PyTorch Encoder Decoder Attempt.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ab68e127",
   "metadata": {},
   "source": [
    "# Trying PyTorch for LSTM Encoder / Decoder\n",
    "borrowed from interwebs\n",
    "https://github.com/lkulowski/LSTM_encoder_decoder"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2016dd97",
   "metadata": {},
   "source": [
    "## Standard header"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3674bd47",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Standard Header used on the projects\n",
    "\n",
    "#first the major packages used for math and graphing\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from cycler import cycler\n",
    "import scipy.special as sp\n",
    "\n",
    "#Custome graph format style sheet\n",
    "#plt.style.use('Prospectus.mplstyle')\n",
    "\n",
    "#If being run by a seperate file, use the seperate file's graph format and saving paramaeters\n",
    "#otherwise set what is needed\n",
    "if not 'Saving' in locals():\n",
    "    Saving = False\n",
    "if not 'Titles' in locals():\n",
    "    Titles = True\n",
    "if not 'Ledgends' in locals():\n",
    "    Ledgends = True\n",
    "if not 'FFormat' in locals():\n",
    "    FFormat = '.png'\n",
    "\n",
    "#Standard cycle to make black and white images and dashed and line styles\n",
    "default_cycler = (cycler('color', ['0.00', '0.40', '0.60', '0.70']) + cycler(linestyle=['-', '-', '-', '-']))\n",
    "plt.rc('axes', prop_cycle=default_cycler)\n",
    "my_cmap = plt.get_cmap('gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "93fad782",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Extra Headers:\n",
    "import os as os\n",
    "import pywt as py\n",
    "import statistics as st\n",
    "import os as os\n",
    "import random\n",
    "import multiprocessing\n",
    "from joblib import Parallel, delayed\n",
    "import platform\n",
    "\n",
    "from time import time as ti"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f7693c3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "#import CoreFunctions as cf\n",
    "from skimage.restoration import denoise_wavelet"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d43eb88e",
   "metadata": {},
   "source": [
    "## Choosing platfrom and folders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c695da0e",
   "metadata": {},
   "outputs": [],
   "source": [
    "HostName = platform.node()\n",
    "\n",
    "if HostName == \"Server\":\n",
    "    Computer = \"Desktop\"   \n",
    "elif HostName[-6:] == 'wm.edu':\n",
    "    Computer = \"SciClone\"\n",
    "elif HostName == \"SchoolLaptop\":\n",
    "    Computer = \"LinLap\"\n",
    "elif HostName == \"WTC-TAB-512\":\n",
    "    Computer = \"PortLap\"\n",
    "else:\n",
    "    Computer = \"WinLap\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9a5a3518",
   "metadata": {},
   "outputs": [],
   "source": [
    "if Computer == \"SciClone\":\n",
    "    location = '/sciclone/home20/dchendrickson01/image/'\n",
    "elif Computer == \"WinLap\":\n",
    "    location = 'C:\\\\Data\\\\'\n",
    "elif Computer == \"Desktop\":\n",
    "    location = \"E:\\\\Backups\\\\Dan\\\\CraneData\\\\\"\n",
    "elif Computer == \"LinLap\":\n",
    "    location = '/home/dan/Output/'\n",
    "elif Computer == 'PortLap':\n",
    "    location = 'C:\\\\users\\\\dhendrickson\\\\Desktop\\\\AccelData\\\\'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "484fbcc9",
   "metadata": {},
   "outputs": [],
   "source": [
    "if Computer ==  \"SciClone\":\n",
    "    rootfolder = '/sciclone/home20/dchendrickson01/'\n",
    "    folder = '/sciclone/scr10/dchendrickson01/Recordings2/'\n",
    "    imageFolder = '/sciclone/scr10/dchendrickson01/Move3Dprint/'\n",
    "elif Computer == \"Desktop\":\n",
    "    rootfolder = location\n",
    "    folder = rootfolder + \"Recordings2\\\\\"\n",
    "elif Computer ==\"WinLap\":\n",
    "    rootfolder = location\n",
    "    folder = rootfolder + \"Recordings2\\\\\"   \n",
    "elif Computer == \"LinLap\":\n",
    "    rootfolder = '/home/dan/Data/'\n",
    "    folder = rootfolder + 'Recordings2/'\n",
    "elif Computer =='PortLap':\n",
    "    rootfolder = location \n",
    "    folder = rootfolder + 'Recordings2\\\\'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d7dd29f",
   "metadata": {},
   "source": [
    "## Start Code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "c0cdb2ac",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Author: Laura Kulowski\n",
    "\n",
    "'''\n",
    "Example of using a LSTM encoder-decoder to model a synthetic time series \n",
    "https://github.com/lkulowski/LSTM_encoder_decoder\n",
    "\n",
    "'''\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from importlib import reload\n",
    "import sys\n",
    "\n",
    "import lstm_encoder_decoder\n",
    "import plotting\n",
    "\n",
    "matplotlib.rcParams.update({'font.size': 17})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "29d1a11c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "f0be3690",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_test_split(t, y, split = 0.8):\n",
    "\n",
    "    '''\n",
    "    \n",
    "    split time series into train/test sets\n",
    "    \n",
    "    : param t:                      time array\n",
    "    : para y:                       feature array\n",
    "    : para split:                   percent of data to include in training set \n",
    "    : return t_train, y_train:      time/feature training and test sets;  \n",
    "    :        t_test, y_test:        (shape: [# samples, 1])\n",
    "    \n",
    "    '''\n",
    "    \n",
    "    indx_split = int(split * len(y))\n",
    "    indx_train = np.arange(0, indx_split)\n",
    "    indx_test = np.arange(indx_split, len(y))\n",
    "\n",
    "    t_train = t[indx_train]\n",
    "    y_train = y[indx_train]\n",
    "    y_train = y_train.reshape(-1, 1)\n",
    "    \n",
    "    t_test = t[indx_test]\n",
    "    y_test = y[indx_test]\n",
    "    y_test = y_test.reshape(-1, 1)\n",
    "    \n",
    "    return t_train, y_train, t_test, y_test "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "57ed2096",
   "metadata": {},
   "outputs": [],
   "source": [
    "def windowed_dataset(y, input_window = 5, output_window = 1, stride = 1, num_features = 1):\n",
    "  \n",
    "    '''\n",
    "    create a windowed dataset\n",
    "    \n",
    "    : param y:                time series feature (array)\n",
    "    : param input_window:     number of y samples to give model \n",
    "    : param output_window:    number of future y samples to predict  \n",
    "    : param stide:            spacing between windows   \n",
    "    : param num_features:     number of features (i.e., 1 for us, but we could have multiple features)\n",
    "    : return X, Y:            arrays with correct dimensions for LSTM\n",
    "    :                         (i.e., [input/output window size # examples, # features])\n",
    "    '''\n",
    "  \n",
    "    L = y.shape[0]\n",
    "    num_samples = (L - input_window - output_window) // stride + 1\n",
    "\n",
    "    X = np.zeros([input_window, num_samples, num_features])\n",
    "    Y = np.zeros([output_window, num_samples, num_features])    \n",
    "    \n",
    "    for ff in np.arange(num_features):\n",
    "        for ii in np.arange(num_samples):\n",
    "            start_x = stride * ii\n",
    "            end_x = start_x + input_window\n",
    "            X[:, ii, ff] = y[start_x:end_x, ff]\n",
    "\n",
    "            start_y = stride * ii + input_window\n",
    "            end_y = start_y + output_window \n",
    "            Y[:, ii, ff] = y[start_y:end_y, ff]\n",
    "\n",
    "    return X, Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "05864bd1",
   "metadata": {},
   "outputs": [],
   "source": [
    "def numpy_to_torch(Xtrain, Ytrain, Xtest, Ytest):\n",
    "    '''\n",
    "    convert numpy array to PyTorch tensor\n",
    "    \n",
    "    : param Xtrain:                           windowed training input data (input window size, # examples, # features); np.array\n",
    "    : param Ytrain:                           windowed training target data (output window size, # examples, # features); np.array\n",
    "    : param Xtest:                            windowed test input data (input window size, # examples, # features); np.array\n",
    "    : param Ytest:                            windowed test target data (output window size, # examples, # features); np.array\n",
    "    : return X_train_torch, Y_train_torch,\n",
    "    :        X_test_torch, Y_test_torch:      all input np.arrays converted to PyTorch tensors \n",
    "\n",
    "    '''\n",
    "    \n",
    "    X_train_torch = torch.from_numpy(Xtrain).type(torch.Tensor)\n",
    "    Y_train_torch = torch.from_numpy(Ytrain).type(torch.Tensor)\n",
    "\n",
    "    X_test_torch = torch.from_numpy(Xtest).type(torch.Tensor)\n",
    "    Y_test_torch = torch.from_numpy(Ytest).type(torch.Tensor)\n",
    "    \n",
    "    return X_train_torch, Y_train_torch, X_test_torch, Y_test_torch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "74fc65b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "def SmoothMoves(file):\n",
    "    #    if file[-3:] =='csv':\n",
    "    ODataSet = np.genfromtxt(open(folder+file,'r'), delimiter=',',skip_header=0,missing_values=0,invalid_raise=False)\n",
    "    SmoothX = denoise_wavelet(ODataSet[:,3], method='VisuShrink', mode='soft', wavelet_levels=3, wavelet='sym2', rescale_sigma='True')\n",
    "    #SmoothY = denoise_wavelet(ODataSet[:,4], method='VisuShrink', mode='soft', wavelet_levels=3, wavelet='sym2', rescale_sigma='True')\n",
    "    #SmoothZ = denoise_wavelet(ODataSet[:,5], method='VisuShrink', mode='soft', wavelet_levels=3, wavelet='sym2', rescale_sigma='True')\n",
    "    SmoothX -= np.average(SmoothX)\n",
    "    Time = []\n",
    "    for r in ODataSet:\n",
    "        hr = int(r[1]/10000)\n",
    "        mint = int((r[1]-10000*hr)/100)\n",
    "        sec = r[1] % 100\n",
    "        t = hr * 3600 + mint * 60 + sec + r[2]/10000.0\n",
    "        Time.append(t)\n",
    "    #SmoothY -= np.average(SmoothY)\n",
    "    #SmoothZ -= np.average(SmoothZ)\n",
    "    #MoveMatrix = np.matrix([SmoothX, SmoothY, SmoothZ])\n",
    "    return SmoothX, np.asarray(Time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "a6f5f2ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "#----------------------------------------------------------------------------------------------------------------\n",
    "# generate dataset for LSTM\n",
    "y, t = SmoothMoves('230418 recording1.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "cf9dc9dd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18659125 18659125\n"
     ]
    }
   ],
   "source": [
    "print(len(t), len(y))\n",
    "\n",
    "#t=np.asarray(t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "810d0853",
   "metadata": {},
   "outputs": [],
   "source": [
    "t_train, y_train, t_test, y_test = train_test_split(t, y, split = 0.8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "9c42f1da",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot time series \n",
    "plt.figure(figsize = (18, 6))\n",
    "plt.plot(t, y, color = 'k', linewidth = 2)\n",
    "plt.xlim([t[0], t[-1]])\n",
    "plt.xlabel('$t$')\n",
    "plt.ylabel('$y$')\n",
    "plt.title('Synthetic Time Series')\n",
    "plt.savefig('plots/synthetic_time_series.png')\n",
    "\n",
    "# plot time series with train/test split\n",
    "plt.figure(figsize = (18, 6))\n",
    "plt.plot(t_train, y_train, color = '0.4', linewidth = 2, label = 'Train') \n",
    "plt.plot(np.concatenate([[t_train[-1]], t_test]), np.concatenate([[y_train[-1]], y_test]),\n",
    "         color = (0.74, 0.37, 0.22), linewidth = 2, label = 'Test')\n",
    "plt.xlim([t[0], t[-1]])\n",
    "plt.xlabel(r'$t$')\n",
    "plt.ylabel(r'$y$')\n",
    "plt.title('Time Series Split into Train and Test Sets')\n",
    "plt.legend(bbox_to_anchor=(1, 1))\n",
    "plt.tight_layout\n",
    "plt.savefig('plots/train_test_split.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "43d921c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "#----------------------------------------------------------------------------------------------------------------\n",
    "# window dataset\n",
    "\n",
    "# set size of input/output windows \n",
    "iw = 80 \n",
    "ow = 20 \n",
    "s = 5\n",
    "\n",
    "# generate windowed training/test datasets\n",
    "Xtrain, Ytrain= windowed_dataset(y_train, input_window = iw, output_window = ow, stride = s)\n",
    "Xtest, Ytest = windowed_dataset(y_test, input_window = iw, output_window = ow, stride = s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "7e050ccd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3731787\n"
     ]
    }
   ],
   "source": [
    "print(np.shape(Xtrain)[1]+np.shape(Xtest)[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c7932011",
   "metadata": {},
   "outputs": [],
   "source": [
    "testing = np.concatenate((Xtrain, Xtest), axis =1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "76c525fd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(80, 3731787, 1)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.shape(testing)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a991fc7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "asdfasdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b5269fad",
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot example of windowed data  \n",
    "plt.figure(figsize = (10, 6)) \n",
    "plt.plot(np.arange(0, iw), Xtrain[:, 0, 0], 'k', linewidth = 2.2, label = 'Input')\n",
    "plt.plot(np.arange(iw - 1, iw + ow), np.concatenate([[Xtrain[-1, 0, 0]], Ytrain[:, 0, 0]]),\n",
    "         color = (0.2, 0.42, 0.72), linewidth = 2.2, label = 'Target')\n",
    "plt.xlim([0, iw + ow - 1])\n",
    "plt.xlabel(r'$t$')\n",
    "plt.ylabel(r'$y$')\n",
    "plt.title('Example of Windowed Training Data')\n",
    "plt.legend(bbox_to_anchor=(1.3, 1))\n",
    "plt.tight_layout() \n",
    "plt.savefig('plots/windowed_data.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d0907cb7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#----------------------------------------------------------------------------------------------------------------\n",
    "# LSTM encoder-decoder\n",
    "\n",
    "# convert windowed data from np.array to PyTorch tensor\n",
    "X_train, Y_train, X_test, Y_test = numpy_to_torch(Xtrain, Ytrain, Xtest, Ytest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97579802",
   "metadata": {},
   "outputs": [],
   "source": [
    "# specify model parameters and train\n",
    "model = lstm_encoder_decoder.lstm_seq2seq(input_size = X_train.shape[2], hidden_size = 15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c0d03ecf",
   "metadata": {},
   "outputs": [],
   "source": [
    "loss = model.train_model(X_train, Y_train, n_epochs = 25, target_len = ow, batch_size = 1, training_prediction = 'mixed_teacher_forcing', teacher_forcing_ratio = 0.6, learning_rate = 0.01, dynamic_tf = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "48cea0f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot predictions on train/test data\n",
    "plotting.plot_train_test_results(model, Xtrain, Ytrain, Xtest, Ytest)\n",
    "\n",
    "plt.close('all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06ef70a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save(\"PyTorch_LSTM_xyz\")"
   ]
  }
 ],
 "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}