      <th>Adj_Volume</th>\n",
       "      <th>sma_diff</th>\n",
       "      <th>ema_diff</th>\n",
       "      <th>rsi</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2000-01-03</th>\n",
       "      <td>148.250000</td>\n",
       "      <td>148.250000</td>\n",
       "      <td>143.875000</td>\n",
       "      <td>145.437500</td>\n",
       "      <td>8164300.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>104.914399</td>\n",
       "      <td>104.914399</td>\n",
       "      <td>101.818275</td>\n",
       "      <td>102.924034</td>\n",
       "      <td>8164300.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-04</th>\n",
       "      <td>143.531204</td>\n",
       "      <td>144.062500</td>\n",
       "      <td>139.640594</td>\n",
       "      <td>139.750000</td>\n",
       "      <td>8089800.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>101.574975</td>\n",
       "      <td>101.950966</td>\n",
       "      <td>98.821646</td>\n",
       "      <td>98.899071</td>\n",
       "      <td>8089800.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-05</th>\n",
       "      <td>139.937500</td>\n",
       "      <td>141.531204</td>\n",
       "      <td>137.250000</td>\n",
       "      <td>140.000000</td>\n",
       "      <td>12177900.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>99.031762</td>\n",
       "      <td>100.159604</td>\n",
       "      <td>97.129857</td>\n",
       "      <td>99.075993</td>\n",
       "      <td>12177900.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-06</th>\n",
       "      <td>139.625000</td>\n",
       "      <td>141.500000</td>\n",
       "      <td>137.750000</td>\n",
       "      <td>137.750000</td>\n",
       "      <td>6227200.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>98.810611</td>\n",
       "      <td>100.137521</td>\n",
       "      <td>97.483700</td>\n",
       "      <td>97.483700</td>\n",
       "      <td>6227200.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-07</th>\n",
       "      <td>140.312500</td>\n",
       "      <td>145.750000</td>\n",
       "      <td>140.062500</td>\n",
       "      <td>145.750000</td>\n",
       "      <td>8066500.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>99.297145</td>\n",
       "      <td>103.145185</td>\n",
       "      <td>99.120223</td>\n",
       "      <td>103.145185</td>\n",
       "      <td>8066500.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-10</th>\n",
       "      <td>146.250000</td>\n",
       "      <td>146.906204</td>\n",
       "      <td>145.031204</td>\n",
       "      <td>146.250000</td>\n",
       "      <td>5741700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>103.499028</td>\n",
       "      <td>103.963414</td>\n",
       "      <td>102.636504</td>\n",
       "      <td>103.499028</td>\n",
       "      <td>5741700.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>52.434457</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-11</th>\n",
       "      <td>145.812500</td>\n",
       "      <td>146.093704</td>\n",
       "      <td>143.500000</td>\n",
       "      <td>144.500000</td>\n",
       "      <td>7503700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>103.189416</td>\n",
       "      <td>103.388420</td>\n",
       "      <td>101.552893</td>\n",
       "      <td>102.260578</td>\n",
       "      <td>7503700.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>46.357616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-12</th>\n",
       "      <td>144.593704</td>\n",
       "      <td>144.593704</td>\n",
       "      <td>142.875000</td>\n",
       "      <td>143.062500</td>\n",
       "      <td>6907700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>102.326891</td>\n",
       "      <td>102.326891</td>\n",
       "      <td>101.110589</td>\n",
       "      <td>101.243280</td>\n",
       "      <td>6907700.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>41.427779</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-13</th>\n",
       "      <td>144.468704</td>\n",
       "      <td>145.750000</td>\n",
       "      <td>143.281204</td>\n",
       "      <td>145.000000</td>\n",
       "      <td>5158300.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>102.238430</td>\n",
       "      <td>103.145185</td>\n",
       "      <td>101.398054</td>\n",
       "      <td>102.614421</td>\n",
       "      <td>5158300.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>50.327412</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-14</th>\n",
       "      <td>146.531204</td>\n",
       "      <td>147.468704</td>\n",
       "      <td>145.968704</td>\n",
       "      <td>146.968704</td>\n",
       "      <td>7437300.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>103.698032</td>\n",
       "      <td>104.361487</td>\n",
       "      <td>103.299959</td>\n",
       "      <td>104.007645</td>\n",
       "      <td>7437300.0</td>\n",
       "      <td>-1.563531</td>\n",
       "      <td>NaN</td>\n",
       "      <td>58.362860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-18</th>\n",
       "      <td>145.343704</td>\n",
       "      <td>146.625000</td>\n",
       "      <td>145.187500</td>\n",
       "      <td>145.812500</td>\n",
       "      <td>6488500.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>102.857655</td>\n",
       "      <td>103.764410</td>\n",
       "      <td>102.747112</td>\n",
       "      <td>103.189416</td>\n",
       "      <td>6488500.0</td>\n",
       "      <td>-2.442698</td>\n",
       "      <td>NaN</td>\n",
       "      <td>52.167617</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-19</th>\n",
       "      <td>145.312500</td>\n",
       "      <td>147.000000</td>\n",
       "      <td>145.000000</td>\n",
       "      <td>147.000000</td>\n",
       "      <td>6157900.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>102.835573</td>\n",
       "      <td>104.029792</td>\n",
       "      <td>102.614421</td>\n",
       "      <td>104.029792</td>\n",
       "      <td>6157900.0</td>\n",
       "      <td>-2.384364</td>\n",
       "      <td>NaN</td>\n",
       "      <td>57.904391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-20</th>\n",
       "      <td>146.968704</td>\n",
       "      <td>146.968704</td>\n",
       "      <td>143.812500</td>\n",
       "      <td>144.750000</td>\n",
       "      <td>5800100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>104.007645</td>\n",
       "      <td>104.007645</td>\n",
       "      <td>101.774044</td>\n",
       "      <td>102.437500</td>\n",
       "      <td>5800100.0</td>\n",
       "      <td>-1.169796</td>\n",
       "      <td>-1.297712</td>\n",
       "      <td>45.094896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-21</th>\n",
       "      <td>145.500000</td>\n",
       "      <td>145.500000</td>\n",
       "      <td>144.062500</td>\n",
       "      <td>144.437500</td>\n",
       "      <td>6244800.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>102.968264</td>\n",
       "      <td>102.968264</td>\n",
       "      <td>101.950966</td>\n",
       "      <td>102.216348</td>\n",
       "      <td>6244800.0</td>\n",
       "      <td>-0.042713</td>\n",
       "      <td>-0.989094</td>\n",
       "      <td>43.427041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-24</th>\n",
       "      <td>145.656204</td>\n",
       "      <td>145.843704</td>\n",
       "      <td>139.406204</td>\n",
       "      <td>140.343704</td>\n",
       "      <td>7896900.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>103.078807</td>\n",
       "      <td>103.211498</td>\n",
       "      <td>98.655772</td>\n",
       "      <td>99.319227</td>\n",
       "      <td>7896900.0</td>\n",
       "      <td>1.635423</td>\n",
       "      <td>-0.174946</td>\n",
       "      <td>27.046502</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-25</th>\n",
       "      <td>140.515594</td>\n",
       "      <td>141.937500</td>\n",
       "      <td>139.000000</td>\n",
       "      <td>141.937500</td>\n",
       "      <td>9942500.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>99.440871</td>\n",
       "      <td>100.447134</td>\n",
       "      <td>98.368307</td>\n",
       "      <td>100.447134</td>\n",
       "      <td>9942500.0</td>\n",
       "      <td>2.141673</td>\n",
       "      <td>0.102968</td>\n",
       "      <td>38.361106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-26</th>\n",
       "      <td>141.000000</td>\n",
       "      <td>141.546799</td>\n",
       "      <td>140.093704</td>\n",
       "      <td>140.812500</td>\n",
       "      <td>5158100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>99.783678</td>\n",
       "      <td>100.170640</td>\n",
       "      <td>99.142306</td>\n",
       "      <td>99.650987</td>\n",
       "      <td>5158100.0</td>\n",
       "      <td>2.981256</td>\n",
       "      <td>0.429631</td>\n",
       "      <td>33.743535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-27</th>\n",
       "      <td>141.843704</td>\n",
       "      <td>142.218704</td>\n",
       "      <td>138.125000</td>\n",
       "      <td>140.250000</td>\n",
       "      <td>10922700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>100.380756</td>\n",
       "      <td>100.646138</td>\n",
       "      <td>97.749082</td>\n",
       "      <td>99.252914</td>\n",
       "      <td>10922700.0</td>\n",
       "      <td>2.731241</td>\n",
       "      <td>0.692450</td>\n",
       "      <td>31.382562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-28</th>\n",
       "      <td>139.437500</td>\n",
       "      <td>140.062500</td>\n",
       "      <td>135.531204</td>\n",
       "      <td>135.875000</td>\n",
       "      <td>11916200.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>98.677920</td>\n",
       "      <td>99.120223</td>\n",
       "      <td>95.913490</td>\n",
       "      <td>96.156789</td>\n",
       "      <td>11916200.0</td>\n",
       "      <td>3.839574</td>\n",
       "      <td>1.450097</td>\n",
       "      <td>18.677364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-31</th>\n",
       "      <td>135.812500</td>\n",
       "      <td>139.671799</td>\n",
       "      <td>135.000000</td>\n",
       "      <td>139.562500</td>\n",
       "      <td>10768700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>96.112559</td>\n",
       "      <td>98.843730</td>\n",
       "      <td>95.537564</td>\n",
       "      <td>98.766380</td>\n",
       "      <td>10768700.0</td>\n",
       "      <td>3.515620</td>\n",
       "      <td>1.327988</td>\n",
       "      <td>42.992984</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Open        High         Low       Close      Volume  \\\n",
       "Date                                                                     \n",
       "2000-01-03  148.250000  148.250000  143.875000  145.437500   8164300.0   \n",
       "2000-01-04  143.531204  144.062500  139.640594  139.750000   8089800.0   \n",
       "2000-01-05  139.937500  141.531204  137.250000  140.000000  12177900.0   \n",
       "2000-01-06  139.625000  141.500000  137.750000  137.750000   6227200.0   \n",
       "2000-01-07  140.312500  145.750000  140.062500  145.750000   8066500.0   \n",
       "2000-01-10  146.250000  146.906204  145.031204  146.250000   5741700.0   \n",
       "2000-01-11  145.812500  146.093704  143.500000  144.500000   7503700.0   \n",
       "2000-01-12  144.593704  144.593704  142.875000  143.062500   6907700.0   \n",
       "2000-01-13  144.468704  145.750000  143.281204  145.000000   5158300.0   \n",
       "2000-01-14  146.531204  147.468704  145.968704  146.968704   7437300.0   \n",
       "2000-01-18  145.343704  146.625000  145.187500  145.812500   6488500.0   \n",
       "2000-01-19  145.312500  147.000000  145.000000  147.000000   6157900.0   \n",
       "2000-01-20  146.968704  146.968704  143.812500  144.750000   5800100.0   \n",
       "2000-01-21  145.500000  145.500000  144.062500  144.437500   6244800.0   \n",
       "2000-01-24  145.656204  145.843704  139.406204  140.343704   7896900.0   \n",
       "2000-01-25  140.515594  141.937500  139.000000  141.937500   9942500.0   \n",
       "2000-01-26  141.000000  141.546799  140.093704  140.812500   5158100.0   \n",
       "2000-01-27  141.843704  142.218704  138.125000  140.250000  10922700.0   \n",
       "2000-01-28  139.437500  140.062500  135.531204  135.875000  11916200.0   \n",
       "2000-01-31  135.812500  139.671799  135.000000  139.562500  10768700.0   \n",
       "\n",
       "            Dividend  Split    Adj_Open    Adj_High     Adj_Low   Adj_Close  \\\n",
       "Date                                                                          \n",
       "2000-01-03       0.0    1.0  104.914399  104.914399  101.818275  102.924034   \n",
       "2000-01-04       0.0    1.0  101.574975  101.950966   98.821646   98.899071   \n",
       "2000-01-05       0.0    1.0   99.031762  100.159604   97.129857   99.075993   \n",
       "2000-01-06       0.0    1.0   98.810611  100.137521   97.483700   97.483700   \n",
       "2000-01-07       0.0    1.0   99.297145  103.145185   99.120223  103.145185   \n",
       "2000-01-10       0.0    1.0  103.499028  103.963414  102.636504  103.499028   \n",
       "2000-01-11       0.0    1.0  103.189416  103.388420  101.552893  102.260578   \n",
       "2000-01-12       0.0    1.0  102.326891  102.326891  101.110589  101.243280   \n",
       "2000-01-13       0.0    1.0  102.238430  103.145185  101.398054  102.614421   \n",
       "2000-01-14       0.0    1.0  103.698032  104.361487  103.299959  104.007645   \n",
       "2000-01-18       0.0    1.0  102.857655  103.764410  102.747112  103.189416   \n",
       "2000-01-19       0.0    1.0  102.835573  104.029792  102.614421  104.029792   \n",
       "2000-01-20       0.0    1.0  104.007645  104.007645  101.774044  102.437500   \n",
       "2000-01-21       0.0    1.0  102.968264  102.968264  101.950966  102.216348   \n",
       "2000-01-24       0.0    1.0  103.078807  103.211498   98.655772   99.319227   \n",
       "2000-01-25       0.0    1.0   99.440871  100.447134   98.368307  100.447134   \n",
       "2000-01-26       0.0    1.0   99.783678  100.170640   99.142306   99.650987   \n",
       "2000-01-27       0.0    1.0  100.380756  100.646138   97.749082   99.252914   \n",
       "2000-01-28       0.0    1.0   98.677920   99.120223   95.913490   96.156789   \n",
       "2000-01-31       0.0    1.0   96.112559   98.843730   95.537564   98.766380   \n",
       "\n",
       "            Adj_Volume  sma_diff  ema_diff        rsi  \n",
       "Date                                                   \n",
       "2000-01-03   8164300.0       NaN       NaN        NaN  \n",
       "2000-01-04   8089800.0       NaN       NaN        NaN  \n",
       "2000-01-05  12177900.0       NaN       NaN        NaN  \n",
       "2000-01-06   6227200.0       NaN       NaN        NaN  \n",
       "2000-01-07   8066500.0       NaN       NaN        NaN  \n",
       "2000-01-10   5741700.0       NaN       NaN  52.434457  \n",
       "2000-01-11   7503700.0       NaN       NaN  46.357616  \n",
       "2000-01-12   6907700.0       NaN       NaN  41.427779  \n",
       "2000-01-13   5158300.0       NaN       NaN  50.327412  \n",
       "2000-01-14   7437300.0 -1.563531       NaN  58.362860  \n",
       "2000-01-18   6488500.0 -2.442698       NaN  52.167617  \n",
       "2000-01-19   6157900.0 -2.384364       NaN  57.904391  \n",
       "2000-01-20   5800100.0 -1.169796 -1.297712  45.094896  \n",
       "2000-01-21   6244800.0 -0.042713 -0.989094  43.427041  \n",
       "2000-01-24   7896900.0  1.635423 -0.174946  27.046502  \n",
       "2000-01-25   9942500.0  2.141673  0.102968  38.361106  \n",
       "2000-01-26   5158100.0  2.981256  0.429631  33.743535  \n",
       "2000-01-27  10922700.0  2.731241  0.692450  31.382562  \n",
       "2000-01-28  11916200.0  3.839574  1.450097  18.677364  \n",
       "2000-01-31  10768700.0  3.515620  1.327988  42.992984  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.drop(columns=['Open','High','Low','Adj_Open','Adj_High','Adj_Low','Adj_Volume','Adj_Close', 'Dividend', 'Split', 'Volume'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.dropna(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>Close</th>\n",
       "      <th>sma_diff</th>\n",
       "      <th>ema_diff</th>\n",
       "      <th>rsi</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-06-18</th>\n",
       "      <td>276.56</td>\n",
       "      <td>0.125667</td>\n",
       "      <td>-0.920804</td>\n",
       "      <td>42.957112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-19</th>\n",
       "      <td>275.50</td>\n",
       "      <td>1.242333</td>\n",
       "      <td>-0.520437</td>\n",
       "      <td>33.402330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-20</th>\n",
       "      <td>275.97</td>\n",
       "      <td>1.486000</td>\n",
       "      <td>-0.321215</td>\n",
       "      <td>40.711352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-21</th>\n",
       "      <td>274.24</td>\n",
       "      <td>1.946333</td>\n",
       "      <td>0.061011</td>\n",
       "      <td>27.051420</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-22</th>\n",
       "      <td>274.74</td>\n",
       "      <td>1.854667</td>\n",
       "      <td>0.221109</td>\n",
       "      <td>34.938091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-25</th>\n",
       "      <td>271.00</td>\n",
       "      <td>2.755333</td>\n",
       "      <td>0.844388</td>\n",
       "      <td>17.374752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-26</th>\n",
       "      <td>271.60</td>\n",
       "      <td>2.903333</td>\n",
       "      <td>1.105809</td>\n",
       "      <td>24.941313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-27</th>\n",
       "      <td>269.35</td>\n",
       "      <td>3.832000</td>\n",
       "      <td>1.542156</td>\n",
       "      <td>17.450428</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-28</th>\n",
       "      <td>270.89</td>\n",
       "      <td>3.084667</td>\n",
       "      <td>1.526362</td>\n",
       "      <td>34.325907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-29</th>\n",
       "      <td>271.28</td>\n",
       "      <td>2.606333</td>\n",
       "      <td>1.398678</td>\n",
       "      <td>38.317594</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Close  sma_diff  ema_diff        rsi\n",
       "Date                                             \n",
       "2018-06-18  276.56  0.125667 -0.920804  42.957112\n",
       "2018-06-19  275.50  1.242333 -0.520437  33.402330\n",
       "2018-06-20  275.97  1.486000 -0.321215  40.711352\n",
       "2018-06-21  274.24  1.946333  0.061011  27.051420\n",
       "2018-06-22  274.74  1.854667  0.221109  34.938091\n",
       "2018-06-25  271.00  2.755333  0.844388  17.374752\n",
       "2018-06-26  271.60  2.903333  1.105809  24.941313\n",
       "2018-06-27  269.35  3.832000  1.542156  17.450428\n",
       "2018-06-28  270.89  3.084667  1.526362  34.325907\n",
       "2018-06-29  271.28  2.606333  1.398678  38.317594"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Save Scoring Test Dataset as pickle for later use\n",
    "test_df = data.iloc[-10:]\n",
    "test_df.to_pickle(TEST_DATA_PATH)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting/Visualization of the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'ema_diff')"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x648 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(16,9))\n",
    "top = plt.subplot2grid((4,4), (0, 0), rowspan=3, colspan=4)\n",
    "bottom = plt.subplot2grid((4,4), (3,0), rowspan=1, colspan=4)\n",
    "top.plot(data.index, data['Close']) #CMT.index gives the dates\n",
    "bottom.bar(data.index, data['ema_diff']) \n",
    " \n",
    "# set the labels\n",
    "top.axes.get_xaxis().set_visible(False)\n",
    "top.set_title('CapitalMall Trust')\n",
    "top.set_ylabel('Closing Price')\n",
    "bottom.set_ylabel('ema_diff')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda/envs/myenv/lib/python3.6/site-packages/matplotlib/axes/_axes.py:6462: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f9a767bfa58>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "plt.figure(figsize=(16,9))\n",
    "sns.distplot(data['Close'].dropna(), bins=50, color='purple')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# simple moving averages\n",
    "if use_features:\n",
    "    sma = pandas.DataFrame({\"RSI\": data['rsi'][:90], 'EMA DIFF': data['ema_diff'][:90], 'SMA DIFF': data['sma_diff'][:90]})\n",
    "    sma.plot(figsize=(16, 9), legend=True, title=TICKER)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Add the Target column -> Price at X days in the future"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "shift_days_target = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "run.log('shift_days_target', shift_days_target)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Add the target (price X days in the future)\n",
    "data['target'] = data.Close.shift(-shift_days_target)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.dropna(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "data.head(15) ## df[::-1] inverts the matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Normalization and Columns pick (without Scikit-learn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "def normalize(df):\n",
    "    result = df.copy()\n",
    "    min_dict = {}\n",
    "    max_dict = {}\n",
    "    mean_dict = {}\n",
    "    std_dict = {}\n",
    "    \n",
    "    for feature_name in df.columns:\n",
    "        min_value = df[feature_name].min()\n",
    "        max_value = df[feature_name].max()     \n",
    "        min_dict[feature_name] = min_value;\n",
    "        max_dict[feature_name] = max_value;\n",
    "        result[feature_name] = (df[feature_name] - min_value) / (max_value - min_value)\n",
    "        \n",
    "    min_max_dict_list = [min_dict, max_dict]\n",
    "    \n",
    "    return result, min_max_dict_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df,  min_max_dict_list = normalize(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "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>Close</th>\n",
       "      <th>sma_diff</th>\n",
       "      <th>ema_diff</th>\n",
       "      <th>rsi</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2000-01-20</th>\n",
       "      <td>0.350803</td>\n",
       "      <td>0.265638</td>\n",
       "      <td>0.207627</td>\n",
       "      <td>0.437583</td>\n",
       "      <td>0.337930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-21</th>\n",
       "      <td>0.349373</td>\n",
       "      <td>0.326314</td>\n",
       "      <td>0.234710</td>\n",
       "      <td>0.419709</td>\n",
       "      <td>0.332780</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-24</th>\n",
       "      <td>0.330634</td>\n",
       "      <td>0.416655</td>\n",
       "      <td>0.306156</td>\n",
       "      <td>0.244165</td>\n",
       "      <td>0.330206</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-25</th>\n",
       "      <td>0.337930</td>\n",
       "      <td>0.443909</td>\n",
       "      <td>0.330545</td>\n",
       "      <td>0.365419</td>\n",
       "      <td>0.310180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-26</th>\n",
       "      <td>0.332780</td>\n",
       "      <td>0.489108</td>\n",
       "      <td>0.359211</td>\n",
       "      <td>0.315935</td>\n",
       "      <td>0.327059</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  sma_diff  ema_diff       rsi    target\n",
       "Date                                                        \n",
       "2000-01-20  0.350803  0.265638  0.207627  0.437583  0.337930\n",
       "2000-01-21  0.349373  0.326314  0.234710  0.419709  0.332780\n",
       "2000-01-24  0.330634  0.416655  0.306156  0.244165  0.330206\n",
       "2000-01-25  0.337930  0.443909  0.330545  0.365419  0.310180\n",
       "2000-01-26  0.332780  0.489108  0.359211  0.315935  0.327059"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'Close': 68.110001, 'sma_diff': -6.104154400000056, 'ema_diff': -3.66368113102817, 'rsi': 4.26271527994781, 'target': 68.110001}\n",
      "{'Close': 286.58, 'sma_diff': 12.471332266666792, 'ema_diff': 7.731594856915947, 'rsi': 97.57572579549975, 'target': 286.58}\n"
     ]
    }
   ],
   "source": [
    "print(min_max_dict_list[0]) # Min\n",
    "print(min_max_dict_list[1]) # Max"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Saving the MIN and MAX values of each column\n",
    "We need to save these values max_dict and min_dict into a pickle file, because we are going to need them later in operationalization of the model in order to transfor the realtime data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min-Max values saved\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<azureml.restclient.models.batch_artifact_content_information_dto.BatchArtifactContentInformationDto at 0x7f965c1094a8>"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pickle\n",
    "\n",
    "with open(MIN_MAX_DICT, 'wb') as handle:\n",
    "    pickle.dump(min_max_dict_list, handle, protocol=pickle.HIGHEST_PROTOCOL)\n",
    "print(\"Min-Max values saved\")\n",
    "\n",
    "run.upload_file(name = 'outputs/'+ TICKER +'-min_max.pkl', path_or_stream = MODELS_DIR + TICKER +'-min_max.pkl' )\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Function to load/prepare data for modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "def prepare_data_for_modeling(stock, seq_len):\n",
    "    data = stock.values #converts to numpy\n",
    "    result = []\n",
    "    for index in range(len(data) - seq_len + 1):\n",
    "        result.append(data[index: index + seq_len])\n",
    "\n",
    "    result = np.array(result)    \n",
    "    row = round(0.95 * result.shape[0])\n",
    "    train = result[:int(row), :]\n",
    "    test =  result[int(row):, :]\n",
    "    x_train = train[:,:,:-1] # All rows, all time steps, all feautures minus de last feature\n",
    "    y_train = train[:,-1,-1] # All rows, only the last step, the last feature\n",
    "    x_test = test[:,:,:-1] \n",
    "    y_test = test[:,-1,-1]\n",
    "\n",
    "    return [x_train, y_train, x_test, y_test]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Function to define the RNN model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "from livelossplot import PlotLossesKeras\n",
    "\n",
    "params = {'Dense': 32,\n",
    "          'Dense_1': 32,\n",
    "          'Dropout': 0.35535607083673604,\n",
    "          'LSTM': 156,\n",
    "          'batch_size': 128,\n",
    "          'epochs': 1500,\n",
    "          'lr': 0.00082,\n",
    "          'patience': 100,\n",
    "          'validation_split': 0.1\n",
    "         }\n",
    "\n",
    "run.log(\"LSTM\", params['LSTM'])\n",
    "run.log(\"Dropout\", params['Dropout'])\n",
    "run.log(\"Dense\", params['Dense'])\n",
    "run.log(\"Dense_1\", params['Dense_1'])\n",
    "run.log(\"lr\", params['lr'])\n",
    "run.log(\"epochs\", params['epochs'])\n",
    "run.log(\"patience\", params['patience'])\n",
    "run.log(\"validation_split\", params['validation_split'])\n",
    "\n",
    "\n",
    "def build_model(X_train, y_train):\n",
    "    model = Sequential()\n",
    "    model.add(LSTM(params['LSTM'], input_shape=(len(X_train[0]), len(X_train[0][0])),return_sequences=False))\n",
    "    model.add(Dropout(params['Dropout']))\n",
    "    model.add(Dense(params['Dense'], kernel_initializer=\"uniform\", activation='relu'))\n",
    "    model.add(Dense(params['Dense_1'], kernel_initializer=\"uniform\", activation='relu'))\n",
    "    model.add(Dense(1, kernel_initializer=\"uniform\", activation='relu'))\n",
    "    \n",
    "    adam = keras.optimizers.Adam(lr=params['lr'])\n",
    "    model.compile(loss='mse', optimizer=adam, metrics=['mae'])\n",
    "\n",
    "    # checkpoint\n",
    "    checkpoint = keras.callbacks.ModelCheckpoint(MODEL_WEIGHTS_BEST, \n",
    "                                                 monitor='val_loss', \n",
    "                                                 verbose=0, save_best_only=True, mode='min')\n",
    "    # Early Stop\n",
    "    earlystop = keras.callbacks.EarlyStopping(monitor='val_loss',min_delta=0, patience=params['patience'],verbose=1, mode='min')\n",
    "\n",
    "    history = model.fit(X_train,y_train,\n",
    "                        batch_size=params['batch_size'],\n",
    "                        epochs=params['epochs'],\n",
    "                        validation_split=params['validation_split'],\n",
    "                        verbose=0,\n",
    "                        shuffle=True,\n",
    "                        callbacks=[PlotLossesKeras(), checkpoint, earlystop])\n",
    "\n",
    "\n",
    "    return history, model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train (4398, 10, 4)\n",
      "y_train (4398,)\n",
      "X_test (231, 10, 4)\n",
      "y_test (231,)\n"
     ]
    }
   ],
   "source": [
    "window = 10\n",
    "X_train, y_train, X_test, y_test = prepare_data_for_modeling(df, window) \n",
    "print(\"X_train\", X_train.shape)\n",
    "print(\"y_train\", y_train.shape)\n",
    "print(\"X_test\", X_test.shape)\n",
    "print(\"y_test\", y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 00563: early stopping\n",
      "CPU times: user 9min 22s, sys: 1min 21s, total: 10min 44s\n",
      "Wall time: 6min 24s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "history, model = build_model(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Saving the model\n",
    "The LSTM network is made up of two components, the architecture and the model weights. We'll save these model components in two files, the architecture in a json file that the keras package can use to rebuild the model, and the weights of the last iteration/epoch in an HDF5 heirachy that rebuild the exact model. (the weights for the lowest loss epoch are saved by the checkpoint callback keras function)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "    \n",
    "# save model\n",
    "# serialize model to JSON\n",
    "model_json = model.to_json()\n",
    "with open(LSTM_MODEL, \"w\") as json_file:\n",
    "    json_file.write(model_json)\n",
    "    \n",
    "testScore= model.evaluate(X_test, y_test, verbose=0)\n",
    "\n",
    "# save the model+weights with a unique identifier\n",
    "modelname = TICKER + \"-model-lstm-\" + os.environ['KERAS_BACKEND'] + \"-\" + str(shift_days_target) + \"day-\" + str(testScore[0]) + \".h5\"\n",
    "model.save(MODELS_DIR + modelname)\n",
    "\n",
    "run.upload_file(name = 'outputs/'+ modelname, path_or_stream = MODELS_DIR + modelname )\n",
    "    \n",
    "# serialize the last epoch weights to HDF5\n",
    "model.save_weights(MODEL_WEIGHTS)\n",
    "print(\"Model saved\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Check and plot results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import random"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Naive evaluation (assume that pred = current day close)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.00025187526\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "\n",
    "y_naive = np.roll(y_test,shift_days_target)\n",
    "naiveScore = mean_squared_error(y_test[shift_days_target:], y_naive[shift_days_target:], multioutput='uniform_average')\n",
    "\n",
    "run.log('mse_naive', naiveScore)\n",
    "print(naiveScore)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Last Weights loaded\n",
      "Train Score: 0.00015654 MSE (0.01251 RMSE)\n",
      "Test Score: 0.00671027 MSE (0.08192 RMSE)\n"
     ]
    }
   ],
   "source": [
    "# load last epoch weights into new model\n",
    "model.load_weights(MODEL_WEIGHTS)\n",
    "print(\"Last Weights loaded\")\n",
    "\n",
    "trainScore = model.evaluate(X_train, y_train, verbose=0)\n",
    "print('Train Score: %.8f MSE (%.5f RMSE)' % (trainScore[0], math.sqrt(trainScore[0])))\n",
    "\n",
    "testScore = model.evaluate(X_test, y_test, verbose=0)\n",
    "print('Test Score: %.8f MSE (%.5f RMSE)' % (testScore[0], math.sqrt(testScore[0])))\n",
    "\n",
    "run.log('mse_last_iter', testScore[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "yhat = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Weights loaded\n",
      "Train Score: 0.00012741 MSE (0.01129 RMSE)\n",
      "Test Score: 0.00071856 MSE (0.02681 RMSE)\n"
     ]
    }
   ],
   "source": [
    "# load best epoch weights into new model\n",
    "model.load_weights(MODEL_WEIGHTS_BEST)\n",
    "print(\"Best Weights loaded\")\n",
    "\n",
    "trainScore_best = model.evaluate(X_train, y_train, verbose=0)\n",
    "print('Train Score: %.8f MSE (%.5f RMSE)' % (trainScore_best[0], math.sqrt(trainScore_best[0])))\n",
    "\n",
    "testScore_best = model.evaluate(X_test, y_test, verbose=0)\n",
    "print('Test Score: %.8f MSE (%.5f RMSE)' % (testScore_best[0], math.sqrt(testScore_best[0])))\n",
    "\n",
    "run.log('mse_best_iter', testScore_best[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "yhatbest = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Let's revert the normalization before plotting with this fucntion\n",
    "def denormalize(array,min_max_dict_list):\n",
    "    result = array\n",
    "    min_dict = min_max_dict_list[0]\n",
    "    max_dict = min_max_dict_list[1]\n",
    "    result = array * (max_dict[\"Close\"] - min_dict[\"Close\"]) + min_dict[\"Close\"]\n",
    "    \n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_naive_inv = denormalize(y_naive, min_max_dict_list)\n",
    "yhat_inv = denormalize(yhat, min_max_dict_list)\n",
    "yhatbest_inv = denormalize(yhatbest, min_max_dict_list)\n",
    "y_test_inv = denormalize(y_test, min_max_dict_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "IPython.OutputArea.prototype._should_scroll = function(lines) {\n",
       "    return false;\n",
       "}"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%javascript\n",
    "IPython.OutputArea.prototype._should_scroll = function(lines) {\n",
    "    return false;\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "run id: 69b6d590-89e7-426a-9bbb-5df04d683a29\n"
     ]
    }
   ],
   "source": [
    "run.complete()\n",
    "run_id = run.id\n",
    "print (\"run id:\", run.id)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams[\"figure.figsize\"] = [16,9]\n",
    "#plt.plot(yhat_inv[shift_days_target:],color='red', label='prediction')\n",
    "plt.plot(yhatbest_inv,color='green', label='prediction_best')\n",
    "plt.plot(y_naive_inv[shift_days_target:],color='orange', label='prediction_naive')\n",
    "plt.plot(y_test_inv[shift_days_target:],color='blue', label='y_test')\n",
    "plt.legend(loc='lower left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can view the UI here: https://aka.ms/mlextensions_dev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "69b6d590-89e7-426a-9bbb-5df04d683a29\n"
     ]
    }
   ],
   "source": [
    "rid = list(Run.list(workspace=ws,history_name=run_history_name))[0].id\n",
    "print (rid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "run = Run(workspace=ws,history_name=run_history_name, run_id=rid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'ticker': 'SPY',\n",
       " 'use_features': 'yes',\n",
       " 'SMA_FAST': 3,\n",
       " 'SMA_SLOW': 10,\n",
       " 'EMA_FAST': 6,\n",
       " 'EMA_SLOW': 13,\n",
       " 'RSI_PERIOD': 5,\n",
       " 'shift_days_target': 3,\n",
       " 'LSTM': 156,\n",
       " 'Dropout': 0.35535607083673604,\n",
       " 'Dense': 32,\n",
       " 'Dense_1': 32,\n",
       " 'lr': 0.00082,\n",
       " 'epochs': 1500,\n",
       " 'patience': 250,\n",
       " 'validation_split': 0.1,\n",
       " 'mse_naive': [0.00025187525898218155, 0.00025187525898218155],\n",
       " 'mse_last_iter': 0.006710271673575366,\n",
       " 'mse_best_iter': 0.0007185635940087112}"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get all metris logged in the run\n",
    "run.get_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:azureml._run_impl.run_base.69b6d590-89e7-426a-9bbb-5df04d683a29:Unstable, may be removed in the future\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "{'ticker': 'SPY', 'use_features': 'yes', 'SMA_FAST': 3, 'SMA_SLOW': 10, 'EMA_FAST': 6, 'EMA_SLOW': 13, 'RSI_PERIOD': 5, 'shift_days_target': 3, 'LSTM': 156, 'Dropout': 0.35535607083673604, 'Dense': 32, 'Dense_1': 32, 'lr': 0.00082, 'epochs': 1500, 'patience': 250, 'validation_split': 0.1, 'mse_naive': [0.00025187525898218155, 0.00025187525898218155], 'mse_last_iter': 0.006710271673575366, 'mse_best_iter': 0.0007185635940087112}\n",
      "['outputs/SPY-min_max.pkl', 'outputs/SPY-model-lstm-tensorflow-3day-0.006710271673575366.h5']\n"
     ]
    }
   ],
   "source": [
    "runs = run.history.get_runs()\n",
    "\n",
    "for r in runs:\n",
    "    print (r.number)\n",
    "    print (r.get_metrics())\n",
    "    print (r.get_file_names())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load saved model "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To test the save operations, we can reload the model files into a test model loaded_model and rescore the test dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Loaded\n",
      "Weights loaded\n",
      "Min_max List loaded\n"
     ]
    }
   ],
   "source": [
    "from keras.models import model_from_json\n",
    "\n",
    "# load json and create model\n",
    "json_file = open(LSTM_MODEL, 'r')\n",
    "loaded_model_json = json_file.read()\n",
    "json_file.close()\n",
    "loaded_model = model_from_json(loaded_model_json)\n",
    "print(\"Model Loaded\")\n",
    "\n",
    "# load weights into new model\n",
    "loaded_model.load_weights(MODEL_WEIGHTS_BEST)\n",
    "print(\"Weights loaded\")\n",
    "\n",
    "# Load Min Max list values\n",
    "with open(MIN_MAX_DICT, 'rb') as handle:\n",
    "    loaded_min_max_dict_list = pickle.load(handle)\n",
    "\n",
    "if (min_max_dict_list == loaded_min_max_dict_list):\n",
    "    print(\"Min_max List loaded\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model constructed from storage can be used to predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[[0.81223965 0.22824828 0.2643392  0.7813765 ]\n",
      "  [0.8121023  0.2210524  0.25132695 0.7750538 ]\n",
      "  [0.8126974  0.21737368 0.2445951  0.78503126]\n",
      "  [0.8187852  0.2144307  0.22693929 0.8662528 ]\n",
      "  [0.8192887  0.22151889 0.21861058 0.87163174]\n",
      "  [0.8182817  0.22530523 0.22046934 0.80040014]\n",
      "  [0.81800705 0.25231215 0.2264836  0.77857995]\n",
      "  [0.8207534  0.2747252  0.22652246 0.8388918 ]\n",
      "  [0.8207992  0.28326693 0.2302979  0.8398372 ]\n",
      "  [0.81974643 0.29141387 0.23921566 0.7275096 ]]]\n",
      "(1, 10, 4)\n",
      "(1, 1)\n",
      "[[249.22452]]\n"
     ]
    }
   ],
   "source": [
    "test_row = X_test[:1]\n",
    "print(test_row)\n",
    "print(test_row.shape)\n",
    "score = loaded_model.predict(test_row)\n",
    "print(score.shape)\n",
    "print(denormalize(score,min_max_dict_list))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Score on Test df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "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>Close</th>\n",
       "      <th>sma_diff</th>\n",
       "      <th>ema_diff</th>\n",
       "      <th>rsi</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-06-18</th>\n",
       "      <td>276.56</td>\n",
       "      <td>0.125667</td>\n",
       "      <td>-0.920804</td>\n",
       "      <td>42.957112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-19</th>\n",
       "      <td>275.50</td>\n",
       "      <td>1.242333</td>\n",
       "      <td>-0.520437</td>\n",
       "      <td>33.402330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-20</th>\n",
       "      <td>275.97</td>\n",
       "      <td>1.486000</td>\n",
       "      <td>-0.321215</td>\n",
       "      <td>40.711352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-21</th>\n",
       "      <td>274.24</td>\n",
       "      <td>1.946333</td>\n",
       "      <td>0.061011</td>\n",
       "      <td>27.051420</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-22</th>\n",
       "      <td>274.74</td>\n",
       "      <td>1.854667</td>\n",
       "      <td>0.221109</td>\n",
       "      <td>34.938091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-25</th>\n",
       "      <td>271.00</td>\n",
       "      <td>2.755333</td>\n",
       "      <td>0.844388</td>\n",
       "      <td>17.374752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-26</th>\n",
       "      <td>271.60</td>\n",
       "      <td>2.903333</td>\n",
       "      <td>1.105809</td>\n",
       "      <td>24.941313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-27</th>\n",
       "      <td>269.35</td>\n",
       "      <td>3.832000</td>\n",
       "      <td>1.542156</td>\n",
       "      <td>17.450428</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-28</th>\n",
       "      <td>270.89</td>\n",
       "      <td>3.084667</td>\n",
       "      <td>1.526362</td>\n",
       "      <td>34.325907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-29</th>\n",
       "      <td>271.28</td>\n",
       "      <td>2.606333</td>\n",
       "      <td>1.398678</td>\n",
       "      <td>38.317594</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Close  sma_diff  ema_diff        rsi\n",
       "Date                                             \n",
       "2018-06-18  276.56  0.125667 -0.920804  42.957112\n",
       "2018-06-19  275.50  1.242333 -0.520437  33.402330\n",
       "2018-06-20  275.97  1.486000 -0.321215  40.711352\n",
       "2018-06-21  274.24  1.946333  0.061011  27.051420\n",
       "2018-06-22  274.74  1.854667  0.221109  34.938091\n",
       "2018-06-25  271.00  2.755333  0.844388  17.374752\n",
       "2018-06-26  271.60  2.903333  1.105809  24.941313\n",
       "2018-06-27  269.35  3.832000  1.542156  17.450428\n",
       "2018-06-28  270.89  3.084667  1.526362  34.325907\n",
       "2018-06-29  271.28  2.606333  1.398678  38.317594"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "def normalize2(df,min_max_dict_list):\n",
    "    # Normalize data\n",
    "    result = df.copy()\n",
    "    min_dict = min_max_dict_list[0]\n",
    "    max_dict = min_max_dict_list[1]\n",
    "    print(min_max_dict_list[0]) # Min\n",
    "    print(min_max_dict_list[1]) # Max\n",
    "    print(df.columns)\n",
    "    for feature_name in df.columns:\n",
    "        result[feature_name] = (df[feature_name] - min_dict[feature_name]) / (max_dict[feature_name] - min_dict[feature_name])\n",
    "        \n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'Close': 68.110001, 'sma_diff': -6.104154400000056, 'ema_diff': -3.66368113102817, 'rsi': 4.26271527994781, 'target': 68.110001}\n",
      "{'Close': 286.58, 'sma_diff': 12.471332266666792, 'ema_diff': 7.731594856915947, 'rsi': 97.57572579549975, 'target': 286.58}\n",
      "Index(['Close', 'sma_diff', 'ema_diff', 'rsi'], dtype='object')\n"
     ]
    },
    {
     "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>Close</th>\n",
       "      <th>ema_diff</th>\n",
       "      <th>rsi</th>\n",
       "      <th>sma_diff</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-06-18</th>\n",
       "      <td>0.954136</td>\n",
       "      <td>0.240703</td>\n",
       "      <td>0.414673</td>\n",
       "      <td>0.335379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-19</th>\n",
       "      <td>0.949284</td>\n",
       "      <td>0.275837</td>\n",
       "      <td>0.312278</td>\n",
       "      <td>0.395494</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-20</th>\n",
       "      <td>0.951435</td>\n",
       "      <td>0.293320</td>\n",
       "      <td>0.390606</td>\n",
       "      <td>0.408611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-21</th>\n",
       "      <td>0.943516</td>\n",
       "      <td>0.326863</td>\n",
       "      <td>0.244218</td>\n",
       "      <td>0.433393</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-22</th>\n",
       "      <td>0.945805</td>\n",
       "      <td>0.340912</td>\n",
       "      <td>0.328736</td>\n",
       "      <td>0.428458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-25</th>\n",
       "      <td>0.928686</td>\n",
       "      <td>0.395609</td>\n",
       "      <td>0.140517</td>\n",
       "      <td>0.476945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-26</th>\n",
       "      <td>0.931432</td>\n",
       "      <td>0.418550</td>\n",
       "      <td>0.221605</td>\n",
       "      <td>0.484913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-27</th>\n",
       "      <td>0.921133</td>\n",
       "      <td>0.456842</td>\n",
       "      <td>0.141328</td>\n",
       "      <td>0.534907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-28</th>\n",
       "      <td>0.928182</td>\n",
       "      <td>0.455456</td>\n",
       "      <td>0.322176</td>\n",
       "      <td>0.494675</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-29</th>\n",
       "      <td>0.929968</td>\n",
       "      <td>0.444251</td>\n",
       "      <td>0.364953</td>\n",
       "      <td>0.468924</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  ema_diff       rsi  sma_diff\n",
       "Date                                              \n",
       "2018-06-18  0.954136  0.240703  0.414673  0.335379\n",
       "2018-06-19  0.949284  0.275837  0.312278  0.395494\n",
       "2018-06-20  0.951435  0.293320  0.390606  0.408611\n",
       "2018-06-21  0.943516  0.326863  0.244218  0.433393\n",
       "2018-06-22  0.945805  0.340912  0.328736  0.428458\n",
       "2018-06-25  0.928686  0.395609  0.140517  0.476945\n",
       "2018-06-26  0.931432  0.418550  0.221605  0.484913\n",
       "2018-06-27  0.921133  0.456842  0.141328  0.534907\n",
       "2018-06-28  0.928182  0.455456  0.322176  0.494675\n",
       "2018-06-29  0.929968  0.444251  0.364953  0.468924"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df_n = normalize2(test_df, min_max_dict_list)\n",
    "test_df_n.reindex(sorted(test_df_n.columns), axis=1) # To make sure columns are always with same order"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "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>Close</th>\n",
       "      <th>sma_diff</th>\n",
       "      <th>ema_diff</th>\n",
       "      <th>rsi</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-06-18</th>\n",
       "      <td>0.954136</td>\n",
       "      <td>0.335379</td>\n",
       "      <td>0.240703</td>\n",
       "      <td>0.414673</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-19</th>\n",
       "      <td>0.949284</td>\n",
       "      <td>0.395494</td>\n",
       "      <td>0.275837</td>\n",
       "      <td>0.312278</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-20</th>\n",
       "      <td>0.951435</td>\n",
       "      <td>0.408611</td>\n",
       "      <td>0.293320</td>\n",
       "      <td>0.390606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-21</th>\n",
       "      <td>0.943516</td>\n",
       "      <td>0.433393</td>\n",
       "      <td>0.326863</td>\n",
       "      <td>0.244218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-22</th>\n",
       "      <td>0.945805</td>\n",
       "      <td>0.428458</td>\n",
       "      <td>0.340912</td>\n",
       "      <td>0.328736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-25</th>\n",
       "      <td>0.928686</td>\n",
       "      <td>0.476945</td>\n",
       "      <td>0.395609</td>\n",
       "      <td>0.140517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-26</th>\n",
       "      <td>0.931432</td>\n",
       "      <td>0.484913</td>\n",
       "      <td>0.418550</td>\n",
       "      <td>0.221605</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-27</th>\n",
       "      <td>0.921133</td>\n",
       "      <td>0.534907</td>\n",
       "      <td>0.456842</td>\n",
       "      <td>0.141328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-28</th>\n",
       "      <td>0.928182</td>\n",
       "      <td>0.494675</td>\n",
       "      <td>0.455456</td>\n",
       "      <td>0.322176</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-29</th>\n",
       "      <td>0.929968</td>\n",
       "      <td>0.468924</td>\n",
       "      <td>0.444251</td>\n",
       "      <td>0.364953</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  sma_diff  ema_diff       rsi\n",
       "Date                                              \n",
       "2018-06-18  0.954136  0.335379  0.240703  0.414673\n",
       "2018-06-19  0.949284  0.395494  0.275837  0.312278\n",
       "2018-06-20  0.951435  0.408611  0.293320  0.390606\n",
       "2018-06-21  0.943516  0.433393  0.326863  0.244218\n",
       "2018-06-22  0.945805  0.428458  0.340912  0.328736\n",
       "2018-06-25  0.928686  0.476945  0.395609  0.140517\n",
       "2018-06-26  0.931432  0.484913  0.418550  0.221605\n",
       "2018-06-27  0.921133  0.534907  0.456842  0.141328\n",
       "2018-06-28  0.928182  0.494675  0.455456  0.322176\n",
       "2018-06-29  0.929968  0.468924  0.444251  0.364953"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df_n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([[0.95413558, 0.33537862, 0.24070298, 0.41467311],\n",
      "       [0.94928365, 0.39549369, 0.27583745, 0.31227816],\n",
      "       [0.95143498, 0.40861133, 0.29332032, 0.39060616],\n",
      "       [0.94351627, 0.4333931 , 0.32686286, 0.24421787],\n",
      "       [0.94580492, 0.42845828, 0.34091231, 0.32873632],\n",
      "       [0.92868586, 0.47694512, 0.39560863, 0.1405167 ],\n",
      "       [0.93143223, 0.48491261, 0.41854977, 0.22160465],\n",
      "       [0.92113334, 0.53490681, 0.45684169, 0.14132769],\n",
      "       [0.92818236, 0.49467458, 0.45545568, 0.32217578],\n",
      "       [0.9299675 , 0.4689238 , 0.44425064, 0.36495317]])]\n",
      "(1, 10, 4)\n",
      "[[1.0272956]]\n",
      "[[292.54327392578125]]\n"
     ]
    }
   ],
   "source": [
    "# Create sequences\n",
    "data = test_df_n.values \n",
    "seq_len = window\n",
    "result = []\n",
    "for index in range(len(data) - seq_len + 1):\n",
    "    result.append(data[index: index + seq_len])\n",
    "\n",
    "print(result)\n",
    "result = np.array(result)\n",
    "print(result.shape)\n",
    "\n",
    "#Predict\n",
    "pred = model.predict(result)\n",
    "print(pred)\n",
    "\n",
    "# de-normalize the target\n",
    "pred = denormalize(pred, min_max_dict_list)\n",
    "\n",
    "# Send results\n",
    "pred = pred.tolist()\n",
    "print(pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# DEPLOY Model in Azure Container Instance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ticker: SPY\n",
      "Model  file: ./models/SPY-model-lstm-tensorflow-3day-0.006710271673575366.h5\n",
      "Min-max file: ./models/SPY-min_max.pkl\n",
      "Test DF file: ./data/SPY-test_score_df.pkl\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "import shutil\n",
    "from urllib.request import urlretrieve\n",
    "from keras.models import load_model\n",
    "import pandas as pd\n",
    "\n",
    "DATA_DIR = \"./data/\"\n",
    "DEPLOY_DIR = \"./deploy/\"\n",
    "MODEL_FILE = \"SPY-model-lstm-tensorflow-3day-0.006710271673575366.h5\"\n",
    "MODEL_PATH = MODELS_DIR + MODEL_FILE\n",
    "\n",
    "TICKER = MODEL_FILE.split(\"-\")[0]\n",
    "\n",
    "# the min_max values dictionary\n",
    "MIN_MAX_DICT = TICKER +'-min_max.pkl'\n",
    "MIN_MAX_DICT_PATH = MODELS_DIR + MIN_MAX_DICT\n",
    "\n",
    "# the pickle test df\n",
    "TEST_DF = TICKER +'-test_score_df.pkl'\n",
    "TEST_DF_PATH = DATA_DIR + TEST_DF\n",
    "\n",
    "# Azure Container Service (ACI) Name\n",
    "ACI_SERVICE_NAME = TICKER + '-aciservice'\n",
    "\n",
    "print(\"Ticker: \" + TICKER)\n",
    "print(\"Model  file: \" + MODEL_PATH)\n",
    "print(\"Min-max file: \" + MIN_MAX_DICT_PATH )\n",
    "print(\"Test DF file: \" + TEST_DF_PATH)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test init() and run() functions to read from the working directory"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The web service requires two functions, an init() function that will initialize the web service by loading the model into the service, and a run() function that will engineer the features to match the model call structure, and score that data set. We create the functions in here for testing and debugging."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "def init():\n",
    "    # read in the model file\n",
    "    global model\n",
    "    global min_max_dict_list\n",
    "    \n",
    "    # load model\n",
    "    model = load_model(MODEL_PATH)\n",
    "    print(\"Model Loaded\")\n",
    "    \n",
    "    # Load Min Max list values\n",
    "    with open(MIN_MAX_DICT_PATH, 'rb') as handle:\n",
    "        min_max_dict_list = pickle.load(handle)\n",
    "        print(\"Min_max List loaded\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run(raw_data):\n",
    "    try:\n",
    "        data = json.loads(raw_data)['data']\n",
    "        data = pd.read_json(data, orient='records')\n",
    "        data_n = data.copy()\n",
    "        \n",
    "        # Normalize data\n",
    "        min_dict = min_max_dict_list[0]\n",
    "        max_dict = min_max_dict_list[1]\n",
    "        for feature_name in data_n.columns:\n",
    "            data_n[feature_name] = (data[feature_name] - min_dict[feature_name]) / (max_dict[feature_name] - min_dict[feature_name])\n",
    "        \n",
    "        # Create sequences\n",
    "        data = data_n.values \n",
    "        seq_len = 10\n",
    "        result = []\n",
    "        for index in range(len(data) - seq_len + 1):\n",
    "            result.append(data[index: index + seq_len])\n",
    "\n",
    "        result = np.array(result)\n",
    "        print(result.shape)\n",
    "        \n",
    "        pred = model.predict(result)\n",
    "        print(pred)\n",
    "        \n",
    "        # de-normalize the target\n",
    "        pred = pred * (max_dict[\"Close\"] - min_dict[\"Close\"]) + min_dict[\"Close\"]\n",
    "        \n",
    "        # Send results\n",
    "        pred = pred.tolist()\n",
    "        return json.dumps({\"result\": pred})\n",
    "\n",
    "    except Exception as e:\n",
    "        result = str(e)\n",
    "        return json.dumps({\"error\": result})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The webservice test requires an initialize of the webservice, then send the entire scoring data set into the model. We expect to get 1  prediction for each input in the scoring data set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'{\"data\": \"[{\\\\\"Close\\\\\":276.56,\\\\\"sma_diff\\\\\":0.1256666667,\\\\\"ema_diff\\\\\":-0.9208042383,\\\\\"rsi\\\\\":42.9571116039},{\\\\\"Close\\\\\":275.5,\\\\\"sma_diff\\\\\":1.2423333333,\\\\\"ema_diff\\\\\":-0.5204372189,\\\\\"rsi\\\\\":33.4023302607},{\\\\\"Close\\\\\":275.97,\\\\\"sma_diff\\\\\":1.486,\\\\\"ema_diff\\\\\":-0.3212150757,\\\\\"rsi\\\\\":40.7113516968},{\\\\\"Close\\\\\":274.24,\\\\\"sma_diff\\\\\":1.9463333333,\\\\\"ema_diff\\\\\":0.0610113416,\\\\\"rsi\\\\\":27.0514200019},{\\\\\"Close\\\\\":274.74,\\\\\"sma_diff\\\\\":1.8546666667,\\\\\"ema_diff\\\\\":0.2211086852,\\\\\"rsi\\\\\":34.9380909623},{\\\\\"Close\\\\\":271.0,\\\\\"sma_diff\\\\\":2.7553333333,\\\\\"ema_diff\\\\\":0.844388337,\\\\\"rsi\\\\\":17.3747515618},{\\\\\"Close\\\\\":271.6,\\\\\"sma_diff\\\\\":2.9033333333,\\\\\"ema_diff\\\\\":1.105809008,\\\\\"rsi\\\\\":24.9413127492},{\\\\\"Close\\\\\":269.35,\\\\\"sma_diff\\\\\":3.832,\\\\\"ema_diff\\\\\":1.5421559899,\\\\\"rsi\\\\\":17.4504277746},{\\\\\"Close\\\\\":270.89,\\\\\"sma_diff\\\\\":3.0846666667,\\\\\"ema_diff\\\\\":1.5263620609,\\\\\"rsi\\\\\":34.3259071505},{\\\\\"Close\\\\\":271.28,\\\\\"sma_diff\\\\\":2.6063333333,\\\\\"ema_diff\\\\\":1.3986775304,\\\\\"rsi\\\\\":38.3175942846}]\"}'"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "json.dumps({\"data\": test_df.to_json(orient='records')})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Loaded\n",
      "Min_max List loaded\n",
      "(1, 10, 4)\n",
      "[[1.0319816]]\n",
      "{\"result\": [[293.5670166015625]]}\n"
     ]
    }
   ],
   "source": [
    "init()\n",
    "pred=run(json.dumps({\"data\": test_df.to_json(orient='records')}))\n",
    "print(pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Persist model assets"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next we persist the assets we have created for use in operationalization. The conda dependencies are defined in this YAML file. This will be used to tell the webservice server which python packages are required to run this web service"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing ./deploy/myenv.yml\n"
     ]
    }
   ],
   "source": [
    "%%writefile {DEPLOY_DIR}myenv.yml\n",
    "name: myenv\n",
    "channels:\n",
    "  - defaults\n",
    "dependencies:\n",
    "  - python=3.5.2\n",
    "  - pip:\n",
    "    - keras\n",
    "    - tensorflow\n",
    "    - h5py\n",
    "    # Required packages for AzureML execution, history, and data preparation.\n",
    "    - --extra-index-url https://azuremlsdktestpypi.azureedge.net/sdk-release/Preview/E7501C02541B433786111FE8E140CAA1\n",
    "    - azureml-core"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The score.py file is python code defining the web service operation. It includes both the init() and run() functions defined earlier imports the required libraries. These should be nearly identical to the previous defined versions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overwriting ./deploy/score.py\n"
     ]
    }
   ],
   "source": [
    "%%writefile {DEPLOY_DIR}score.py\n",
    "\n",
    "import pickle\n",
    "import json\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from azureml.core.model import Model\n",
    "from keras.models import load_model\n",
    "\n",
    "TICKER = \"SPY\"\n",
    "LSTM_MODEL = TICKER +'-modellstm.h5'\n",
    "MIN_MAX_DICT = TICKER +'-min_max.pkl'\n",
    "\n",
    "\n",
    "def init():\n",
    "    global model\n",
    "    global min_max_dict_list\n",
    "    \n",
    "    # load model\n",
    "    model_path = Model.get_model_path(model_name = LSTM_MODEL)\n",
    "    model = load_model(model_path)\n",
    "\n",
    "    # Load Min Max list values\n",
    "    model_path = Model.get_model_path(model_name = MIN_MAX_DICT)\n",
    "    with open(model_path, 'rb') as handle:\n",
    "        min_max_dict_list = pickle.load(handle)\n",
    "        print(\"Min_max List loaded\")\n",
    "\n",
    "def run(raw_data):\n",
    "    try:\n",
    "        data = json.loads(raw_data)['data']\n",
    "        data = pd.read_json(data, orient='records')\n",
    "        data_n = data.copy()\n",
    "        \n",
    "        # Normalize data\n",
    "        min_dict = min_max_dict_list[0]\n",
    "        max_dict = min_max_dict_list[1]\n",
    "        for feature_name in data.columns:\n",
    "            data_n[feature_name] = (data[feature_name] - min_dict[feature_name]) / (max_dict[feature_name] - min_dict[feature_name])\n",
    "        \n",
    "        # Create sequences\n",
    "        data = data_n.values \n",
    "        seq_len = 10\n",
    "        result = []\n",
    "        for index in range(len(data) - seq_len + 1):\n",
    "            result.append(data[index: index + seq_len])\n",
    "\n",
    "        result = np.array(result)\n",
    "        print(result.shape)\n",
    "        \n",
    "        pred = model.predict(result)\n",
    "        print(pred)\n",
    "        \n",
    "        # De-normalize the target\n",
    "        pred = pred * (max_dict[\"Close\"] - min_dict[\"Close\"]) + min_dict[\"Close\"]\n",
    "        \n",
    "        # Send results\n",
    "        pred = pred.tolist()\n",
    "        return json.dumps({\"result\": pred})\n",
    "\n",
    "    except Exception as e:\n",
    "        result = str(e)\n",
    "        return json.dumps({\"error\": result})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We also include a python file test_service.py which can test the web service you create. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Writing ./deploy/test_service.py\n"
     ]
    }
   ],
   "source": [
    "%%writefile {DEPLOY_DIR}test_service.py\n",
    "\n",
    "import urllib\n",
    "import json \n",
    "import requests\n",
    "import pandas as pd\n",
    "\n",
    "# The URL will need to be editted after service create.\n",
    "url = 'http://40.76.29.12:5001/score'\n",
    "\n",
    "## Sequence length will need to match the training sequence length from the model training\n",
    "sequence_length = 10\n",
    "\n",
    "# We'll read in this data to test the service\n",
    "body = pd.read_pickle('test_dataframe.pkl')\n",
    "headers = {'Content-Type':'application/json'}\n",
    "\n",
    "try:\n",
    "    if body.shape[0] < sequence_length : \n",
    "        print(\"Skipping scoring as we need {} records to score and only have {} records.\".format(sequence_length, body.shape[0]))\n",
    "    else:\n",
    "        #print('{}'.format(body.shape))\n",
    "        body = json.dumps({\"data\": body.to_json(orient='records')})\n",
    "        print (body + '\\n')\n",
    "        req = urllib.request.Request(url, str.encode(body), headers) \n",
    "        \n",
    "        with urllib.request.urlopen(req) as response:\n",
    "            the_page = response.read()\n",
    "            print('{}'.format(the_page))\n",
    "        \n",
    "except urllib.error.HTTPError as error:\n",
    "    print(\"The request failed with status code {}: \\n{}\".format(error, error.read))\n",
    "\n",
    "    # Print the headers - they include the requert ID and the timestamp, which are useful for debugging the failure\n",
    "    print(error.info())\n",
    "    print(error.reason)      "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Creating a web service out of the scoring script\n",
    "\n",
    "Let's now see how we can create a scoring web service from the above model. We are going to be using the Preview of the Azure ML Python SDK."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found the config file in: /home/sshuser/notebooks/Meetups-Data-AI-DFW/aml_config/config.json\n",
      "meetup_aml_workspace\n",
      "meetup_aml_rg\n",
      "eastus2\n",
      "b1395605-1fe9-4af4-b3ff-82a4725a3791\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'id': '/subscriptions/b1395605-1fe9-4af4-b3ff-82a4725a3791/resourceGroups/meetup_aml_rg/providers/Microsoft.MachineLearningServices/workspaces/meetup_aml_workspace',\n",
       " 'name': 'meetup_aml_workspace',\n",
       " 'location': 'eastus2',\n",
       " 'type': 'Microsoft.MachineLearningServices/workspaces',\n",
       " 'description': '',\n",
       " 'friendlyName': 'meetup_aml_workspace',\n",
       " 'containerRegistry': '/subscriptions/b1395605-1fe9-4af4-b3ff-82a4725a3791/resourcegroups/meetup_aml_rg/providers/microsoft.containerregistry/registries/meetupamacroifgeoka',\n",
       " 'keyVault': '/subscriptions/b1395605-1fe9-4af4-b3ff-82a4725a3791/resourcegroups/meetup_aml_rg/providers/microsoft.keyvault/vaults/meetupamkeyvaultxeeptbrq',\n",
       " 'applicationInsights': '/subscriptions/b1395605-1fe9-4af4-b3ff-82a4725a3791/resourcegroups/meetup_aml_rg/providers/microsoft.insights/components/meetupaminsightsddedjwla',\n",
       " 'identityPrincipalId': '7768c804-0fab-4f64-9d97-59a9cecf7025',\n",
       " 'identityTenantId': '72f988bf-86f1-41af-91ab-2d7cd011db47',\n",
       " 'identityType': 'SystemAssigned',\n",
       " 'storageAccount': '/subscriptions/b1395605-1fe9-4af4-b3ff-82a4725a3791/resourcegroups/meetup_aml_rg/providers/microsoft.storage/storageaccounts/meetupamstoragemjbffvgf'}"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# load workspace configuratio from ./aml_config/config.json file\n",
    "ws = Workspace.from_config()\n",
    "print(ws.name, ws.resource_group, ws.location, ws.subscription_id, sep = '\\n')\n",
    "ws.get_details()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Link (Preview whitelisted) to see the portal UI\n",
    "\n",
    "https://aka.ms/mlextensions_dev"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Register Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Registering model SPY-modellstm.h5\n"
     ]
    }
   ],
   "source": [
    "from azureml.core.model import Model\n",
    "\n",
    "LSTM_MODEL = TICKER +'-modellstm.h5' # need to shrinken the name\n",
    "\n",
    "model = Model.register(model_path = MODEL_PATH,\n",
    "                       model_name = LSTM_MODEL,\n",
    "                       tags = [TICKER, \"Close\", \"lstm\"],\n",
    "                       description = \"LSTM regression model to predict \"+ TICKER +\" Close price\",\n",
    "                       workspace = ws)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Registering model SPY-min_max.pkl\n"
     ]
    }
   ],
   "source": [
    "min_max_dict_model = Model.register(model_path = MIN_MAX_DICT_PATH,\n",
    "                       model_name = MIN_MAX_DICT,\n",
    "                       tags = [TICKER, \"MinMaxDict\"],\n",
    "                       description = \"MIN_MAX dictionary use to normalization of \"+ TICKER +\" stock data\",\n",
    "                       workspace = ws)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Name: SPY-min_max.pkl \tVersion: 1 \tDescription: MIN_MAX dictionary use to normalization of SPY stock data ['SPY', 'MinMaxDict']\n",
      "Name: SPY-modellstm.h5 \tVersion: 1 \tDescription: LSTM regression model to predict SPY Close price ['SPY', 'Close', 'lstm']\n"
     ]
    }
   ],
   "source": [
    "regression_models = ws.models(tag = TICKER)\n",
    "for m in regression_models:\n",
    "    print(\"Name:\", m.name,\"\\tVersion:\", m.version, \"\\tDescription:\", m.description, m.tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create Docker Image"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that following command can take few minutes.<br>\n",
    "Note that the score.py and the conda yml file must be in the same directory than this notebook.<br>\n",
    "You can add tags and descriptions to images. Also, an image can contain multiple models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "!cp ./deploy/score.py ./\n",
    "!cp ./deploy/myenv.yml ./"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Creating image\n",
      "Running.............................\n",
      "SucceededImage creation operation finished for image spy.image:1, operation \"Succeeded\"\n"
     ]
    }
   ],
   "source": [
    "from azureml.core.image import ContainerImage\n",
    "\n",
    "image_config = ContainerImage.image_configuration(execution_script = \"score.py\",\n",
    "                                                  runtime = \"python\",\n",
    "                                                  conda_file = \"myenv.yml\",\n",
    "                                                  description = \"Image with \"+ TICKER + \"regression LSTM model\",\n",
    "                                                  tags = [TICKER, \"Close\", \"lstm\"]\n",
    "                                                 )\n",
    "\n",
    "image = ContainerImage.create(name = TICKER.lower() + \".image\",\n",
    "                              # this is the model object\n",
    "                              models = [model, min_max_dict_model],\n",
    "                              image_config = image_config,\n",
    "                              workspace = ws)\n",
    "\n",
    "image.wait_for_creation(show_output = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "!rm score.py myenv.yml"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "spy.image:1 spy.image(v.1 [Succeeded]) stored at meetupamacroifgeoka.azurecr.io/spy.image:1 with build log https://eastus2ice.blob.core.windows.net/logs/meetupamacroifgeoka_612b8edee1504a948e560c4b7a7e6637.txt?se=2018-09-05T22%3A35%3A01Z&sp=r&sig=/WZHsHsl%2Baehnte2nZeN4K5Y0AKKIyRmX5k8WJSnqeM%3D&sr=b&sv=2017-04-17\n"
     ]
    }
   ],
   "source": [
    "for i in image.list(workspace = ws,tag = TICKER):\n",
    "    print('{} {}(v.{} [{}]) stored at {} with build log {}'.format(i.id, i.name, i.version, i.creation_state, i.image_location, i.image_build_log_uri))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Deploy image as web service on Azure Container Instance (ACI)\n",
    "Note that the service creation can take few minutes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "from azureml.core.webservice import AciWebservice\n",
    "\n",
    "aciconfig = AciWebservice.deploy_configuration(cpu_cores = 1, \n",
    "                                               memory_gb = 4, \n",
    "                                               tags = [TICKER, \"Close\", \"lstm\"], \n",
    "                                               description = \"ACI Service to predict \"+ TICKER +\" Close price\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "spy-aciservice\n",
      "Creating service\n",
      "Running...................................\n",
      "SucceededACI service creation operation finished, operation \"Succeeded\"\n",
      "Healthy\n",
      "CPU times: user 583 ms, sys: 61.4 ms, total: 644 ms\n",
      "Wall time: 3min 9s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "from azureml.core.webservice import Webservice\n",
    "\n",
    "aci_service_name = ACI_SERVICE_NAME.lower()\n",
    "print(aci_service_name)\n",
    "aci_service = Webservice.deploy_from_image(deployment_config = aciconfig,\n",
    "                                           image = image,\n",
    "                                           name = aci_service_name,\n",
    "                                           workspace = ws)\n",
    "aci_service.wait_for_deployment(True)\n",
    "print(aci_service.state)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test ACI web service"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "web service hosted in ACI: http://40.76.201.164:5001/score\n"
     ]
    }
   ],
   "source": [
    "print('web service hosted in ACI:', aci_service.scoring_uri)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{\"result\": [[293.5670166015625]]}\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "\n",
    "test_sample = json.dumps({\"data\": test_df.to_json(orient='records')})\n",
    "\n",
    "prediction = aci_service.run(input_data = test_sample)\n",
    "print(prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "b'\"{\\\\\"result\\\\\": [[293.5670166015625]]}\"'\n"
     ]
    }
   ],
   "source": [
    "import urllib\n",
    "import requests\n",
    "\n",
    "# The URL will need to be editted after service create.\n",
    "url_aci = aci_service.scoring_uri\n",
    "\n",
    "headers = {'Content-Type':'application/json'}\n",
    "\n",
    "body = json.dumps({\"data\": test_df.to_json(orient='records')})\n",
    "\n",
    "#Send Request to ACI service and print response\n",
    "req_aci = urllib.request.Request(url_aci, str.encode(body), headers) \n",
    "print(urllib.request.urlopen(req_aci).read())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "myenv",
   "language": "python",
   "name": "myenv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  },
  "widgets": {
   "state": {},
   "version": "1.1.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}