Main program with sentiment analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "426fd135-65f2-45e0-8669-844f3ad94473",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Mihan Perera\\AppData\\Local\\Programs\\Python\\Python311\\Lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import json\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import yfinance as yf\n",
    "from sklearn.metrics import mean_absolute_percentage_error\n",
    "from keras import backend as K\n",
    "from keras.models import *\n",
    "from keras.layers import *\n",
    "from keras import layers\n",
    "import time\n",
    "import timeit\n",
    "from transformers import pipeline\n",
    "from transformers import AutoTokenizer, AutoModelForSequenceClassification\n",
    "from statistics import mean \n",
    "import requests\n",
    "from datetime import date, timedelta"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6886f6c7-cdc3-4836-b904-d4ad4cc01301",
   "metadata": {},
   "outputs": [],
   "source": [
    "def sentiment_score(symbol, date):\n",
    "    pipe = pipeline(\"text-classification\", model=\"ProsusAI/finbert\")\n",
    "    tokenizer = AutoTokenizer.from_pretrained(\"ProsusAI/finbert\")\n",
    "    model = AutoModelForSequenceClassification.from_pretrained(\"ProsusAI/finbert\")\n",
    "    score = []\n",
    "    stock_news_tables = {'AAPL': 'AAPL.json', 'TSLA': 'TSLA.json', 'MSFT': 'MSFT.json', 'NVDA': 'NVDA.json'}\n",
    "    filename = stock_news_tables.get(symbol)\n",
    "    \n",
    "    if filename is not None:\n",
    "        # Open and read the JSON data from the file\n",
    "        with open(filename, 'r') as file:\n",
    "            data = json.load(file)\n",
    "            score = []  # Initialize the score list as an empty list \n",
    "            titles = []\n",
    "\n",
    "            for i in range(len(data)):   \n",
    "                for j in range(len(data[i]['articles'])):\n",
    "                    #print(data[i]['articles'][j]['publishedAt'])\n",
    "                    if date == data[i]['articles'][j]['publishedAt']:\n",
    "                        titles.append(data[i]['articles'][j]['title'])\n",
    "         \n",
    "            for k in range(len(titles)): \n",
    "\n",
    "                res = pipe(titles[k])\n",
    "                if res[0]['label'] == 'positive':\n",
    "                    score.append(res[0]['score'])  # Append 'score' from 'res' to the 'score' list\n",
    "                if res[0]['label'] == 'negative':\n",
    "                    score.append(-res[0]['score'])\n",
    "                if res[0]['label'] == 'neutral':\n",
    "                    score.append(0*res[0]['score'])\n",
    "    else:\n",
    "        # Handle the case where the symbol is not found in the dictionary\n",
    "        print(f\"Symbol '{symbol}' not found in stock_news_tables\")\n",
    "    if score == []:\n",
    "        return 0\n",
    "    else:\n",
    "        return mean(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4fe52636-845c-41c2-a2e3-30cdb7fc2101",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "\n",
    "import keras\n",
    "class Time2Vec(keras.layers.Layer):\n",
    "    def __init__(self, kernel_size=1):\n",
    "        super(Time2Vec, self).__init__(trainable=True, name='Time2VecLayer')\n",
    "        self.k = kernel_size\n",
    "    \n",
    "    def build(self, input_shape):\n",
    "        # trend\n",
    "        self.wb = self.add_weight(name='wb',shape=(input_shape[1],),initializer='uniform',trainable=True)\n",
    "        self.bb = self.add_weight(name='bb',shape=(input_shape[1],),initializer='uniform',trainable=True)\n",
    "        # periodic\n",
    "        self.wa = self.add_weight(name='wa',shape=(1, input_shape[1], self.k),initializer='uniform',trainable=True)\n",
    "        self.ba = self.add_weight(name='ba',shape=(1, input_shape[1], self.k),initializer='uniform',trainable=True)\n",
    "        super(Time2Vec, self).build(input_shape)\n",
    "    \n",
    "    def call(self, inputs, **kwargs):\n",
    "        bias = self.wb * inputs + self.bb\n",
    "        dp = K.dot(inputs, self.wa) + self.ba\n",
    "        wgts = K.sin(dp) # or K.cos(.)\n",
    "        #print(wgts,bias)\n",
    "        ret = K.concatenate([K.expand_dims(bias, -1), wgts], -1)\n",
    "        ret = K.reshape(ret, (-1, inputs.shape[1]*(self.k+1)))\n",
    "        #print(ret)\n",
    "        return ret\n",
    "    \n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return (input_shape[0], input_shape[1]*(self.k + 1))\n",
    "        \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6c211af8-32f1-4ffd-9b08-9ac324c826c6",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "class ETL:\n",
    "    \"\"\"\n",
    "    ticker: str\n",
    "    period: string\n",
    "    test_size: float betwee 0 and 1\n",
    "    n_input: int\n",
    "    timestep: int\n",
    "    Extracts data for stock with ticker `ticker` from yf api,\n",
    "    splits the data into train and test sets by date,\n",
    "    reshapes the data into np.array of shape [#weeks, 5, 1],\n",
    "    converts our problem into supervised learning problem.\n",
    "    \"\"\"\n",
    "    def __init__(self, ticker=[], test_size=0.2, period='max', n_input=1, timestep=1) -> None:\n",
    "        self.ticker = ticker\n",
    "        self.period = period\n",
    "        self.test_size = test_size\n",
    "        self.n_input = n_input\n",
    "        self.df = self.extract_historic_data()\n",
    "        self.timestep = timestep\n",
    "        self.train, self.test = self.etl()\n",
    "        self.X_train, self.y_train = self.to_supervised(self.train)\n",
    "        self.X_test, self.y_test = self.to_supervised(self.test)\n",
    "        \n",
    "\n",
    "    def extract_historic_data(self) -> pd.Series:\n",
    "        \"\"\"\n",
    "        gets historical data from yf api.\n",
    "        \"\"\"\n",
    "        # Define the list of stock symbols\n",
    "        #stock_symbols = ['AAPL', 'MSFT', 'TSLA', 'NVDA']\n",
    "\n",
    "        # Fetch historical data using yfinance\n",
    "        start_date = '2023-10-18'\n",
    "        end_date = '2023-11-02'\n",
    "        delta = timedelta(days=1)\n",
    "        \n",
    "        # Create an empty DataFrame\n",
    "        df = pd.DataFrame()\n",
    "\n",
    "        # Fetch and append closing prices for each stock\n",
    "        for symbol in self.ticker:\n",
    "            stock_data = yf.download(symbol, start=start_date, end=end_date)\n",
    "            df[symbol] = stock_data['Close']\n",
    "        #print(df)  \n",
    "        return df\n",
    "\n",
    "    def split_data(self) -> tuple:\n",
    "        \"\"\"\n",
    "        Splits our pd.Series into train and test series with\n",
    "        test series representing test_size * 100 % of data.\n",
    "        \"\"\"\n",
    "        data = self.extract_historic_data()\n",
    "        \n",
    "        if len(data) != 0:\n",
    "            train_idx = round(len(data) * (1-self.test_size))\n",
    "            train = data[:train_idx]\n",
    "            test = data[train_idx:]\n",
    "            train = np.array(train)\n",
    "            test = np.array(test)\n",
    "            #print(train[:, np.newaxis])\n",
    "            return train[:, np.newaxis], test[:, np.newaxis]\n",
    "        else:\n",
    "            raise Exception('Data set is empty, cannot split.')\n",
    "\n",
    "    def window_and_reshape(self, data) -> np.array:\n",
    "        \"\"\"\n",
    "        Reformats data into shape our model needs,\n",
    "        namely, [# samples, timestep, # feautures]\n",
    "        samples\n",
    "        \"\"\"\n",
    "        NUM_FEATURES = 4\n",
    "        samples = int(data.shape[0] / self.timestep)\n",
    "        result = np.array(np.array_split(data, samples))\n",
    "       # print(result)\n",
    "        return result.reshape((samples, self.timestep, NUM_FEATURES))\n",
    "\n",
    "    def transform(self, train, test) -> np.array:\n",
    "        train_remainder = train.shape[0] % self.timestep\n",
    "        test_remainder = test.shape[0] % self.timestep\n",
    "        if train_remainder != 0 and test_remainder != 0:\n",
    "            train = train[train_remainder:]\n",
    "            test = test[test_remainder:]\n",
    "        elif train_remainder != 0:\n",
    "            train = train[train_remainder:]\n",
    "        elif test_remainder != 0:\n",
    "            test = test[test_remainder:]\n",
    "        return self.window_and_reshape(train), self.window_and_reshape(test)\n",
    "    \n",
    "    def sentiment_ana(self, train, stocks, i):\n",
    "        sentiment = [] \n",
    "        num_rows = train.shape[0]*train.shape[1]\n",
    "        # Create a column of twos\n",
    "        stock_ticker_val = np.full((num_rows, 1), i+1)\n",
    "\n",
    "        #print(train.shape)\n",
    "        #print(stock_ticker_val.shape)\n",
    "        start_date = date(2023, 10, 18)\n",
    "        end_date = date(2023, 11, 2)\n",
    "        delta = timedelta(days=1)\n",
    "        past_data = self.df\n",
    "       # print(past_data)\n",
    "        while start_date <= end_date:\n",
    "            #print(start_date.strftime(\"%Y-%m-%d\"))\n",
    "            dates = start_date.strftime(\"%Y-%m-%d\")\n",
    "            if dates in past_data.index:                     #figure out a way to get the date\n",
    "                res = sentiment_score(stocks[i], dates)\n",
    "               # print(res)\n",
    "                sentiment.append(res)\n",
    "            start_date += delta\n",
    "\n",
    "        senti_array = []\n",
    "        if (np.array(sentiment).shape[0]/2) > train.shape[0]:\n",
    "            senti_array = sentiment[-train.shape[0]:]\n",
    "            #print(train.shape[0])\n",
    "        else:\n",
    "            senti_array = sentiment[:train.shape[0]]\n",
    "            #print(np.array(senti_array).shape[0])\n",
    "\n",
    "        new_sent = np.array(senti_array).reshape((num_rows,1))\n",
    "        # Concatenate the original array and the twos column horizontally\n",
    "        new_array = train[:, :, i].reshape((train.shape[0]*train.shape[1],1))\n",
    "        reshaped_array = np.hstack((new_array, new_sent, stock_ticker_val)) #add the sentiment analysis here\n",
    "        return reshaped_array\n",
    "    \n",
    "    def etl(self) -> tuple[np.array, np.array]:\n",
    "        \"\"\"\n",
    "        Runs complete ETL\n",
    "        \"\"\"\n",
    "        train, test = self.split_data()\n",
    "        return self.transform(train, test)\n",
    "\n",
    "\n",
    "    def to_supervised(self, train, n_out=1) -> tuple:\n",
    "        \"\"\"\n",
    "        Converts our time series prediction problem to a\n",
    "        supervised learning problem.\n",
    "        \"\"\"\n",
    "        # flatted the data\n",
    "        #print(train.shape[0], train.shape[1], train.shape[2])\n",
    "        stocks = self.ticker\n",
    "        num_stocks = len(stocks)\n",
    "        data_x = []\n",
    "        data_y = [] \n",
    "        data_new = []\n",
    "        \n",
    "        for i in range(len(self.ticker)):\n",
    "                                          \n",
    "            #reshaped_array = new_array.reshape((new_array.shape[0]*train.shape[1], train.shape[2]))\n",
    "            '''\n",
    "            sentiment = [] \n",
    "            num_rows = train.shape[0]*train.shape[1]\n",
    "            # Create a column of twos\n",
    "            stock_ticker_val = np.full((num_rows, 1), i+1)\n",
    "\n",
    "            #print(train.shape)\n",
    "            #print(stock_ticker_val.shape)\n",
    "            start_date = date(2023, 10, 18)\n",
    "            end_date = date(2023, 11, 2)\n",
    "            delta = timedelta(days=1)\n",
    "\n",
    "            while start_date <= end_date:\n",
    "                #print(start_date.strftime(\"%Y-%m-%d\"))\n",
    "                dates = start_date.strftime(\"%Y-%m-%d\")\n",
    "                if dates in self.df.index:                     #figure out a way to get the date\n",
    "                    res = sentiment_score(stocks[i], dates)\n",
    "                    print(res)\n",
    "                    sentiment.append(res)\n",
    "                start_date += delta\n",
    "\n",
    "            senti_array = []\n",
    "            if (np.array(sentiment).shape[0]/2) > train.shape[0]:\n",
    "                senti_array = sentiment[-train.shape[0]:]\n",
    "                #print(train.shape[0])\n",
    "            else:\n",
    "                senti_array = sentiment[:train.shape[0]]\n",
    "                #print(np.array(senti_array).shape[0])\n",
    "\n",
    "            new_sent = np.array(senti_array).reshape((num_rows,1))\n",
    "            # Concatenate the original array and the twos column horizontally\n",
    "            new_array = train[:,:, i].reshape((train.shape[0]*train.shape[1],1))\n",
    "            reshaped_array = np.hstack((new_array, new_sent, stock_ticker_val)) #add the sentiment analysis here\n",
    "            '''\n",
    "            \n",
    "            reshaped_array = self.sentiment_ana(train, stocks, i)\n",
    "            X, y = [], []\n",
    "            in_start = 0\n",
    "            \n",
    "            # step over the entire history one time step at a time\n",
    "            for _ in range(len(reshaped_array)):\n",
    "                # define the end of the input sequence\n",
    "                in_end = in_start + self.n_input\n",
    "                out_end = in_end + n_out\n",
    "                # ensure we have enough data for this instance\n",
    "                if out_end <= len(reshaped_array):\n",
    "                    x_input = reshaped_array[in_start:in_end, :]\n",
    "                    x_input = x_input.reshape((len(x_input), 3))\n",
    "                    #print(x_input)\n",
    "                    X.append(x_input)\n",
    "                    y.append(reshaped_array[in_end:out_end, 0])\n",
    "                    # move along one time step\n",
    "                    in_start += 1\n",
    "                    #print(X)\n",
    "            data_x.append(X)\n",
    "            data_y.append(y)\n",
    "        data_x = np.array(data_x)\n",
    "        #print(data_x.shape)\n",
    "        data_y = np.array(data_y)\n",
    "       # print(data_y.shape)\n",
    "        return data_x, data_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "46dc076a-6380-43f9-8232-ba6a92122433",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "class PredictAndForecast():\n",
    "    \"\"\"\n",
    "    model: tf.keras.Model\n",
    "    train: np.array\n",
    "    test: np.array\n",
    "    Takes a trained model, train, and test datasets and returns predictions\n",
    "    of len(test) with same shape.\n",
    "    \"\"\"\n",
    "    def __init__(self, model, train, test, n_input=1) -> None:\n",
    "        self.model = model\n",
    "        self.train = train\n",
    "        self.test = test\n",
    "        self.n_input = n_input\n",
    "        self.predictions = self.get_predictions()\n",
    "\n",
    "\n",
    "    def forecast(self, history) -> np.array:\n",
    "        \n",
    "        \"\"\"\n",
    "        Given last weeks actual data, forecasts next weeks prices.\n",
    "        \"\"\"\n",
    "        # flatten data\n",
    "        data = np.array(history)\n",
    "                            \n",
    "        data = data.reshape((data.shape[0]*data.shape[1], data.shape[2]))\n",
    "        # retrieve last observations for input data\n",
    "        input_x = data[-self.n_input:, :]\n",
    "      \n",
    "        # reshape into [1, n_input, 1]\n",
    "        input_x = input_x.reshape((1, len(input_x), input_x.shape[1]))\n",
    "       # print(input_x)\n",
    "        # forecast the next week\n",
    "        yhat = self.model.predict(input_x, verbose=0)\n",
    "        # we only want the vector forecast\n",
    "        yhat = yhat[0]\n",
    "        return yhat\n",
    "        \n",
    "    def get_predictions(self) -> np.array:\n",
    "        \"\"\"\n",
    "        compiles models predictions week by week over entire\n",
    "        test set.\n",
    "        \"\"\"\n",
    "        # history is a list of weekly data\n",
    "        history = [x for x in self.train]\n",
    "        # walk-forward validation over each week\n",
    "        predictions = []\n",
    "        for i in range(len(self.test)):\n",
    "            yhat_sequence = self.forecast(history)\n",
    "            # store the predictions\n",
    "            predictions.append(yhat_sequence)\n",
    "        # get real observation and add to history for predicting the next week\n",
    "            history.append(self.test[i, :])\n",
    "        #print(np.array(predictions))\n",
    "        return np.array(predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5265bec5-2ad1-40b9-823c-4a24954cb428",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "class Evaluate:\n",
    "\n",
    "  def __init__(self, actual, predictions) -> None:\n",
    "    \n",
    "    #print(predictions.shape)\n",
    "    self.actual = actual[:,:,0]\n",
    "    self.predictions = predictions\n",
    "    self.var_ratio = self.compare_var()\n",
    "    self.mape = self.evaluate_model_with_mape()\n",
    "\n",
    "  def compare_var(self):\n",
    "    return abs( 1 - (np.var(self.predictions) / np.var(self.actual)))\n",
    "\n",
    "  def evaluate_model_with_mape(self):\n",
    "    return mean_absolute_percentage_error(self.actual.flatten(), self.predictions.flatten())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "67ba09b7-428f-4af6-a919-abc518ec6ff6",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[*********************100%%**********************]  1 of 1 completed\n",
      "[*********************100%%**********************]  1 of 1 completed\n",
      "[*********************100%%**********************]  1 of 1 completed\n",
      "[*********************100%%**********************]  1 of 1 completed\n",
      "[*********************100%%**********************]  1 of 1 completed\n",
      "[*********************100%%**********************]  1 of 1 completed\n",
      "[*********************100%%**********************]  1 of 1 completed\n",
      "[*********************100%%**********************]  1 of 1 completed\n"
     ]
    }
   ],
   "source": [
    "data = ETL(['AAPL', 'MSFT', 'TSLA', 'NVDA'])\n",
    "#data.df.reset_index(inplace=True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "6ab60e1b-7f67-411f-9fad-6e139dfe8b23",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def gated_residual_network(inputs, hidden_dim):\n",
    "    x = layers.LayerNormalization()(inputs)\n",
    "    \n",
    "    g1 = layers.Dense(hidden_dim, activation=\"linear\")(x)\n",
    "    g2 = layers.Dense(hidden_dim, activation=\"elu\")(x)\n",
    "    \n",
    "    g = layers.multiply([g1, g2])\n",
    "    g = layers.Dense(x.shape[-1])(g)\n",
    "    grn_output = layers.add([x, g])\n",
    "    \n",
    "    return grn_output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "f0f3be0d-ea38-427f-93b8-5a6de05a355b",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def transformer_encoder(inputs, head_size, num_heads, ff_dim, dropout=0, epsilon=1e-6, attention_axes=None, kernel_size=1):\n",
    "  \"\"\"\n",
    "  Creates a single transformer block.\n",
    "  \"\"\"\n",
    "  x = layers.LayerNormalization(epsilon=epsilon)(inputs)\n",
    "  x = layers.MultiHeadAttention(\n",
    "      key_dim=head_size, num_heads=num_heads, dropout=dropout,\n",
    "      attention_axes=attention_axes\n",
    "      )(x, x)\n",
    "  x = layers.Dropout(dropout)(x)\n",
    "  res = x + inputs\n",
    "\n",
    "  grn_output = gated_residual_network(res, ff_dim)\n",
    "#1 Gated Feed Forward Part\n",
    " # x = layers.LayerNormalization(epsilon=epsilon)(res)\n",
    " # x = gated_residual_block(x, 3)\n",
    " # x = layers.LayerNormalization(epsilon=epsilon)(x)\n",
    " # ff_output = layers.Conv1D(filters=ff_dim, kernel_size=kernel_size, activation=\"relu\")(x)\n",
    "  #ff_output = layers.Conv1D(filters=inputs.shape[-1], kernel_size=kernel_size)(ff_output)\n",
    "    \n",
    "# Gated Residual Connection\n",
    " # gated_output = keras.layers.multiply([ff_output, res])\n",
    " \n",
    "  #3 Gated Residual Connection\n",
    "  #print(res.shape[-2])\n",
    "  #x = GatedResidualNetwork(input_dim=res.shape[-2], hidden_dim=256, output_dim=res.shape[-2])(res)\n",
    "    \n",
    " # Feed Forward Part\n",
    "  x = layers.LayerNormalization(epsilon=epsilon)(grn_output)\n",
    "  x = layers.Conv1D(filters=ff_dim, kernel_size=kernel_size, activation=\"relu\")(x)\n",
    "  x = layers.Dropout(dropout)(x)\n",
    "  x = layers.Conv1D(filters=inputs.shape[-1], kernel_size=kernel_size)(x)\n",
    "  return x + res\n",
    "  \n",
    "\n",
    "def build_transfromer(head_size, num_heads, ff_dim, num_trans_blocks, mlp_units, dropout=0, mlp_dropout=0, attention_axes=None, epsilon=1e-6, kernel_size=1):\n",
    "  \"\"\"\n",
    "  Creates final model by building many transformer blocks.\n",
    "  \"\"\"\n",
    "\n",
    "  n_timesteps, n_features, n_outputs = 1, 3, 1\n",
    "  time2vec = Time2Vec(kernel_size=1)\n",
    "\n",
    "  inputs = tf.keras.Input(shape=(n_timesteps, n_features))\n",
    "  time_embedding = keras.layers.TimeDistributed(time2vec)(inputs)\n",
    "  x = K.concatenate([inputs, time_embedding], -1)\n",
    "  \n",
    "  for _ in range(num_trans_blocks):\n",
    "    x = transformer_encoder(x, head_size=head_size, num_heads=num_heads, ff_dim=ff_dim, dropout=dropout, attention_axes=attention_axes, kernel_size=kernel_size, epsilon=epsilon)\n",
    "\n",
    "  x = layers.GlobalAveragePooling1D(data_format=\"channels_first\")(x)\n",
    "  for dim in mlp_units:\n",
    "    x = layers.Dense(dim, activation=\"relu\")(x)\n",
    "    x = layers.Dropout(mlp_dropout)(x)\n",
    "    \n",
    "  #portfolio_weights = CustomOutputLayer()(x)\n",
    "  #return tf.keras.Model(inputs, portfolio_weights)\n",
    "  outputs = layers.Dense(n_outputs)(x)\n",
    "  return tf.keras.Model(inputs, outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "9508f15e-a08a-4d6e-8995-360eea75ec7b",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "transformer = build_transfromer(head_size=256, num_heads=12, ff_dim=128, num_trans_blocks=4, mlp_units=[512], mlp_dropout=0.10, dropout=0.10, attention_axes=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "976d9158-d053-4821-bc1c-b7b9f435a42b",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " input_2 (InputLayer)           [(None, 1, 3)]       0           []                               \n",
      "                                                                                                  \n",
      " time_distributed_1 (TimeDistri  (None, 1, 6)        12          ['input_2[0][0]']                \n",
      " buted)                                                                                           \n",
      "                                                                                                  \n",
      " tf.concat_1 (TFOpLambda)       (None, 1, 9)         0           ['input_2[0][0]',                \n",
      "                                                                  'time_distributed_1[0][0]']     \n",
      "                                                                                                  \n",
      " layer_normalization_12 (LayerN  (None, 1, 9)        18          ['tf.concat_1[0][0]']            \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " multi_head_attention_4 (MultiH  (None, 1, 9)        119817      ['layer_normalization_12[0][0]', \n",
      " eadAttention)                                                    'layer_normalization_12[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_9 (Dropout)            (None, 1, 9)         0           ['multi_head_attention_4[0][0]'] \n",
      "                                                                                                  \n",
      " tf.__operators__.add_8 (TFOpLa  (None, 1, 9)        0           ['dropout_9[0][0]',              \n",
      " mbda)                                                            'tf.concat_1[0][0]']            \n",
      "                                                                                                  \n",
      " layer_normalization_13 (LayerN  (None, 1, 9)        18          ['tf.__operators__.add_8[0][0]'] \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " dense_14 (Dense)               (None, 1, 128)       1280        ['layer_normalization_13[0][0]'] \n",
      "                                                                                                  \n",
      " dense_15 (Dense)               (None, 1, 128)       1280        ['layer_normalization_13[0][0]'] \n",
      "                                                                                                  \n",
      " multiply_4 (Multiply)          (None, 1, 128)       0           ['dense_14[0][0]',               \n",
      "                                                                  'dense_15[0][0]']               \n",
      "                                                                                                  \n",
      " dense_16 (Dense)               (None, 1, 9)         1161        ['multiply_4[0][0]']             \n",
      "                                                                                                  \n",
      " add_4 (Add)                    (None, 1, 9)         0           ['layer_normalization_13[0][0]', \n",
      "                                                                  'dense_16[0][0]']               \n",
      "                                                                                                  \n",
      " layer_normalization_14 (LayerN  (None, 1, 9)        18          ['add_4[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " conv1d_8 (Conv1D)              (None, 1, 128)       1280        ['layer_normalization_14[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_10 (Dropout)           (None, 1, 128)       0           ['conv1d_8[0][0]']               \n",
      "                                                                                                  \n",
      " conv1d_9 (Conv1D)              (None, 1, 9)         1161        ['dropout_10[0][0]']             \n",
      "                                                                                                  \n",
      " tf.__operators__.add_9 (TFOpLa  (None, 1, 9)        0           ['conv1d_9[0][0]',               \n",
      " mbda)                                                            'tf.__operators__.add_8[0][0]'] \n",
      "                                                                                                  \n",
      " layer_normalization_15 (LayerN  (None, 1, 9)        18          ['tf.__operators__.add_9[0][0]'] \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " multi_head_attention_5 (MultiH  (None, 1, 9)        119817      ['layer_normalization_15[0][0]', \n",
      " eadAttention)                                                    'layer_normalization_15[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_11 (Dropout)           (None, 1, 9)         0           ['multi_head_attention_5[0][0]'] \n",
      "                                                                                                  \n",
      " tf.__operators__.add_10 (TFOpL  (None, 1, 9)        0           ['dropout_11[0][0]',             \n",
      " ambda)                                                           'tf.__operators__.add_9[0][0]'] \n",
      "                                                                                                  \n",
      " layer_normalization_16 (LayerN  (None, 1, 9)        18          ['tf.__operators__.add_10[0][0]']\n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " dense_17 (Dense)               (None, 1, 128)       1280        ['layer_normalization_16[0][0]'] \n",
      "                                                                                                  \n",
      " dense_18 (Dense)               (None, 1, 128)       1280        ['layer_normalization_16[0][0]'] \n",
      "                                                                                                  \n",
      " multiply_5 (Multiply)          (None, 1, 128)       0           ['dense_17[0][0]',               \n",
      "                                                                  'dense_18[0][0]']               \n",
      "                                                                                                  \n",
      " dense_19 (Dense)               (None, 1, 9)         1161        ['multiply_5[0][0]']             \n",
      "                                                                                                  \n",
      " add_5 (Add)                    (None, 1, 9)         0           ['layer_normalization_16[0][0]', \n",
      "                                                                  'dense_19[0][0]']               \n",
      "                                                                                                  \n",
      " layer_normalization_17 (LayerN  (None, 1, 9)        18          ['add_5[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " conv1d_10 (Conv1D)             (None, 1, 128)       1280        ['layer_normalization_17[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_12 (Dropout)           (None, 1, 128)       0           ['conv1d_10[0][0]']              \n",
      "                                                                                                  \n",
      " conv1d_11 (Conv1D)             (None, 1, 9)         1161        ['dropout_12[0][0]']             \n",
      "                                                                                                  \n",
      " tf.__operators__.add_11 (TFOpL  (None, 1, 9)        0           ['conv1d_11[0][0]',              \n",
      " ambda)                                                           'tf.__operators__.add_10[0][0]']\n",
      "                                                                                                  \n",
      " layer_normalization_18 (LayerN  (None, 1, 9)        18          ['tf.__operators__.add_11[0][0]']\n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " multi_head_attention_6 (MultiH  (None, 1, 9)        119817      ['layer_normalization_18[0][0]', \n",
      " eadAttention)                                                    'layer_normalization_18[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_13 (Dropout)           (None, 1, 9)         0           ['multi_head_attention_6[0][0]'] \n",
      "                                                                                                  \n",
      " tf.__operators__.add_12 (TFOpL  (None, 1, 9)        0           ['dropout_13[0][0]',             \n",
      " ambda)                                                           'tf.__operators__.add_11[0][0]']\n",
      "                                                                                                  \n",
      " layer_normalization_19 (LayerN  (None, 1, 9)        18          ['tf.__operators__.add_12[0][0]']\n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " dense_20 (Dense)               (None, 1, 128)       1280        ['layer_normalization_19[0][0]'] \n",
      "                                                                                                  \n",
      " dense_21 (Dense)               (None, 1, 128)       1280        ['layer_normalization_19[0][0]'] \n",
      "                                                                                                  \n",
      " multiply_6 (Multiply)          (None, 1, 128)       0           ['dense_20[0][0]',               \n",
      "                                                                  'dense_21[0][0]']               \n",
      "                                                                                                  \n",
      " dense_22 (Dense)               (None, 1, 9)         1161        ['multiply_6[0][0]']             \n",
      "                                                                                                  \n",
      " add_6 (Add)                    (None, 1, 9)         0           ['layer_normalization_19[0][0]', \n",
      "                                                                  'dense_22[0][0]']               \n",
      "                                                                                                  \n",
      " layer_normalization_20 (LayerN  (None, 1, 9)        18          ['add_6[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " conv1d_12 (Conv1D)             (None, 1, 128)       1280        ['layer_normalization_20[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_14 (Dropout)           (None, 1, 128)       0           ['conv1d_12[0][0]']              \n",
      "                                                                                                  \n",
      " conv1d_13 (Conv1D)             (None, 1, 9)         1161        ['dropout_14[0][0]']             \n",
      "                                                                                                  \n",
      " tf.__operators__.add_13 (TFOpL  (None, 1, 9)        0           ['conv1d_13[0][0]',              \n",
      " ambda)                                                           'tf.__operators__.add_12[0][0]']\n",
      "                                                                                                  \n",
      " layer_normalization_21 (LayerN  (None, 1, 9)        18          ['tf.__operators__.add_13[0][0]']\n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " multi_head_attention_7 (MultiH  (None, 1, 9)        119817      ['layer_normalization_21[0][0]', \n",
      " eadAttention)                                                    'layer_normalization_21[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_15 (Dropout)           (None, 1, 9)         0           ['multi_head_attention_7[0][0]'] \n",
      "                                                                                                  \n",
      " tf.__operators__.add_14 (TFOpL  (None, 1, 9)        0           ['dropout_15[0][0]',             \n",
      " ambda)                                                           'tf.__operators__.add_13[0][0]']\n",
      "                                                                                                  \n",
      " layer_normalization_22 (LayerN  (None, 1, 9)        18          ['tf.__operators__.add_14[0][0]']\n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " dense_23 (Dense)               (None, 1, 128)       1280        ['layer_normalization_22[0][0]'] \n",
      "                                                                                                  \n",
      " dense_24 (Dense)               (None, 1, 128)       1280        ['layer_normalization_22[0][0]'] \n",
      "                                                                                                  \n",
      " multiply_7 (Multiply)          (None, 1, 128)       0           ['dense_23[0][0]',               \n",
      "                                                                  'dense_24[0][0]']               \n",
      "                                                                                                  \n",
      " dense_25 (Dense)               (None, 1, 9)         1161        ['multiply_7[0][0]']             \n",
      "                                                                                                  \n",
      " add_7 (Add)                    (None, 1, 9)         0           ['layer_normalization_22[0][0]', \n",
      "                                                                  'dense_25[0][0]']               \n",
      "                                                                                                  \n",
      " layer_normalization_23 (LayerN  (None, 1, 9)        18          ['add_7[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " conv1d_14 (Conv1D)             (None, 1, 128)       1280        ['layer_normalization_23[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_16 (Dropout)           (None, 1, 128)       0           ['conv1d_14[0][0]']              \n",
      "                                                                                                  \n",
      " conv1d_15 (Conv1D)             (None, 1, 9)         1161        ['dropout_16[0][0]']             \n",
      "                                                                                                  \n",
      " tf.__operators__.add_15 (TFOpL  (None, 1, 9)        0           ['conv1d_15[0][0]',              \n",
      " ambda)                                                           'tf.__operators__.add_14[0][0]']\n",
      "                                                                                                  \n",
      " global_average_pooling1d_1 (Gl  (None, 1)           0           ['tf.__operators__.add_15[0][0]']\n",
      " obalAveragePooling1D)                                                                            \n",
      "                                                                                                  \n",
      " dense_26 (Dense)               (None, 512)          1024        ['global_average_pooling1d_1[0][0\n",
      "                                                                 ]']                              \n",
      "                                                                                                  \n",
      " dropout_17 (Dropout)           (None, 512)          0           ['dense_26[0][0]']               \n",
      "                                                                                                  \n",
      " dense_27 (Dense)               (None, 1)            513         ['dropout_17[0][0]']             \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 505,681\n",
      "Trainable params: 505,681\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "transformer.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "0b30ee7a-8370-408d-9f43-b0004d2b9026",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def fit_transformer(transformer: tf.keras.Model):\n",
    "  \"\"\"\n",
    "  Compiles and fits our transformer.\n",
    "  \"\"\"\n",
    "  transformer.compile(\n",
    "    loss=\"mse\",\n",
    "    optimizer=tf.keras.optimizers.Adam(learning_rate=1e-3),\n",
    "    metrics=[\"mae\", 'mape'])\n",
    "  \n",
    "  print(data.X_train.shape[0])\n",
    "  hist = 0 \n",
    "  for stocks in range(data.X_train.shape[0]):\n",
    "      callbacks = [tf.keras.callbacks.EarlyStopping(monitor='loss', patience=10, restore_best_weights=True)]\n",
    "      start = time.time()\n",
    "      hist = transformer.fit(data.X_train[stocks], data.y_train[stocks], batch_size=32, epochs=25, verbose=1, callbacks=callbacks)\n",
    "      print(time.time() - start)\n",
    "  return hist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "4d2be5df-c7bf-4dfc-aac8-53a58a706afb",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4\n",
      "Epoch 1/25\n",
      "1/1 [==============================] - 10s 10s/step - loss: 29957.8711 - mae: 173.0642 - mape: 100.9662\n",
      "Epoch 2/25\n",
      "1/1 [==============================] - 0s 58ms/step - loss: 29815.5586 - mae: 172.6536 - mape: 100.7275\n",
      "Epoch 3/25\n",
      "1/1 [==============================] - 0s 57ms/step - loss: 29613.5820 - mae: 172.0639 - mape: 100.3808\n",
      "Epoch 4/25\n",
      "1/1 [==============================] - 0s 57ms/step - loss: 29469.8633 - mae: 171.6473 - mape: 100.1389\n",
      "Epoch 5/25\n",
      "1/1 [==============================] - 0s 59ms/step - loss: 29314.8828 - mae: 171.1945 - mape: 99.8741\n",
      "Epoch 6/25\n",
      "1/1 [==============================] - 0s 53ms/step - loss: 29265.9609 - mae: 171.0508 - mape: 99.7896\n",
      "Epoch 7/25\n",
      "1/1 [==============================] - 0s 51ms/step - loss: 29193.3906 - mae: 170.8404 - mape: 99.6680\n",
      "Epoch 8/25\n",
      "1/1 [==============================] - 0s 57ms/step - loss: 29120.3164 - mae: 170.6270 - mape: 99.5438\n",
      "Epoch 9/25\n",
      "1/1 [==============================] - 0s 62ms/step - loss: 28999.4863 - mae: 170.2715 - mape: 99.3358\n",
      "Epoch 10/25\n",
      "1/1 [==============================] - 0s 60ms/step - loss: 28846.6094 - mae: 169.8242 - mape: 99.0761\n",
      "Epoch 11/25\n",
      "1/1 [==============================] - 0s 58ms/step - loss: 28701.5957 - mae: 169.3956 - mape: 98.8255\n",
      "Epoch 12/25\n",
      "1/1 [==============================] - 0s 61ms/step - loss: 28586.3594 - mae: 169.0550 - mape: 98.6265\n",
      "Epoch 13/25\n",
      "1/1 [==============================] - 0s 59ms/step - loss: 28343.7383 - mae: 168.3347 - mape: 98.2057\n",
      "Epoch 14/25\n",
      "1/1 [==============================] - 0s 54ms/step - loss: 28190.3867 - mae: 167.8816 - mape: 97.9430\n",
      "Epoch 15/25\n",
      "1/1 [==============================] - 0s 59ms/step - loss: 27900.3789 - mae: 167.0127 - mape: 97.4345\n",
      "Epoch 16/25\n",
      "1/1 [==============================] - 0s 60ms/step - loss: 27742.4863 - mae: 166.5446 - mape: 97.1642\n",
      "Epoch 17/25\n",
      "1/1 [==============================] - 0s 61ms/step - loss: 27387.2949 - mae: 165.4713 - mape: 96.5359\n",
      "Epoch 18/25\n",
      "1/1 [==============================] - 0s 59ms/step - loss: 27048.2305 - mae: 164.4419 - mape: 95.9345\n",
      "Epoch 19/25\n",
      "1/1 [==============================] - 0s 64ms/step - loss: 26685.4805 - mae: 163.3403 - mape: 95.2950\n",
      "Epoch 20/25\n",
      "1/1 [==============================] - 0s 58ms/step - loss: 26282.4180 - mae: 162.0973 - mape: 94.5662\n",
      "Epoch 21/25\n",
      "1/1 [==============================] - 0s 52ms/step - loss: 26032.7422 - mae: 161.3218 - mape: 94.1133\n",
      "Epoch 22/25\n",
      "1/1 [==============================] - 0s 60ms/step - loss: 25423.2148 - mae: 159.4226 - mape: 93.0046\n",
      "Epoch 23/25\n",
      "1/1 [==============================] - 0s 54ms/step - loss: 25001.8457 - mae: 158.0937 - mape: 92.2281\n",
      "Epoch 24/25\n",
      "1/1 [==============================] - 0s 66ms/step - loss: 24493.8242 - mae: 156.4787 - mape: 91.2859\n",
      "Epoch 25/25\n",
      "1/1 [==============================] - 0s 67ms/step - loss: 23854.0664 - mae: 154.4227 - mape: 90.0872\n",
      "11.579726934432983\n",
      "Epoch 1/25\n",
      "1/1 [==============================] - 0s 91ms/step - loss: 91774.9688 - mae: 302.9121 - mape: 91.3230\n",
      "Epoch 2/25\n",
      "1/1 [==============================] - 0s 68ms/step - loss: 90309.1328 - mae: 300.4708 - mape: 90.5844\n",
      "Epoch 3/25\n",
      "1/1 [==============================] - 0s 66ms/step - loss: 88082.0859 - mae: 296.7562 - mape: 89.4674\n",
      "Epoch 4/25\n",
      "1/1 [==============================] - 0s 64ms/step - loss: 86814.7969 - mae: 294.6054 - mape: 88.8188\n",
      "Epoch 5/25\n",
      "1/1 [==============================] - 0s 67ms/step - loss: 83727.5469 - mae: 289.3002 - mape: 87.2148\n",
      "Epoch 6/25\n",
      "1/1 [==============================] - 0s 67ms/step - loss: 80856.5312 - mae: 284.2940 - mape: 85.7039\n",
      "Epoch 7/25\n",
      "1/1 [==============================] - 0s 65ms/step - loss: 79332.3438 - mae: 281.6123 - mape: 84.8978\n",
      "Epoch 8/25\n",
      "1/1 [==============================] - 0s 64ms/step - loss: 75363.7422 - mae: 274.4542 - mape: 82.7374\n",
      "Epoch 9/25\n",
      "1/1 [==============================] - 0s 70ms/step - loss: 73236.7031 - mae: 270.5617 - mape: 81.5650\n",
      "Epoch 10/25\n",
      "1/1 [==============================] - 0s 65ms/step - loss: 68599.1562 - mae: 261.8522 - mape: 78.9432\n",
      "Epoch 11/25\n",
      "1/1 [==============================] - 0s 73ms/step - loss: 64680.2617 - mae: 254.2913 - mape: 76.6670\n",
      "Epoch 12/25\n",
      "1/1 [==============================] - 0s 73ms/step - loss: 60789.0156 - mae: 246.5085 - mape: 74.3206\n",
      "Epoch 13/25\n",
      "1/1 [==============================] - 0s 62ms/step - loss: 56732.3047 - mae: 238.1011 - mape: 71.7765\n",
      "Epoch 14/25\n",
      "1/1 [==============================] - 0s 68ms/step - loss: 52904.1211 - mae: 229.7410 - mape: 69.2342\n",
      "Epoch 15/25\n",
      "1/1 [==============================] - 0s 65ms/step - loss: 46346.9922 - mae: 215.1651 - mape: 64.8640\n",
      "Epoch 16/25\n",
      "1/1 [==============================] - 0s 104ms/step - loss: 42543.7031 - mae: 206.0468 - mape: 62.1210\n",
      "Epoch 17/25\n",
      "1/1 [==============================] - 0s 67ms/step - loss: 35426.1641 - mae: 188.0383 - mape: 56.6931\n",
      "Epoch 18/25\n",
      "1/1 [==============================] - 0s 62ms/step - loss: 31810.0312 - mae: 177.9682 - mape: 53.6289\n",
      "Epoch 19/25\n",
      "1/1 [==============================] - 0s 73ms/step - loss: 25556.8906 - mae: 158.9582 - mape: 47.9007\n",
      "Epoch 20/25\n",
      "1/1 [==============================] - 0s 63ms/step - loss: 20262.1602 - mae: 142.0048 - mape: 42.8165\n",
      "Epoch 21/25\n",
      "1/1 [==============================] - 0s 66ms/step - loss: 15691.3330 - mae: 124.4842 - mape: 37.5024\n",
      "Epoch 22/25\n",
      "1/1 [==============================] - 0s 70ms/step - loss: 10669.6748 - mae: 102.1917 - mape: 30.7888\n",
      "Epoch 23/25\n",
      "1/1 [==============================] - 0s 69ms/step - loss: 8922.2129 - mae: 93.7450 - mape: 28.2383\n",
      "Epoch 24/25\n",
      "1/1 [==============================] - 0s 74ms/step - loss: 3484.3569 - mae: 55.5050 - mape: 16.7121\n",
      "Epoch 25/25\n",
      "1/1 [==============================] - 0s 63ms/step - loss: 2557.6582 - mae: 49.2370 - mape: 14.8589\n",
      "1.9773869514465332\n",
      "Epoch 1/25\n",
      "1/1 [==============================] - 0s 98ms/step - loss: 5195.6919 - mae: 69.7931 - mape: 33.2650\n",
      "Epoch 2/25\n",
      "1/1 [==============================] - 0s 36ms/step - loss: 8770.5977 - mae: 92.6353 - mape: 44.0807\n",
      "Epoch 3/25\n",
      "1/1 [==============================] - 0s 41ms/step - loss: 16453.7910 - mae: 124.8878 - mape: 59.4762\n",
      "Epoch 4/25\n",
      "1/1 [==============================] - 0s 36ms/step - loss: 12448.4043 - mae: 110.4006 - mape: 52.4791\n",
      "Epoch 5/25\n",
      "1/1 [==============================] - 0s 34ms/step - loss: 13177.1748 - mae: 111.0533 - mape: 52.6267\n",
      "Epoch 6/25\n",
      "1/1 [==============================] - 0s 43ms/step - loss: 14812.3301 - mae: 119.7828 - mape: 57.0726\n",
      "Epoch 7/25\n",
      "1/1 [==============================] - 0s 35ms/step - loss: 16448.1055 - mae: 126.0793 - mape: 60.1841\n",
      "Epoch 8/25\n",
      "1/1 [==============================] - 0s 36ms/step - loss: 10950.7207 - mae: 102.3802 - mape: 48.9209\n",
      "Epoch 9/25\n",
      "1/1 [==============================] - 0s 33ms/step - loss: 11809.3320 - mae: 107.8755 - mape: 51.3437\n",
      "Epoch 10/25\n",
      "1/1 [==============================] - 0s 34ms/step - loss: 8244.8105 - mae: 86.9301 - mape: 41.4268\n",
      "Epoch 11/25\n",
      "1/1 [==============================] - 0s 103ms/step - loss: 6276.1885 - mae: 74.9968 - mape: 35.7201\n",
      "0.6959896087646484\n",
      "Epoch 1/25\n",
      "1/1 [==============================] - 0s 77ms/step - loss: 4809.2637 - mae: 66.0649 - mape: 15.8649\n",
      "Epoch 2/25\n",
      "1/1 [==============================] - 0s 26ms/step - loss: 5933.0518 - mae: 74.2205 - mape: 17.7500\n",
      "Epoch 3/25\n",
      "1/1 [==============================] - 0s 34ms/step - loss: 8472.4131 - mae: 88.6880 - mape: 21.1713\n",
      "Epoch 4/25\n",
      "1/1 [==============================] - 0s 35ms/step - loss: 8189.0752 - mae: 86.9661 - mape: 20.7362\n",
      "Epoch 5/25\n",
      "1/1 [==============================] - 0s 31ms/step - loss: 10431.8604 - mae: 99.5825 - mape: 23.7748\n",
      "Epoch 6/25\n",
      "1/1 [==============================] - 0s 31ms/step - loss: 9172.5439 - mae: 93.7859 - mape: 22.4061\n",
      "Epoch 7/25\n",
      "1/1 [==============================] - 0s 33ms/step - loss: 6519.6787 - mae: 78.8432 - mape: 18.8788\n",
      "Epoch 8/25\n",
      "1/1 [==============================] - 0s 35ms/step - loss: 7454.8286 - mae: 81.5689 - mape: 19.5035\n",
      "Epoch 9/25\n",
      "1/1 [==============================] - 0s 33ms/step - loss: 5540.6016 - mae: 67.7995 - mape: 16.1311\n",
      "Epoch 10/25\n",
      "1/1 [==============================] - 0s 57ms/step - loss: 4066.7229 - mae: 59.5519 - mape: 14.2544\n",
      "Epoch 11/25\n",
      "1/1 [==============================] - 0s 56ms/step - loss: 2461.9558 - mae: 46.3471 - mape: 11.0170\n",
      "Epoch 12/25\n",
      "1/1 [==============================] - 0s 26ms/step - loss: 3288.2129 - mae: 49.3792 - mape: 11.6979\n",
      "Epoch 13/25\n",
      "1/1 [==============================] - 0s 62ms/step - loss: 1681.3667 - mae: 37.9355 - mape: 9.0577\n",
      "Epoch 14/25\n",
      "1/1 [==============================] - 0s 58ms/step - loss: 972.8364 - mae: 23.9590 - mape: 5.6577\n",
      "Epoch 15/25\n",
      "1/1 [==============================] - 0s 53ms/step - loss: 898.9604 - mae: 22.5503 - mape: 5.3497\n",
      "Epoch 16/25\n",
      "1/1 [==============================] - 0s 61ms/step - loss: 660.9006 - mae: 22.4777 - mape: 5.3918\n",
      "Epoch 17/25\n",
      "1/1 [==============================] - 0s 29ms/step - loss: 1014.6204 - mae: 30.0054 - mape: 7.2489\n",
      "Epoch 18/25\n",
      "1/1 [==============================] - 0s 39ms/step - loss: 777.7084 - mae: 25.1000 - mape: 5.9883\n",
      "Epoch 19/25\n",
      "1/1 [==============================] - 0s 28ms/step - loss: 2573.5435 - mae: 41.8806 - mape: 10.0260\n",
      "Epoch 20/25\n",
      "1/1 [==============================] - 0s 36ms/step - loss: 1123.8530 - mae: 27.1108 - mape: 6.5497\n",
      "Epoch 21/25\n",
      "1/1 [==============================] - 0s 34ms/step - loss: 2074.8596 - mae: 37.4552 - mape: 9.0365\n",
      "Epoch 22/25\n",
      "1/1 [==============================] - 0s 36ms/step - loss: 1700.2738 - mae: 33.4113 - mape: 7.8901\n",
      "Epoch 23/25\n",
      "1/1 [==============================] - 0s 31ms/step - loss: 839.0907 - mae: 21.8387 - mape: 5.3099\n",
      "Epoch 24/25\n",
      "1/1 [==============================] - 0s 37ms/step - loss: 1426.4460 - mae: 28.4030 - mape: 6.8272\n",
      "Epoch 25/25\n",
      "1/1 [==============================] - 0s 35ms/step - loss: 2299.7432 - mae: 42.4145 - mape: 10.1861\n",
      "1.2318246364593506\n"
     ]
    }
   ],
   "source": [
    "hist = fit_transformer(transformer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "6abdfae5-044d-441d-a943-bdefecfc6b19",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def convert_preds(dataa, timestep, NUM_FEATURES, ticker):\n",
    "    \"\"\"\n",
    "    Given last weeks actual data, forecasts next weeks prices.\n",
    "    \"\"\"\n",
    "    data_new = []\n",
    "\n",
    "    for i in range(len(ticker)):\n",
    "        reshaped_data = data.sentiment_ana(dataa, ticker, i)\n",
    "        samples = int(reshaped_data.shape[0] / timestep)\n",
    "        result = np.array(np.array_split(reshaped_data, samples))\n",
    "    \n",
    "        data_new.append(result.reshape((samples, timestep, NUM_FEATURES)))\n",
    "        \n",
    "        \n",
    "    return np.array(data_new)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "613fe8c6-a21e-406b-9469-e6f890c2b5de",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "413.4799265861511\n",
      "413.6057207584381\n",
      "413.72831106185913\n",
      "413.8509211540222\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "train = convert_preds(data.train, data.timestep, 3, data.ticker)\n",
    "test = convert_preds(data.test, data.timestep, 3, data.ticker)\n",
    "predict = []\n",
    "for i in range(len(data.ticker)):\n",
    "    predict.append(PredictAndForecast(transformer, train[i], test[i]))\n",
    "    print(time.time() - start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "212ddf75-0000-4541-a5db-8ea9511c8eaf",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[395.71426]\n",
      " [395.69742]]\n"
     ]
    }
   ],
   "source": [
    "print(predict[0].predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "7a203f32-7770-4e44-aaae-06761b18763c",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AAPL transformer_evals.var_ratio: 0.9999722872797813\n",
      "AAPL transformer_evals.mape: 1.2958752143234713\n",
      "MSFT transformer_evals.var_ratio: 0.9998997959201611\n",
      "MSFT transformer_evals.mape: 0.25226842580510617\n",
      "TSLA transformer_evals.var_ratio: 0.9755670137165503\n",
      "TSLA transformer_evals.mape: 0.9735102316621203\n",
      "NVDA transformer_evals.var_ratio: 0.9978212535084322\n",
      "NVDA transformer_evals.mape: 0.06623266268927176\n"
     ]
    }
   ],
   "source": [
    "for i in range(len(data.ticker)):\n",
    "    transformer_evals = Evaluate(test[i], predict[i].predictions)\n",
    "    print(f'{data.ticker[i]} transformer_evals.var_ratio: {transformer_evals.var_ratio}')\n",
    "    print(f'{data.ticker[i]} transformer_evals.mape: {transformer_evals.mape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "5cdde5f4-a42f-44f9-95e5-c43b39ef2d65",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def plot_results(test, preds, df, image_path=None, title_suffix=None, xlabel=None):\n",
    "  \"\"\"\n",
    "  Plots training data in blue, actual values in red, and predictions in green,\n",
    "  over time.\n",
    "  \"\"\"\n",
    "  fig, ax = plt.subplots(figsize=(20,6))\n",
    "  # x = df.Close[-498:].index\n",
    "  plot_test = test[1:,:,0]\n",
    "  print(plot_test)\n",
    "  #print(plot_test)\n",
    " \n",
    "  plot_preds = preds[1:]\n",
    "  #print(plot_preds)\n",
    "  x = df.index[-(plot_test.shape[0] * plot_test.shape[1]):]\n",
    "  #print(x)\n",
    "  plot_test = plot_test.reshape((plot_test.shape[0]*plot_test.shape[1], 1))\n",
    "  plot_preds = plot_preds.reshape((plot_test.shape[0]*plot_test.shape[1], 1))\n",
    "  ax.plot(x, plot_test, label='actual')\n",
    "  ax.plot(x, plot_preds, label='preds')\n",
    "    \n",
    "  if title_suffix is None:\n",
    "    ax.set_title('Predictions vs. Actual')\n",
    "  else:\n",
    "    ax.set_title(f'Predictions vs. Actual, {title_suffix}')\n",
    "\n",
    "  ax.set_xlabel('Date')\n",
    "  ax.set_ylabel(xlabel)\n",
    "  ax.legend()\n",
    "\n",
    "  if image_path is not None:\n",
    "    imagedir = '/content/drive/MyDrive/Colab Notebooks/images'\n",
    "    plt.savefig(f'{imagedir}/{image_path}.png')\n",
    "  plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "db500747-1aee-43d7-820e-dbb3e3aed16f",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[173.97000122]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[346.07000732]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[205.66000366]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[423.25]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(len(data.ticker)):\n",
    "    plot_results(test[i], predict[i].predictions, data.df[data.df.columns[i]], title_suffix='Transformer', xlabel = data.df.columns[i])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ade25b15-ee4d-4b42-a4b5-c58d0f301856",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}