Chapter_08_RNN_cl_app.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "a16ad7d8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "------------------------------------------------------------------\n",
      "Date:  2022-02-17 08:55:23\n",
      "Balance: 8990.52 USD, \tEquity: 9027.21 USD, \tProfit: 36.69 USD\n",
      "------------------------------------------------------------------\n",
      "Warnings\n",
      "(1, 15, 5)\n",
      "WARNING:tensorflow:Detecting that an object or model or tf.train.Checkpoint is being deleted with unrestored values. See the following logs for the specific values in question. To silence these warnings, use `status.expect_partial()`. See https://www.tensorflow.org/api_docs/python/tf/train/Checkpoint#restorefor details about the status object returned by the restore function.\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.iter\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.beta_1\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.beta_2\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.decay\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.learning_rate\n",
      "WARNING:tensorflow:Detecting that an object or model or tf.train.Checkpoint is being deleted with unrestored values. See the following logs for the specific values in question. To silence these warnings, use `status.expect_partial()`. See https://www.tensorflow.org/api_docs/python/tf/train/Checkpoint#restorefor details about the status object returned by the restore function.\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.iter\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.beta_1\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.beta_2\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.decay\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.learning_rate\n",
      "WARNING:tensorflow:Detecting that an object or model or tf.train.Checkpoint is being deleted with unrestored values. See the following logs for the specific values in question. To silence these warnings, use `status.expect_partial()`. See https://www.tensorflow.org/api_docs/python/tf/train/Checkpoint#restorefor details about the status object returned by the restore function.\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.iter\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.beta_1\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.beta_2\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.decay\n",
      "WARNING:tensorflow:Value in checkpoint could not be found in the restored object: (root).optimizer.learning_rate\n",
      "------------------------------------------------------------------\n",
      "Date:  2022-02-17 08:55:28\n",
      "SYMBOL: EURUSD\n",
      "BUY: True \t  SHORT: False\n",
      "POSITION: None \t ID: None\n",
      "OPEN LONG TRADE: Request executed\n",
      "------------------------------------------------------------------\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m~\\AppData\\Local\\Temp/ipykernel_8548/2055564425.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m    180\u001b[0m                      \u001b[1;34mf\"Buy: {buy}\\t\"\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    181\u001b[0m                      f\"Sell: {sell}\")\n\u001b[1;32m--> 182\u001b[1;33m     \u001b[0mtime\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msleep\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "from MT5 import *\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "import time\n",
    "import pickle\n",
    "import ta\n",
    "from joblib import dump, load\n",
    "import os\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "import tensorflow\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, LSTM, Dropout\n",
    "\n",
    "\n",
    "path = \"\" # Ex: C:/Desktop/Python_for_finance_and_algorithmic_trading/ChapterN/Models\n",
    "\n",
    "\n",
    "\n",
    "def feature_engineering(df):\n",
    "    \"\"\" DON'T PUT THE SHIFT HERE\"\"\"\n",
    "    \n",
    "    # We copy the dataframe to avoid interferences in the data\n",
    "    df_copy = df.copy()\n",
    "    df_copy[[\"returns\"]] = df_copy[[\"close\"]].pct_change(1)\n",
    "\n",
    "    # Create the SMAs\n",
    "    df_copy[\"SMA 15\"] = df_copy[[\"close\"]].rolling(15).mean()\n",
    "    df_copy[\"SMA 60\"] = df_copy[[\"close\"]].rolling(60).mean()\n",
    "    \n",
    "    # Create the volatilities\n",
    "    df_copy[\"MSD 10\"] = df_copy[[\"returns\"]].rolling(10).std()\n",
    "    df_copy[\"MSD 30\"] = df_copy[[\"returns\"]].rolling(30).std()\n",
    "    \n",
    "    # Create the Rsi\n",
    "    RSI = ta.momentum.RSIIndicator(df_copy[\"close\"], window= 14, fillna = False)\n",
    "    df_copy[\"rsi\"] = RSI.rsi()\n",
    "    \n",
    "    # STANDARDIZATION\n",
    "\n",
    "    sc = StandardScaler()\n",
    "    df_copy = df_copy[[\"SMA 15\", \"SMA 60\", \"MSD 10\", \"MSD 30\", \"rsi\"]].dropna()\n",
    "    df_copy_sc = sc.fit_transform(df_copy)\n",
    "\n",
    "    return df_copy_sc\n",
    "\n",
    "def X_3d_RNN(X_s, y_s, lag):\n",
    "\n",
    "    # Simple verification\n",
    "    if len(X_s) != len(y_s):\n",
    "        print(\"Warnings\")\n",
    "\n",
    "    # Create the X_train\n",
    "    X_train = []\n",
    "    for variable in range(0, X_s.shape[1]):\n",
    "        X = []\n",
    "        for i in range(lag, X_s.shape[0]):\n",
    "            X.append(X_s[i-lag:i, variable])\n",
    "        X_train.append(X)\n",
    "    X_train, np.array(X_train)\n",
    "    X_train = np.swapaxes(np.swapaxes(X_train, 0, 1), 1, 2)\n",
    "\n",
    "    # Create the y_train\n",
    "    y_train = []\n",
    "    for i in range(lag, y_s.shape[0]):\n",
    "        y_train.append(y_s[i, :].reshape(-1,1).transpose())\n",
    "    y_train = np.concatenate(y_train, axis=0)\n",
    "    return X_train, y_train\n",
    "\n",
    "   \n",
    "def RNN_weights():\n",
    "    # INITIALIZATION OF THE MODEL\n",
    "    classifier = Sequential()\n",
    "\n",
    "    # ADD LSTM LAYER\n",
    "    classifier.add(LSTM(units = 10, return_sequences = True,\n",
    "                      input_shape = (15,5,)))\n",
    "\n",
    "\n",
    "    # LOOP WHICH ADD LSTM LAYER\n",
    "    for _ in range(1):\n",
    "        classifier.add(LSTM(units = 10, return_sequences = True))\n",
    "        classifier.add(Dropout(0.50))\n",
    "\n",
    "    # LAST LSTM LAYER BUT WITH return_sequences = False \n",
    "    classifier.add(LSTM(units = 10, return_sequences = False))\n",
    "\n",
    "\n",
    "    # OUTPUT DENSE LAYER\n",
    "    classifier.add(Dense(1, activation=\"sigmoid\"))\n",
    "\n",
    "    # COMPILE THE MODEL\n",
    "    classifier.compile(loss=\"binary_crossentropy\", optimizer=\"adam\")\n",
    "\n",
    "    return classifier\n",
    "\n",
    "def RNN_cl_sig(symbol):\n",
    "    \"\"\" Function for predict the value of tommorow using DNN model\"\"\"\n",
    "    \n",
    "    # Create the weights if there is not in the folder\n",
    "    rnn_1 = RNN_weights()\n",
    "    rnn_2 = RNN_weights()\n",
    "    rnn_3 = RNN_weights()\n",
    "\n",
    "    # Import trained weights\n",
    "    rnn_1.load_weights(f\"RNN_weights/RNN n3\")\n",
    "    rnn_2.load_weights(f\"RNN_weights/RNN n34\")\n",
    "    rnn_3.load_weights(f\"RNN_weights/RNN n31\")\n",
    "    \n",
    "    # Take the lastest percentage of change \n",
    "    df = MT5.get_data(symbol, 700)\n",
    "    \n",
    "    # Features engeeniring\n",
    "    data = feature_engineering(df)\n",
    "    X_data, _ = X_3d_RNN(data, np.zeros([700,1]),15)\n",
    "    X = X_data[-1:,:,:]\n",
    "    print(np.shape(X))\n",
    "    # Bagging\n",
    "    pr1 = np.where(rnn_1.predict(X)==0, -1,1)\n",
    "    pr2 = np.where(rnn_2.predict(X)==0, -1,1)\n",
    "    pr3 = np.where(rnn_3.predict(X)==0, -1,1)\n",
    "    \n",
    "    # Find the signal\n",
    "    buy = (pr1 + pr2 + pr3)[0][0] >=1\n",
    "    sell = not buy\n",
    "    \n",
    "    \n",
    "    return buy, sell\n",
    "\n",
    "\n",
    "\n",
    "# True = Live Trading and False = Screener\n",
    "live = True\n",
    "\n",
    "if live:\n",
    "    current_account_info = mt5.account_info()\n",
    "    print(\"------------------------------------------------------------------\")\n",
    "    print(\"Date: \", datetime.now().strftime(\"%Y-%m-%d %H:%M:%S\"))\n",
    "    print(f\"Balance: {current_account_info.balance} USD, \\t\"\n",
    "          f\"Equity: {current_account_info.equity} USD, \\t\"\n",
    "          f\"Profit: {current_account_info.profit} USD\")\n",
    "    print(\"------------------------------------------------------------------\")\n",
    "\n",
    "\n",
    "\n",
    "info_order = {\n",
    "    \"Bitcoin\": [\"BTCUSD\", 1.00]\n",
    "}\n",
    "\n",
    "\n",
    "start = datetime.now().strftime(\"%H:%M:%S\")\n",
    "while True:\n",
    "    # Verfication for launch\n",
    "    if datetime.now().weekday() not in (5,6):\n",
    "        is_time = datetime.now().strftime(\"%H:%M:%S\") == start #\"23:59:59\"\n",
    "    else:\n",
    "        is_time = False\n",
    "\n",
    "    \n",
    "    # Launch the algorithm\n",
    "    if is_time:\n",
    "\n",
    "        # Open the trades\n",
    "        for asset in info_order.keys():\n",
    "\n",
    "            # Initialize the inputs\n",
    "            symbol = info_order[asset][0]\n",
    "            lot = info_order[asset][1]\n",
    "\n",
    "            # Create the signals\n",
    "            buy, sell = RNN_cl_sig(symbol)\n",
    "\n",
    "             # Run the algorithm\n",
    "            if live:\n",
    "                MT5.run(symbol, buy, sell,lot)\n",
    "\n",
    "            else:\n",
    "                print(f\"Symbol: {symbol}\\t\"\n",
    "                     f\"Buy: {buy}\\t\"\n",
    "                     f\"Sell: {sell}\")\n",
    "    time.sleep(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3a4f5ea8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "07b14047",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}