regressDensity.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import glob\n",
    "import os\n",
    "\n",
    "import cv2\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.neural_network import MLPRegressor\n",
    "from sklearn.linear_model import LinearRegression, LassoCV, RidgeCV\n",
    "from sklearn.preprocessing import PolynomialFeatures, StandardScaler\n",
    "from lib.getDensity import getFilteredDensity\n",
    "import src.utils as utils\n",
    "import pickle\n",
    "from scipy.io import savemat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'np' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m~\\AppData\\Local\\Temp\\ipykernel_28740\\3085132345.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0m__name__\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;34m\"__main__\"\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m     \u001b[0mfolders\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0marray\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m33\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      3\u001b[0m     \u001b[0mux\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m53.69132701153139\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m     \u001b[0muy\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m56.463734106758785\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m     \u001b[0menableShear\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;32mTrue\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'np' is not defined"
     ]
    }
   ],
   "source": [
    "if __name__ == \"__main__\":\n",
    "    folders = np.array(range(33))\n",
    "    ux = 53.69132701153139\n",
    "    uy = 56.463734106758785\n",
    "    enableShear = True\n",
    "    outputshape = (111, 111)\n",
    "    feature = \"raw\"  # \"raw\" \"cart\" \"mlp\"\n",
    "    polyExtend = (\n",
    "        1, 1\n",
    "    )  # first is for coordinate poly order, second is for density poly order\n",
    "    path = \"flowdata1\"\n",
    "    exportPath = os.path.join(\"data\", \"densityResult\")\n",
    "    if feature == \"mlp\":\n",
    "        polyExtend = (0, 0)\n",
    "        regressor = MLPRegressor(hidden_layer_sizes=(\n",
    "            100,\n",
    "            100,\n",
    "            100,\n",
    "            10,\n",
    "            10,\n",
    "        ),\n",
    "                                 early_stopping=True,\n",
    "                                 max_iter=500)\n",
    "    else:\n",
    "        regressor = LinearRegression()\n",
    "        # regressor = RidgeCV()\n",
    "        # regressor = LassoCV()\n",
    "    total_densityList = []\n",
    "    total_depthList = []\n",
    "    exportList = []\n",
    "    densityMax = -1e9\n",
    "    densityMin = 1e9\n",
    "    depthList = []\n",
    "    densityList = []\n",
    "    flow_file_list = sorted(glob.glob(\n",
    "            os.path.join(path,\"*.npy\")))\n",
    "    h, w, _ = np.load(flow_file_list[0]).shape\n",
    "    croppedShape = (min(h, w), min(h, w))\n",
    "    x = np.linspace(0.0,\n",
    "                        croppedShape[1] - 30,\n",
    "                        croppedShape[1] - 30,\n",
    "                        endpoint=False)\n",
    "    y = np.linspace(0.0,\n",
    "                        croppedShape[0] - 30,\n",
    "                        croppedShape[0] - 30,\n",
    "                        endpoint=False)\n",
    "    axis = np.array(np.meshgrid(x, y)).transpose((1, 2, 0))\n",
    "    for i in range(len(flow_file_list)):\n",
    "            flow = np.load(flow_file_list[i])\n",
    "            if h > w:\n",
    "                flow = flow[(flow.shape[0] - croppedShape[0]) //\n",
    "                            2:-(flow.shape[0] - croppedShape[0]) //\n",
    "                            2, :, :].copy()\n",
    "            elif h < w:\n",
    "                flow = flow[:, (flow.shape[1] - croppedShape[1]) //\n",
    "                            2:-(flow.shape[1] - croppedShape[1]) //\n",
    "                            2, :].copy()\n",
    "            \n",
    "            density, zero_level = getFilteredDensity(flow, use_cuda=True)\n",
    "            # remapping\n",
    "            \"\"\"\n",
    "            density = (density + 1.0) / 1.0\n",
    "            density = (density * 255.0).astype('uint8')\n",
    "\n",
    "            density = density[20:-20, 20:-20]\n",
    "\n",
    "            \"\"\"\n",
    "            map = (axis + flow[15:-15, 15:-15, :]).astype(np.float32)\n",
    "            density = cv2.remap(density,\n",
    "                                map[:, :, 0],\n",
    "                                map[:, :, 1],\n",
    "                                interpolation=cv2.INTER_LINEAR,\n",
    "                                borderValue=float(zero_level))\n",
    "            # resizing\n",
    "            density = cv2.resize(density, (outputshape[1], outputshape[0]),\n",
    "                                 interpolation=cv2.INTER_CUBIC)\n",
    "            \n",
    "            \"\"\"\n",
    "            loc_ind = i // 46\n",
    "            # if loc_ind == 8 or loc_ind == 1 or loc_ind == 2 or loc_ind == 3 or loc_ind == 4 or loc_ind == 5 or loc_ind == 6 or loc_ind == 7:\n",
    "            #     continue\n",
    "            indent_ind = i - loc_ind * 46\n",
    "            max_loc = np.argmax(density)\n",
    "            max_y = max_loc // density.shape[1]\n",
    "            max_x = max_loc - max_y * density.shape[1]\n",
    "            \"\"\"\n",
    "            \n",
    "\n",
    "\n",
    "\n",
    "            densityList.append(density.max())\n",
    "\n",
    "            if i == 0:\n",
    "                depthList.append(0.0)\n",
    "            elif i == 1 or i == 5 or i == 9 or i == 13 or i == 17 or i == 21 or i == 25 or i == 29 or i == 33 or i == 37 or i == 41 or i == 45 or i == 49 or i == 53 or i ==57:\n",
    "                depthList.append(0.5)\n",
    "            elif i == 2 or i == 6 or i == 10 or i == 14 or i == 18 or i == 22 or i == 26 or i == 30 or i == 34 or i == 38 or i == 42 or i == 46 or i == 50 or i == 54 or i ==58:\n",
    "                depthList.append(1.0)\n",
    "            elif i == 3 or i == 7 or i == 11 or i == 15 or i== 19 or i == 23 or i == 27 or i == 31 or i == 35 or i == 39 or i == 43 or i == 47 or i == 51 or i == 55 or i ==59:\n",
    "                depthList.append(1.5)\n",
    "            elif i ==4 or i == 8 or i == 12 or i == 16 or i == 20 or i ==24 or i == 28 or i == 32 or i == 36 or i == 40 or i == 44 or i == 48 or i == 52 or i == 56 or i ==60:\n",
    "                depthList.append(2.0)\n",
    "            elif i>=61 and i<=70:\n",
    "                depthList.append((i-60)*0.2)\n",
    "            elif i>=71 and i<=80:\n",
    "                depthList.append((i-70)*0.2)\n",
    "            elif i>=81 and i<=90:\n",
    "                depthList.append((i-80)*0.2)\n",
    "            elif i>=91 and i<=100:\n",
    "                depthList.append((i-90)*0.2)\n",
    "            elif i>=101 and i<=110:\n",
    "                depthList.append((i-100)*0.2)\n",
    "            \"\"\"\n",
    "            if indent_ind == 0:\n",
    "                if len(depthList) != 0:\n",
    "                    # plt.plot(np.array(densityList)[:, -1],\n",
    "                    #          depthList,\n",
    "                    #          'x',\n",
    "                    #          label=str(folder))\n",
    "                    total_densityList.extend(densityList)\n",
    "                    total_depthList.extend(depthList)\n",
    "                depthList = []\n",
    "                densityList = []\n",
    "                depthList.append(0.0)\n",
    "                if feature == \"raw\":\n",
    "                    densityList.append((density.max()))\n",
    "                elif feature == \"cart\":\n",
    "                    densityList.append(\n",
    "                        (abs(max_x - ux), abs(max_y - uy), density.max()))\n",
    "                elif feature == \"mlp\":\n",
    "                    densityList.append((max_x, max_y, density.max()))\n",
    "                continue\n",
    "            indent_level = ((indent_ind - 1) // 9 + 1.0) * 0.5\n",
    "            if not enableShear and (indent_ind -\n",
    "                                    1) % 9 != 0:  # skip shear step\n",
    "                continue\n",
    "            else:\n",
    "                depthList.append(indent_level)\n",
    "                if feature == \"raw\":\n",
    "                    densityList.append((density.max()))\n",
    "                elif feature == \"cart\":\n",
    "                    densityList.append(\n",
    "                        (abs(max_x - ux), abs(max_y - uy), density.max()))\n",
    "                elif feature == \"mlp\":\n",
    "                    densityList.append((max_x, max_y, density.max()))\n",
    "                exportList.append(density)\n",
    "                if density.max() > densityMax:\n",
    "                    densityMax = density.max()\n",
    "                if density.min() < densityMin:\n",
    "                    densityMin = density.min()\n",
    "                # print(\"File: \", folder, \" Row: \", max_y, \" column: \", max_x,\n",
    "                #       \" r: \", r)\n",
    "                # plt.imshow(\n",
    "                #     cv2.applyColorMap(((density - density.min()) /\n",
    "                #                        (density.max() - density.min()) *\n",
    "                #                        255.0).astype('uint8'),\n",
    "                #                       cv2.COLORMAP_HOT))\n",
    "                # plt.show(block=True)\n",
    "                \"\"\"\n",
    "    #plt.plot(np.array(densityList)[:, -1], depthList, 'x', label=str(folder))\n",
    "    total_densityList.extend(densityList)\n",
    "    total_depthList.extend(depthList)\n",
    "    # plt.legend()\n",
    "    # plt.savefig(\"regress.jpg\")\n",
    "    X = np.array(total_densityList)\n",
    "    if len(X.shape) == 1:\n",
    "        X = X.reshape((-1, 1))\n",
    "    # s = StandardScaler()\n",
    "    # X = s.fit_transform(X)\n",
    "    y = np.array(total_depthList)\n",
    "    savemat(os.path.join(exportPath, \"density2matlab\"), {'X': X, 'y': y})\n",
    "    if polyExtend[0] > 1 or polyExtend[1] > 1:\n",
    "        poly1 = PolynomialFeatures(polyExtend[0], include_bias=False)\n",
    "        poly2 = PolynomialFeatures(polyExtend[1], include_bias=False)\n",
    "        if X.shape[1] == 1:\n",
    "            X = poly2.fit_transform(X)\n",
    "        else:\n",
    "            X = np.hstack(\n",
    "                (poly1.fit_transform(X[:, :-1]), poly2.fit_transform(X[:,\n",
    "                                                                       -1:])))\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X,\n",
    "                                                        y,\n",
    "                                                        test_size=0.2,\n",
    "                                                        random_state=42)\n",
    "    regressor.fit(X_train, y_train)\n",
    "    print(\"Model type: \", type(regressor), \" feature order: \", polyExtend)\n",
    "    print(\"Mean test error: \",\n",
    "          np.abs(regressor.predict(X_test) - y_test).mean())\n",
    "    plt.plot(X_train, y_train , 'x')\n",
    "    plt.plot(X_test,regressor.predict(X_test),color='black',label='best line')\n",
    "    plt.savefig(\"prediction.jpg\")\n",
    "    if feature != \"mlp\":\n",
    "        print(regressor.coef_, regressor.intercept_)\n",
    "    with open('depthPredictor.pkl', 'wb') as f:\n",
    "        pickle.dump((regressor, feature, polyExtend), f)\n",
    "    # for i in range(len(exportList)):\n",
    "    #     im = (exportList[i] - densityMin) / (densityMax - densityMin) * 255.0\n",
    "    #     im = cv2.applyColorMap(im.astype('uint8'), cv2.COLORMAP_HOT)\n",
    "    #     cv2.imwrite(os.path.join(exportPath, str(i) + \".jpg\"), im)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'pickle' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m~\\AppData\\Local\\Temp\\ipykernel_24552\\4251089993.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'depthPredictor.pkl'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'rb'\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0mf\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      2\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 3\u001b[1;33m     \u001b[0mmodel\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mpickle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mload\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mf\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      4\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m \u001b[0mtest\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0marray\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m1.65\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'pickle' is not defined"
     ]
    }
   ],
   "source": [
    "with open('depthPredictor.pkl','rb') as f:\n",
    "\n",
    "    model = pickle.load(f)\n",
    "\n",
    "test = np.array(1.65)\n",
    "\n",
    "test.resize(1,1)\n",
    "\n",
    "regressor1 = model[0]\n",
    "\n",
    "regressor1.predict(test)\n",
    " \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = np.linspace(0,\n",
    "                                19,\n",
    "                                20,\n",
    "                                dtype='float32')\n",
    "y = np.linspace(0,\n",
    "                                19,\n",
    "                                20,\n",
    "                                dtype='float32')\n",
    "map = np.array(np.meshgrid(x, y)).transpose((1, 2, 0))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Vision",
   "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": "3.7.13"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}