python/gray_rotate_two.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6db6b936",
   "metadata": {},
   "source": [
    "# python libs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0ca016ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "import gluoncv as gcv\n",
    "import mxnet as mx\n",
    "# mxnet NDarray \n",
    "from mxnet import nd\n",
    "\n",
    "# MXNet Gluon Neural network layers.\n",
    "from mxnet.gluon import nn\n",
    "http://localhost:8888/notebooks/gray_rotate_two.ipynb#\n",
    "# optional for displaying the image\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import math\n",
    "# pip install pillow but import PIL\n",
    "#from PIL import Image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7586cc84",
   "metadata": {},
   "source": [
    "# Load image and convert to grayscale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "bfcc30cb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "nd_array_f3201_HWC3.shape =  (427, 640, 3)\n",
      "nd_array_f3201_HWC1.shape =  (427, 640, 1)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f1a1c184908>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "a_file_name = '0.jpg'\n",
    "nd_array_uint8_HWC3 = mx.image.imread(a_file_name, flag=1)\n",
    "np_array = nd_array_uint8_HWC3.asnumpy()\n",
    "gray_aug = mx.image.RandomGrayAug(p=1)\n",
    "nd_array_f32_HWC3 = nd_array_uint8_HWC3.astype(\"float32\")\n",
    "nd_array_f3201_HWC3 = nd_array_f32_HWC3/255\n",
    "aug = mx.image.RandomGrayAug(p=1)\n",
    "nd_array_f3201_HWC3 = aug(nd_array_f3201_HWC3)\n",
    "print('nd_array_f3201_HWC3.shape = ', nd_array_f3201_HWC3.shape)\n",
    "gray_np_array_f3201_HWC3 = nd_array_f3201_HWC3.asnumpy()\n",
    "#nd_array_f3201_HWC1 = nd_array_f3201_HWC3[:,:,0]  # this makes a shape like (y,x)\n",
    "nd_array_f3201_HWC1 = nd_array_f3201_HWC3.slice(begin=(0,0,0), end=(428,641,1))  # this give (y,x,1)\n",
    "print('nd_array_f3201_HWC1.shape = ', nd_array_f3201_HWC1.shape)\n",
    "gray_np_array_f3201_HWC1 = nd_array_f3201_HWC1.asnumpy()\n",
    "plt.imshow(gray_np_array_f3201_HWC1, cmap='gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39df3169",
   "metadata": {},
   "source": [
    "# rotate"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eac80374",
   "metadata": {},
   "source": [
    "From this source\n",
    "\n",
    "https://en.wikipedia.org/wiki/Rotation_matrix\n",
    "\n",
    "\n",
    "\\begin{vmatrix}\n",
    "\\cos(\\theta) & -\\sin(\\theta) \\\\\n",
    "\\sin(\\theta) & \\cos(\\theta)\n",
    "\\end{vmatrix}\n",
    "\n",
    "\n",
    "rotates points in the xy plane counterclockwise through an angle θ with respect to the positive x axis about the origin of a two-dimensional Cartesian coordinate system. To perform the rotation on a plane point with standard coordinates v = (x, y), it should be written as a column vector, and multiplied by the matrix R:\n",
    "\n",
    "\n",
    "\\begin{align}\n",
    "\\begin{vmatrix}\n",
    "\\cos(\\theta) & -\\sin(\\theta) \\\\\n",
    "\\sin(\\theta) & \\cos(\\theta)\n",
    "\\end{vmatrix}\n",
    "\\begin{vmatrix} x \\\\ y \\end{vmatrix} = \n",
    "\\begin{vmatrix}\n",
    "x \\cos(\\theta) - y \\sin(\\theta) \\\\\n",
    "x \\sin(\\theta) + y \\cos(\\theta)\n",
    "\\end{vmatrix}\n",
    "\\end{align}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "792f8866",
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_rot_matrix(angle):\n",
    "    #  a  b\n",
    "    #  c  d\n",
    "    \n",
    "    # calculate terms\n",
    "    a = math.cos(x)\n",
    "    b = -math.sin(x)\n",
    "    c = math.sin(x)\n",
    "    d = math.cos(x)\n",
    "\n",
    "    # print(\"a = \", a)\n",
    "    # print(\"b = \", b)\n",
    "    # print(\"c = \", c)\n",
    "    # print(\"d = \", d)\n",
    "    \n",
    "    # build np array\n",
    "    np_array = np.array([\n",
    "        [a,b],\n",
    "        [c,d]\n",
    "    ])\n",
    "    \n",
    "    # build mx.nd array\n",
    "    nd_array = mx.nd.array(np_array)\n",
    "    \n",
    "    return nd_array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b04007fa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rotate by 90 nd array is  \n",
      "[[ 6.123234e-17 -1.000000e+00]\n",
      " [ 1.000000e+00  6.123234e-17]]\n",
      "<NDArray 2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "x=0.5*math.pi\n",
    "\n",
    "rotate_90 = build_rot_matrix(x)\n",
    "print('rotate by 90 nd array is ', rotate_90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2eaf8fc2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "image shape =  (427, 640, 1)\n",
      "image shape =  (427, 640, 1)\n"
     ]
    }
   ],
   "source": [
    "print(\"image shape = \", nd_array_f3201_HWC1.shape)\n",
    "print(\"image shape = \", nd_array_f3201_HWC1.shape)\n",
    "#print(\"image = \", nd_array_f3201_HWC1)\n",
    "#gray_np_array_f3201_HWC1 = nd_array_f3201_HWC1.asnumpy()\n",
    "#plt.imshow(gray_np_array_f3201_HWC1, cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "24e7477c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rotate shape =  (2, 2)\n",
      "rotate shape =  (1, 2, 2)\n",
      "rotate =  \n",
      "[[[ 6.123234e-17 -1.000000e+00]\n",
      "  [ 1.000000e+00  6.123234e-17]]]\n",
      "<NDArray 1x2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "print('rotate shape = ', rotate_90.shape)\n",
    "# add a dimension\n",
    "#rotate_90_221 = rotate_90.reshape(2,2,1)\n",
    "#ndarray.stack([a, a]).T\n",
    "rotate_90_221=nd.expand_dims(rotate_90, axis=0)\n",
    "print('rotate shape = ', rotate_90_221.shape)\n",
    "print('rotate = ', rotate_90_221)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d70efdbf",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "5ac2642d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# chance from HWC order to CHW order\n",
    "nd_array_f3201_C1HW = nd.transpose(nd_array_f3201_HWC1, axes=(2,0,1) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "1710e9ed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 427, 640)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nd_array_f3201_C1HW.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "235b5331",
   "metadata": {},
   "outputs": [],
   "source": [
    "# I didn't know what I was doing.  I need to rethink this on a small scale."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "0ed0bfda",
   "metadata": {},
   "outputs": [],
   "source": [
    "# rotate_img = nd_array_f3201_C1HW * rotate_90_221"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2773a3bd",
   "metadata": {},
   "source": [
    "# Reattempt of rotation with small image\n",
    "\n",
    "This image is already grayscale so the augmentation step is skipped.\n",
    "\n",
    "Also, I am trying to rotate a grayscale image.  As a result, I am removing the channel dimension.  FWIW, I am using grayscale so that I can learn how to do mxnet.  Later, I will try to rotate a color image. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "51969016",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "nd_array_uint8_HWC1.shape =  (480, 640, 1)\n",
      "nd_array_f3201_HWC1.shape =  (480, 640, 1)\n",
      "nd_array_f3201_HW.shape =  (480, 640)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f1a0cbfb518>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Setup the sample image file paths\n",
    "a_file_name = 'imgs/testy-grid.png'\n",
    "nd_array_uint8_HWC1 = mx.image.imread(a_file_name, flag=0)\n",
    "print('nd_array_uint8_HWC1.shape = ', nd_array_uint8_HWC1.shape)\n",
    "np_array = nd_array_uint8_HWC1.asnumpy()\n",
    "nd_array_f32_HWC1 = nd_array_uint8_HWC1.astype(\"float32\")\n",
    "nd_array_f3201_HWC1 = nd_array_f32_HWC1/255\n",
    "print('nd_array_f3201_HWC1.shape = ', nd_array_f3201_HWC1.shape)\n",
    "nd_array_f3201_HW = nd_array_f3201_HWC1[:,:,0]  # this makes a shape like (y,x)\n",
    "print('nd_array_f3201_HW.shape = ', nd_array_f3201_HW.shape)\n",
    "np_array_f3201_HW = nd_array_f3201_HW.asnumpy()\n",
    "plt.imshow(np_array_f3201_HW, cmap='gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8bf9a4bd",
   "metadata": {},
   "source": [
    "# Subset the image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "eac44e61",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f1a0cb68fd0>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "subset_nd_array_f3201_HW = nd_array_f3201_HW[52:56,94:98]\n",
    "subset_np_array_f3201_HW = subset_nd_array_f3201_HW.asnumpy()\n",
    "plt.imshow(subset_np_array_f3201_HW, cmap='gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7b89c4f",
   "metadata": {},
   "source": [
    "# Iterate the pixels\n",
    "\n",
    "A couple of links found\n",
    "\n",
    "* https://mxnet.apache.org/versions/1.5.0/api/python/io/io.html\n",
    "* https://mxnet.apache.org/versions/1.1.0/tutorials/basic/data.html\n",
    "* https://mxnet.apache.org/versions/1.5.0/tutorials/basic/data.html\n",
    "* https://mxnet.apache.org/versions/0.11.0/tutorials/r/CustomIteratorTutorial.html\n",
    "\n",
    "Nothing to do with iterators, but it looks interesting for beginners\n",
    "\n",
    "* https://mxnet.apache.org/versions/1.1.0/tutorials/basic/ndarray.html\n",
    "\n",
    "A book, what is Julia?  It shows some mxnet.\n",
    "\n",
    "* https://books.google.com/books?id=JMNiDwAAQBAJ&pg=PA147&lpg=PA147&dq=iterate+mxnet+iterate+a+matrix&source=bl&ots=8tWjjWCzsz&sig=ACfU3U3967qA3cNGQjQxuSoACFgRySqBhA&hl=en&sa=X&ved=2ahUKEwihrNXzsof2AhW1GDQIHVogD4YQ6AF6BAgREAM#v=onepage&q=iterate%20mxnet%20iterate%20a%20matrix&f=false\n",
    "\n",
    "A channel RGB only image based iterator\n",
    "\n",
    "* https://mxnet.apache.org/versions/1.5.0/api/python/image/image.html#mxnet.image.ImageIter\n",
    "\n",
    "Finally one which looks appropriate\n",
    "\n",
    "* https://mxnet.apache.org/versions/1.6/api/python/docs/api/ndarray/ndarray.html#mxnet.ndarray.broadcast_mul\n",
    "* https://mxnet.incubator.apache.org/api/faq/new_op\n",
    "* https://mxnet.apache.org/versions/1.1.0/api/python/io/io.html#mxnet.io.NDArrayIter\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "cb824f4a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1.         1.         1.         0.01176471]\n",
       " [1.         1.         1.         0.01176471]\n",
       " [1.         1.         1.         0.01176471]\n",
       " [0.01176471 0.01176471 0.01176471 0.01176471]]\n",
       "<NDArray 4x4 @cpu(0)>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "subset_nd_array_f3201_HW"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c3521015",
   "metadata": {},
   "source": [
    "# Iteration routines"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93130c7f",
   "metadata": {},
   "source": [
    "## Numpy NDArrayIter\n",
    "\n",
    "I found this via stackoverflow.  Not really interested in it though.  I want to stick with MXNet/NDArray API."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "be577dd5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(0, 0) 1 -1\n",
      "(0, 1) 2 -2\n",
      "(0, 2) 3 -3\n",
      "(0, 3) 4 -4\n",
      "(1, 0) 5 -5\n",
      "(1, 1) 6 -6\n",
      "(1, 2) 7 -7\n",
      "(1, 3) 8 -8\n"
     ]
    }
   ],
   "source": [
    "a = np.arange(1,9).reshape(2,4)\n",
    "b = -a\n",
    " \n",
    "it = np.nditer((a, b), order='C', flags=('multi_index',))\n",
    "for i, j in it:\n",
    "    print(it.multi_index, i, j)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6164ecff",
   "metadata": {},
   "source": [
    "## NDArrayIter\n",
    "\n",
    "\n",
    "[docs](https://mxnet.apache.org/versions/1.6/api/python/docs/api/mxnet/io/index.html#mxnet.io.NDArrayIter)\n",
    "\n",
    "```\n",
    "mxnet.io.NDArrayIter(data, \n",
    "                     label=None, \n",
    "                     batch_size=1, \n",
    "                     shuffle=False, \n",
    "                     last_batch_handle='pad', \n",
    "                     data_name='data', \n",
    "                     label_name='softmax_label')\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "f794e7f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "data is  [[[ 0  1  2  3]\n",
      "  [ 4  5  6  7]\n",
      "  [ 8  9 10 11]\n",
      "  [12 13 14 15]]]\n",
      "loop entry\n",
      "batch.data[0] =  [[[ 0.  1.  2.  3.]\n",
      "  [ 4.  5.  6.  7.]\n",
      "  [ 8.  9. 10. 11.]\n",
      "  [12. 13. 14. 15.]]]\n",
      "batch.data[0].shape =  (1, 4, 4)\n"
     ]
    }
   ],
   "source": [
    "# batch 1\n",
    "data = np.arange(16).reshape((1,4,4))\n",
    "print('data is ', data)\n",
    "labels = np.ones([1, 1])\n",
    "dataiter = mx.io.NDArrayIter(data, labels, 1, False, last_batch_handle='discard')\n",
    "for batch in dataiter:\n",
    "    print('loop entry')\n",
    "    #print ('batch.data[0] = ', batch.data[0].asnumpy())\n",
    "    print ('batch.data[0] = ', batch.data[0].asnumpy())\n",
    "    print ('batch.data[0].shape = ', batch.data[0].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "1363e5de",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loop entry\n",
      "[[[12. 13.]\n",
      "  [14. 15.]]\n",
      "\n",
      " [[32. 33.]\n",
      "  [34. 35.]]\n",
      "\n",
      " [[ 0.  1.]\n",
      "  [ 2.  3.]]]\n",
      "(3, 2, 2)\n",
      "loop entry\n",
      "[[[ 4.  5.]\n",
      "  [ 6.  7.]]\n",
      "\n",
      " [[28. 29.]\n",
      "  [30. 31.]]\n",
      "\n",
      " [[24. 25.]\n",
      "  [26. 27.]]]\n",
      "(3, 2, 2)\n",
      "loop entry\n",
      "[[[ 8.  9.]\n",
      "  [10. 11.]]\n",
      "\n",
      " [[16. 17.]\n",
      "  [18. 19.]]\n",
      "\n",
      " [[20. 21.]\n",
      "  [22. 23.]]]\n",
      "(3, 2, 2)\n"
     ]
    }
   ],
   "source": [
    "# batch 3\n",
    "data = np.arange(40).reshape((10,2,2))\n",
    "labels = np.ones([10, 1])\n",
    "dataiter = mx.io.NDArrayIter(data, labels, 3, True, last_batch_handle='discard')\n",
    "for batch in dataiter:\n",
    "    print('loop entry')\n",
    "    print (batch.data[0].asnumpy())\n",
    "    print (batch.data[0].shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0b5c8c0",
   "metadata": {},
   "source": [
    "## NDArray Broadcast Multiply\n",
    "\n",
    "[docs](https://mxnet.apache.org/versions/master/api/python/docs/api/legacy/ndarray/ndarray.html#mxnet.ndarray.broadcast_mul)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "b3395e89",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x is  \n",
      "[[1. 1. 1.]\n",
      " [1. 1. 1.]]\n",
      "<NDArray 2x3 @cpu(0)>\n",
      "y is  \n",
      "[[0.]\n",
      " [1.]]\n",
      "<NDArray 2x1 @cpu(0)>\n",
      "z is  \n",
      "[[0. 0. 0.]\n",
      " [1. 1. 1.]]\n",
      "<NDArray 2x3 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# mxnet.ndarray.broadcast_mul(lhs=None, rhs=None, out=None, name=None, **kwargs)\n",
    "\n",
    "# This is a good choice for submititing a patch\n",
    "\n",
    "#\n",
    "# x = [[ 1.,  1.,  1.],\n",
    "#      [ 1.,  1.,  1.]]\n",
    "# \n",
    "# y = [[ 0.],\n",
    "#     [ 1.]]\n",
    "#\n",
    "#mx.nd.broadcast_mul(x, y) = [[ 0.,  0.,  0.],\n",
    "#                            [ 1.,  1.,  1.]]\n",
    "\n",
    "\n",
    "x = nd.array([[ 1.,  1.,  1.],[ 1.,  1.,  1.]])\n",
    "print('x is ', x)\n",
    "y = nd.array([[ 0.],[ 1.]])\n",
    "print('y is ', y)\n",
    "z = mx.ndarray.broadcast_mul(x, y)\n",
    "print ('z is ', z)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6df6a32",
   "metadata": {},
   "source": [
    "# array indexes\n",
    "\n",
    "[docs](https://mxnet.apache.org/versions/master/api/python/docs/api/legacy/ndarray/contrib/index.html#mxnet.ndarray.contrib.index_array)\n",
    "\n",
    "```\n",
    "index_array(data=None, \n",
    "            axes=_Null, \n",
    "            out=None, \n",
    "            name=None, \n",
    "            **kwargs)\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "763a98a3",
   "metadata": {},
   "source": [
    "#### index_array ex 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "9f8145c4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[[0 0]\n",
       "  [0 1]]\n",
       "\n",
       " [[1 0]\n",
       "  [1 1]]\n",
       "\n",
       " [[2 0]\n",
       "  [2 1]]]\n",
       "<NDArray 3x2x2 @cpu(0)>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = mx.nd.ones((3, 2))\n",
    "\n",
    "mx.nd.contrib.index_array(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9109a632",
   "metadata": {},
   "source": [
    "#### index_array ex 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "8ea2bcc9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[[[0 0]\n",
       "   [0 0]]\n",
       "\n",
       "  [[1 0]\n",
       "   [1 0]]]\n",
       "\n",
       "\n",
       " [[[0 1]\n",
       "   [0 1]]\n",
       "\n",
       "  [[1 1]\n",
       "   [1 1]]]\n",
       "\n",
       "\n",
       " [[[0 2]\n",
       "   [0 2]]\n",
       "\n",
       "  [[1 2]\n",
       "   [1 2]]]]\n",
       "<NDArray 3x2x2x2 @cpu(0)>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = mx.nd.ones((3, 2, 2))\n",
    "\n",
    "mx.nd.contrib.index_array(x, axes=(1, 0))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b16a0123",
   "metadata": {},
   "source": [
    "# index_copy\n",
    "\n",
    "[docs](https://mxnet.apache.org/versions/master/api/python/docs/api/contrib/ndarray/index.html#mxnet.contrib.ndarray.index_copy)\n",
    "\n",
    "\n",
    "```\n",
    "index_copy(old_tensor=None, \n",
    "           index_vector=None, \n",
    "           new_tensor=None, \n",
    "           out=None, \n",
    "           name=None, \n",
    "           **kwargs)\n",
    "```\n",
    "\n",
    "I've learned a little more about this api.  I've elaborated on the example below.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "33954904",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1. 2. 3.]\n",
       " [0. 0. 0.]\n",
       " [7. 8. 9.]\n",
       " [0. 0. 0.]\n",
       " [4. 5. 6.]]\n",
       "<NDArray 5x3 @cpu(0)>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = mx.nd.zeros((5,3))\n",
    "index = mx.nd.array([0,4,2])\n",
    "t = mx.nd.array([[1,2,3],[4,5,6],[7,8,9]])\n",
    "\n",
    "mx.nd.contrib.index_copy(x, index, t)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c86e86f",
   "metadata": {},
   "source": [
    "####  A better example of this api"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "37fa65f0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[0. 0. 0.]\n",
       " [0. 0. 0.]\n",
       " [0. 0. 0.]\n",
       " [0. 0. 0.]\n",
       " [0. 0. 0.]]\n",
       "<NDArray 5x3 @cpu(0)>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "template_tensor = nd.zeros([5,3])\n",
    "template_tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "7957a8fe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[0. 2. 4.]\n",
       "<NDArray 3 @cpu(0)>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "source_indices = nd.array([0,2,4])\n",
    "source_indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "45a46173",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[0. 1. 2.]\n",
       " [3. 4. 5.]\n",
       " [6. 7. 8.]]\n",
       "<NDArray 3x3 @cpu(0)>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "source_tensor = nd.array([[0,1,2],[3,4,5],[6,7,8]])\n",
    "source_tensor"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "90fd0e9a",
   "metadata": {},
   "source": [
    "This api uses the src_indices to specify which elements to copy from the source tensor to output.  The output uses\n",
    "the template tensor for unspecified values misssing from the src_indices array."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "9bb8452a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[0. 1. 2.]\n",
       " [0. 0. 0.]\n",
       " [3. 4. 5.]\n",
       " [0. 0. 0.]\n",
       " [6. 7. 8.]]\n",
       "<NDArray 5x3 @cpu(0)>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nd.contrib.index_copy(template_tensor, source_indices, source_tensor)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "87e4e4bb",
   "metadata": {},
   "source": [
    "Doing again using the named parameter form"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "e1aaeabc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[0. 1. 2.]\n",
       " [0. 0. 0.]\n",
       " [3. 4. 5.]\n",
       " [0. 0. 0.]\n",
       " [6. 7. 8.]]\n",
       "<NDArray 5x3 @cpu(0)>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out_tensor = nd.ones([5,3])*9 # testing the out parameter with all nines\n",
    "\n",
    "nd.contrib.index_copy(old_tensor=template_tensor,  # old_tensor is confusing  \n",
    "           index_vector=source_indices,            # index_vector is not as confusing, but not clear \n",
    "           new_tensor=source_tensor,               # new_tensor is confusing.  Its not new nor is it modifed. \n",
    "           out=out_tensor,                         # using out paramter as foo, since not sure what it is ...\n",
    "           name=None,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "75730693",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[0. 1. 2.]\n",
       " [0. 0. 0.]\n",
       " [3. 4. 5.]\n",
       " [0. 0. 0.]\n",
       " [6. 7. 8.]]\n",
       "<NDArray 5x3 @cpu(0)>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Hmm, this shows that out_tensor is the name of the output.  So, what happens is\n",
    "# all the values from the template_tensor is copied to the output, with the exception\n",
    "# of the rows specified by the index_vector.  In that case, the values in the new_tensor\n",
    "# are are copied to output\n",
    "out_tensor"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "208bde1f",
   "metadata": {},
   "source": [
    "## Testing to rotate the images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "1d45938c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "in_data is  \n",
      "[[[0. 1.]\n",
      "  [2. 3.]]]\n",
      "<NDArray 1x2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# batch 1\n",
    "in_data = nd.arange(4).reshape((1,2,2))\n",
    "print('in_data is ', in_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "612efa79",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "out_data is  \n",
      "[[[9. 9.]\n",
      "  [9. 9.]]]\n",
      "<NDArray 1x2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# setup an ouptut array with all nines, so I can see overwrites with new values\n",
    "out_data = nd.ones(4).reshape((2,2))*9\n",
    "# out_data also needs to be in batch form\n",
    "out_data = nd.expand_dims(out_data, axis=0)\n",
    "print('out_data is ', out_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "c2cabe17",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rotate_90  \n",
      "[[ 6.123234e-17 -1.000000e+00]\n",
      " [ 1.000000e+00  6.123234e-17]]\n",
      "<NDArray 2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# setup a rotation matrix\n",
    "rotate_data = rotate_90\n",
    "print('rotate_90 ', rotate_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "1f7a9abf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loop entry - a single batch - a single image in batch.data[0] from what is in in_data\n",
      "input_img.shape =  (1, 2, 2)\n",
      "indexes are:  \n",
      "[[[[0 0]\n",
      "   [0 1]]\n",
      "\n",
      "  [[1 0]\n",
      "   [1 1]]]]\n",
      "<NDArray 1x2x2x2 @cpu(0)>\n",
      "input_img_indexes.shape =  (1, 2, 2, 2)\n",
      "new_indexes  \n",
      "[[0 1]\n",
      " [1 1]\n",
      " [0 0]\n",
      " [1 0]]\n",
      "<NDArray 4x2 @cpu(0)>\n",
      "new_indexes =  \n",
      "[[ 0  0]\n",
      " [ 1  0]\n",
      " [ 0 -1]\n",
      " [ 1 -1]]\n",
      "<NDArray 4x2 @cpu(0)>\n",
      "new_indexes after shift to positive \n",
      "[[0 1]\n",
      " [1 1]\n",
      " [0 0]\n",
      " [1 0]]\n",
      "<NDArray 4x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# NDArrayIter(data, label=None, batch_size=1, shuffle=False, \n",
    "#            last_batch_handle='pad', data_name='data', \n",
    "#            label_name='softmax_label')\n",
    "#\n",
    "# Ignore the label parameter.\n",
    "dataiter = mx.io.NDArrayIter(in_data, batch_size=1, shuffle=False, last_batch_handle='discard')\n",
    "#batch_index = [0]\n",
    "for batch in dataiter:\n",
    "    print('loop entry - a single batch - a single image in batch.data[0] from what is in in_data')\n",
    "    #print ('batch.data[0] = ', batch.data[0].asnumpy())\n",
    "    #print ('batch.data[0].shape = ', batch.data[0].shape)\n",
    "    # Does this copy or get an alias to the input image?\n",
    "    input_img = batch.data[0]\n",
    "    #print ('input_img = ', input_img.asnumpy())\n",
    "    print ('input_img.shape = ', input_img.shape)\n",
    "    \n",
    "    # this will print the axis including the batch axis\n",
    "    # print('indicies are ', mx.nd.contrib.index_array(entire_img) )\n",
    "    # this will print the x,y axis and ignore the batch\n",
    "    input_img_indexes = mx.nd.contrib.index_array(input_img, axes=(1, 2))\n",
    "    print('indexes are: ', input_img_indexes)\n",
    "    print ('input_img_indexes.shape = ', input_img_indexes.shape)\n",
    "    \n",
    "    \n",
    "    # copy data from input to output\n",
    "    # Note, this seems to also use the batch axis?  Its shape (1,2,2)\n",
    "    #out_data = input_img\n",
    "    \n",
    "    # Try to assign input data to output data based upon indicies\n",
    "    orig_indexes = mx.nd.reshape(input_img_indexes, shape=(4,2))\n",
    "    print('new_indexes ', new_indexes)\n",
    "    orig_indexes = orig_indexes.astype(\"float32\")\n",
    "    \n",
    "    # I've seen one variant where matrix-matrix multiply \n",
    "    # is done with .T.  why?\n",
    "    new_indexes = nd.dot(orig_indexes, rotate_data) \n",
    "    #print('result = ', result)\n",
    "    new_indexes = new_indexes.astype('int64')\n",
    "    print('new_indexes = ', new_indexes)\n",
    "    new_indexes = new_indexes + nd.array(nd.array([0, 1])).astype('int64')\n",
    "    print('new_indexes after shift to positive', new_indexes)\n",
    "    \n",
    "    \n",
    "    \n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "5bba5b09",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "out_data is  \n",
      "[[[9. 9.]\n",
      "  [9. 9.]]]\n",
      "<NDArray 1x2x2 @cpu(0)>\n",
      "in_data is  \n",
      "[[[0. 1.]\n",
      "  [2. 3.]]]\n",
      "<NDArray 1x2x2 @cpu(0)>\n",
      "rotate_data is  \n",
      "[[ 6.123234e-17 -1.000000e+00]\n",
      " [ 1.000000e+00  6.123234e-17]]\n",
      "<NDArray 2x2 @cpu(0)>\n",
      "orig_indexes is  \n",
      "[[0. 0.]\n",
      " [0. 1.]\n",
      " [1. 0.]\n",
      " [1. 1.]]\n",
      "<NDArray 4x2 @cpu(0)>\n",
      "new_indexes is  \n",
      "[[0 1]\n",
      " [1 1]\n",
      " [0 0]\n",
      " [1 0]]\n",
      "<NDArray 4x2 @cpu(0)>\n",
      "new_indexes.dtype is  <class 'numpy.int64'>\n"
     ]
    }
   ],
   "source": [
    "print('out_data is ', out_data)\n",
    "print('in_data is ', in_data)\n",
    "print('rotate_data is ', rotate_data)\n",
    "print('orig_indexes is ', orig_indexes)\n",
    "print('new_indexes is ', new_indexes)\n",
    "print('new_indexes.dtype is ', new_indexes.dtype)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "496f68d9",
   "metadata": {},
   "source": [
    "# rewriting to match notes\n",
    "\n",
    "To avoid the batch dimension, and simply problem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "cf534ddc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "orig_image  \n",
      "[[0. 1.]\n",
      " [2. 3.]]\n",
      "<NDArray 2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "orig_image = in_data[0]\n",
    "print('orig_image ', orig_image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "c3560cf6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rotated_image \n",
      "[[9. 9.]\n",
      " [9. 9.]]\n",
      "<NDArray 2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "rotated_image=out_data[0]\n",
    "print('rotated_image', rotated_image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "638bfb41",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "orig_indexes \n",
      "[[0. 0.]\n",
      " [0. 1.]\n",
      " [1. 0.]\n",
      " [1. 1.]]\n",
      "<NDArray 4x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "print('orig_indexes', orig_indexes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "41b33c6e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "new_indexes \n",
      "[[0 1]\n",
      " [1 1]\n",
      " [0 0]\n",
      " [1 0]]\n",
      "<NDArray 4x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "print('new_indexes', new_indexes)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "383e2c45",
   "metadata": {},
   "source": [
    "#### do pixel 0,0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "03dabbae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[0.]\n",
       "<NDArray 1 @cpu(0)>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "orig_indexes[0,0]  # row 0, column 0 = the first row index is in the first row column 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "9ab4277d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[1]\n",
       "<NDArray 1 @cpu(0)>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_indexes[0,1] # row 0, column 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "912e1c2f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[1.]\n",
       "<NDArray 1 @cpu(0)>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "orig_image[new_indexes[0,0], new_indexes[0,1]]  # the new 0,0 pixel should be nine, pulled from 1,1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "58f9c135",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1. 9.]\n",
       " [9. 9.]]\n",
       "<NDArray 2x2 @cpu(0)>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Did a 1 move ot pixel 0,0?\n",
    "rotated_image[orig_indexes[0,0], orig_indexes[0,1]] = orig_image[new_indexes[0,0], new_indexes[0,1]]\n",
    "rotated_image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "154c2f6c",
   "metadata": {},
   "source": [
    "#### do pixel 0,1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "fddc5377",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1. 3.]\n",
       " [9. 9.]]\n",
       "<NDArray 2x2 @cpu(0)>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rotated_image[orig_indexes[1,0], orig_indexes[1,1]] = orig_image[new_indexes[1,0], new_indexes[1,1]]\n",
    "rotated_image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1daab96",
   "metadata": {},
   "source": [
    "#### do pixel 1,0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "a943c4f1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1. 3.]\n",
       " [0. 9.]]\n",
       "<NDArray 2x2 @cpu(0)>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rotated_image[orig_indexes[2,0], orig_indexes[2,1]] = orig_image[new_indexes[2,0], new_indexes[2,1]]\n",
    "rotated_image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "be098af7",
   "metadata": {},
   "source": [
    "#### do pixel 1,1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "2494928f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1. 3.]\n",
       " [0. 2.]]\n",
       "<NDArray 2x2 @cpu(0)>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rotated_image[orig_indexes[3,0], orig_indexes[3,1]] = orig_image[new_indexes[3,0], new_indexes[3,1]]\n",
    "rotated_image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c88c8001",
   "metadata": {},
   "source": [
    "# Is that rotated 90?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "ca7b7504",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "orig image \n",
      "[[0. 1.]\n",
      " [2. 3.]]\n",
      "<NDArray 2x2 @cpu(0)>\n",
      "rotated_image \n",
      "[[1. 3.]\n",
      " [0. 2.]]\n",
      "<NDArray 2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "print('orig image', orig_image)\n",
    "print('rotated_image', rotated_image)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b3fb85e",
   "metadata": {},
   "source": [
    "# Looks like it to me.  I call that success, but not DFD success"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8732307",
   "metadata": {},
   "source": [
    "## How to do this implicitly?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "62cddb43",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[[0. 1.]\n",
       "  [2. 3.]]\n",
       "\n",
       " [[2. 3.]\n",
       "  [2. 3.]]\n",
       "\n",
       " [[0. 1.]\n",
       "  [0. 1.]]\n",
       "\n",
       " [[2. 3.]\n",
       "  [0. 1.]]]\n",
       "<NDArray 4x2x2 @cpu(0)>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "orig_image[new_indexes] # nope"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c872f50c",
   "metadata": {},
   "source": [
    "# this looks promising"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "8adadae8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rotated_image \n",
      "[[9. 9.]\n",
      " [9. 9.]]\n",
      "<NDArray 2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# reset rotated_image\n",
    "# setup an ouptut array with all nines, so I can see overwrites with new values\n",
    "out_data = nd.ones(4).reshape((2,2))*9\n",
    "# out_data also needs to be in batch form\n",
    "out_data = nd.expand_dims(out_data, axis=0)\n",
    "#print('out_data is ', out_data)\n",
    "rotated_image=out_data[0]\n",
    "print('rotated_image', rotated_image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "e42c1a70",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[1. 3. 0. 2.]\n",
       "<NDArray 4 @cpu(0)>"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rotated_image = orig_image[new_indexes[:,0],new_indexes[:,1]]\n",
    "rotated_image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "47f99914",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1. 3.]\n",
       " [0. 2.]]\n",
       "<NDArray 2x2 @cpu(0)>"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rotated_image.reshape(2,2)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "660fe36b",
   "metadata": {},
   "source": [
    "# another try with a 3x3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "id": "2bb1f477",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "in_data is  \n",
      "[[[0. 1. 2.]\n",
      "  [3. 4. 5.]\n",
      "  [6. 7. 8.]]]\n",
      "<NDArray 1x3x3 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# batch 1\n",
    "in_data = nd.arange(9).reshape((1,3,3))\n",
    "print('in_data is ', in_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "id": "75c9f4ba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "out_data is  \n",
      "[[[9. 9. 9.]\n",
      "  [9. 9. 9.]\n",
      "  [9. 9. 9.]]]\n",
      "<NDArray 1x3x3 @cpu(0)>\n",
      "rotated_image \n",
      "[[9. 9. 9.]\n",
      " [9. 9. 9.]\n",
      " [9. 9. 9.]]\n",
      "<NDArray 3x3 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# setup an ouptut array with all nines, so I can see overwrites with new values\n",
    "out_data = nd.ones(9).reshape((3,3))*9\n",
    "# out_data also needs to be in batch form\n",
    "out_data = nd.expand_dims(out_data, axis=0)\n",
    "print('out_data is ', out_data)\n",
    "rotated_image=out_data[0]\n",
    "print('rotated_image', rotated_image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "id": "1afcc417",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rotate_90  \n",
      "[[ 6.123234e-17 -1.000000e+00]\n",
      " [ 1.000000e+00  6.123234e-17]]\n",
      "<NDArray 2x2 @cpu(0)>\n"
     ]
    }
   ],
   "source": [
    "# setup a rotation matrix\n",
    "rotate_data = rotate_90\n",
    "print('rotate_90 ', rotate_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "fe9b4228",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loop entry - a single batch - a single image in batch.data[0] from what is in in_data\n",
      "input_img.shape =  (1, 3, 3)\n",
      "indexes are:  \n",
      "[[[[0 0]\n",
      "   [0 1]\n",
      "   [0 2]]\n",
      "\n",
      "  [[1 0]\n",
      "   [1 1]\n",
      "   [1 2]]\n",
      "\n",
      "  [[2 0]\n",
      "   [2 1]\n",
      "   [2 2]]]]\n",
      "<NDArray 1x3x3x2 @cpu(0)>\n",
      "input_img_indexes.shape =  (1, 3, 3, 2)\n",
      "orig_indexes  \n",
      "[[0 0]\n",
      " [0 1]\n",
      " [0 2]\n",
      " [1 0]\n",
      " [1 1]\n",
      " [1 2]\n",
      " [2 0]\n",
      " [2 1]\n",
      " [2 2]]\n",
      "<NDArray 9x2 @cpu(0)>\n",
      "new_indexes =  \n",
      "[[ 0.0000000e+00  0.0000000e+00]\n",
      " [ 1.0000000e+00  6.1232343e-17]\n",
      " [ 2.0000000e+00  1.2246469e-16]\n",
      " [ 6.1232343e-17 -1.0000000e+00]\n",
      " [ 1.0000000e+00 -1.0000000e+00]\n",
      " [ 2.0000000e+00 -1.0000000e+00]\n",
      " [ 1.2246469e-16 -2.0000000e+00]\n",
      " [ 1.0000000e+00 -2.0000000e+00]\n",
      " [ 2.0000000e+00 -2.0000000e+00]]\n",
      "<NDArray 9x2 @cpu(0)>\n",
      "new_indexes =  \n",
      "[[ 0  0]\n",
      " [ 1  0]\n",
      " [ 2  0]\n",
      " [ 0 -1]\n",
      " [ 1 -1]\n",
      " [ 2 -1]\n",
      " [ 0 -2]\n",
      " [ 1 -2]\n",
      " [ 2 -2]]\n",
      "<NDArray 9x2 @cpu(0)>\n",
      "new_indexes after shift to positive \n",
      "[[0 2]\n",
      " [1 2]\n",
      " [2 2]\n",
      " [0 1]\n",
      " [1 1]\n",
      " [2 1]\n",
      " [0 0]\n",
      " [1 0]\n",
      " [2 0]]\n",
      "<NDArray 9x2 @cpu(0)>\n",
      "rotated_image  \n",
      "[[9. 9. 9.]\n",
      " [9. 9. 9.]\n",
      " [9. 9. 9.]]\n",
      "<NDArray 3x3 @cpu(0)>\n",
      "rotated_image.shape  (3, 3)\n",
      "rotated_image  "
     ]
    },
    {
     "ename": "IndexError",
     "evalue": "Traceback (most recent call last):\n  File \"../src/operator/tensor/indexing_op.cc\", line 454\nIndexError: index 2 is out of bounds for axis 0 with size 2",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m/tmp/ipykernel_169413/3004159683.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     42\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'rotated_image.shape '\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrotated_image\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     43\u001b[0m     \u001b[0mrotated_image\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0morig_image\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_indexes\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mnew_indexes\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 44\u001b[0;31m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'rotated_image '\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrotated_image\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     45\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.pyenv/versions/3.7.2/envs/devpost/lib/python3.7/site-packages/mxnet/ndarray/ndarray.py\u001b[0m in \u001b[0;36m__repr__\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    255\u001b[0m         \u001b[0;34m\"\"\"Returns a string representation of the array.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    256\u001b[0m         \u001b[0mshape_info\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m'x'\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjoin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'%d'\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 257\u001b[0;31m         return '\\n%s\\n<%s %s @%s>' % (str(self.asnumpy()),\n\u001b[0m\u001b[1;32m    258\u001b[0m                                       \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__class__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    259\u001b[0m                                       shape_info, self.ctx)\n",
      "\u001b[0;32m~/.pyenv/versions/3.7.2/envs/devpost/lib/python3.7/site-packages/mxnet/ndarray/ndarray.py\u001b[0m in \u001b[0;36masnumpy\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m   2569\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhandle\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2570\u001b[0m             \u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mctypes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata_as\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mctypes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mc_void_p\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2571\u001b[0;31m             ctypes.c_size_t(data.size)))\n\u001b[0m\u001b[1;32m   2572\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2573\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.pyenv/versions/3.7.2/envs/devpost/lib/python3.7/site-packages/mxnet/base.py\u001b[0m in \u001b[0;36mcheck_call\u001b[0;34m(ret)\u001b[0m\n\u001b[1;32m    244\u001b[0m     \"\"\"\n\u001b[1;32m    245\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mret\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 246\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mget_last_ffi_error\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    247\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    248\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mIndexError\u001b[0m: Traceback (most recent call last):\n  File \"../src/operator/tensor/indexing_op.cc\", line 454\nIndexError: index 2 is out of bounds for axis 0 with size 2"
     ]
    }
   ],
   "source": [
    "# NDArrayIter(data, label=None, batch_size=1, shuffle=False, \n",
    "#            last_batch_handle='pad', data_name='data', \n",
    "#            label_name='softmax_label')\n",
    "#\n",
    "# Ignore the label parameter.\n",
    "dataiter = mx.io.NDArrayIter(in_data, batch_size=1, shuffle=False, last_batch_handle='discard')\n",
    "#batch_index = [0]\n",
    "for batch in dataiter:\n",
    "    \n",
    "    print('loop entry - a single batch - a single image in batch.data[0] from what is in in_data')\n",
    "    #print ('batch.data[0] = ', batch.data[0].asnumpy())\n",
    "    #print ('batch.data[0].shape = ', batch.data[0].shape)\n",
    "    # Does this copy or get an alias to the input image?\n",
    "    input_img = batch.data[0]\n",
    "    #print ('input_img = ', input_img.asnumpy())\n",
    "    print ('input_img.shape = ', input_img.shape)\n",
    "    \n",
    "    # this will print the axis including the batch axis\n",
    "    # print('indicies are ', mx.nd.contrib.index_array(entire_img) )\n",
    "    # this will print the x,y axis and ignore the batch\n",
    "    input_img_indexes = mx.nd.contrib.index_array(input_img, axes=(1, 2))\n",
    "    print('indexes are: ', input_img_indexes)\n",
    "    print ('input_img_indexes.shape = ', input_img_indexes.shape)\n",
    "    \n",
    "    \n",
    "    \n",
    "    # Try to assign input data to output data based upon indicies\n",
    "    orig_indexes = mx.nd.reshape(input_img_indexes, shape=(9,2))\n",
    "    print('orig_indexes ', orig_indexes)\n",
    "    orig_indexes = orig_indexes.astype(\"float32\")\n",
    "    \n",
    "    # I've seen one variant where matrix-matrix multiply \n",
    "    # is done with .T.  why?\n",
    "    new_indexes = nd.dot(orig_indexes, rotate_data) \n",
    "    print('new_indexes = ', new_indexes)\n",
    "    new_indexes = new_indexes.astype('int64')\n",
    "    print('new_indexes = ', new_indexes)\n",
    "    new_indexes = new_indexes + nd.array(nd.array([0, 2])).astype('int64')\n",
    "    print('new_indexes after shift to positive', new_indexes)\n",
    "\n",
    "    print('rotated_image ', rotated_image)\n",
    "    print('rotated_image.shape ', rotated_image.shape)\n",
    "    rotated_image = orig_image[new_indexes[:,0],new_indexes[:,1]]\n",
    "    print('rotated_image ', rotated_image)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9218bee4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "54df567c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "551a896f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5f3c7814",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0c130eca",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "805b7f87",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "38d384fc",
   "metadata": {},
   "source": [
    "### Hmm, try this contrib api again"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "65ff1888",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[1. 2. 3.]\n",
       " [0. 0. 0.]\n",
       " [7. 8. 9.]\n",
       " [0. 0. 0.]\n",
       " [4. 5. 6.]]\n",
       "<NDArray 5x3 @cpu(0)>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# for reference\n",
    "x = mx.nd.zeros((5,3))\n",
    "t = mx.nd.array([[1,2,3],[4,5,6],[7,8,9]])\n",
    "index = mx.nd.array([0,4,2])  # <--- this is interesting. It has a zero length axes?????\n",
    "# x is a template for output\n",
    "# index is the indexes?\n",
    "# t is \"new_tensor?\", its the source tensor?\n",
    "mx.nd.contrib.index_copy(x, index, t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "25a70eff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[0. 4. 2.]\n",
       "<NDArray 3 @cpu(0)>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# hmm, the t has three axis specification. perhaps it requires a batch parameter?\n",
    "index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "3cc2cdde",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[0. 0. 0.]\n",
       " [0. 0. 0.]\n",
       " [0. 0. 0.]\n",
       " [0. 0. 0.]\n",
       " [0. 0. 0.]]\n",
       "<NDArray 5x3 @cpu(0)>"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Did x change? No, it is just a template for output array?\n",
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "414eb26a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[[0. 0.]\n",
       "  [0. 0.]]]\n",
       "<NDArray 1x2x2 @cpu(0)>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "template_img = nd.zeros([2,2])\n",
    "template_img_batch = nd.expand_dims(template_img, axis=0)\n",
    "template_img_batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "6d3bec2d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[0 1]\n",
       " [1 1]\n",
       " [0 0]\n",
       " [1 0]]\n",
       "<NDArray 4x2 @cpu(0)>"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# this is the wrong shape.  It gives an error. possibly due to the lack of batch\n",
    "new_indexes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "abe553b9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[[0 1]\n",
       "  [1 1]\n",
       "  [0 0]\n",
       "  [1 0]]]\n",
       "<NDArray 1x4x2 @cpu(0)>"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_indexes_batch = nd.expand_dims(new_indexes, axis=0)\n",
    "new_indexes_batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "9ccfb3bb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "[[[0. 1.]\n",
       "  [2. 3.]]]\n",
       "<NDArray 1x2x2 @cpu(0)>"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# perhaps the orignal image also needs a batch?\n",
    "orig_image_batch = nd.expand_dims(orig_image, axis=0)\n",
    "orig_image_batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "e16ad5fb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "orig_image_batch.shape =  (1, 2, 2)\n",
      "new_indexes_batch.shape =  (1, 4, 2)\n",
      "template_img_batch.shape =  (1, 2, 2)\n"
     ]
    }
   ],
   "source": [
    "# Hmm. adding back batch dims? Or running in circles of confusion?\n",
    "print('orig_image_batch.shape = ', orig_image_batch.shape )\n",
    "print('new_indexes_batch.shape = ', new_indexes_batch.shape )\n",
    "print('template_img_batch.shape = ', template_img_batch.shape )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "f5e6d856",
   "metadata": {},
   "outputs": [
    {
     "ename": "MXNetError",
     "evalue": "Traceback (most recent call last):\n  File \"../src/operator/contrib/./index_copy-inl.h\", line 64\nMXNetError: Check failed: in_attrs->at(1).ndim() == 1 (3 vs. 1) : ",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mMXNetError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m/tmp/ipykernel_169413/1642385947.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;31m# * template img is not batched, but it is three columns?\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;31m# *\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m \u001b[0mmx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcontrib\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex_copy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtemplate_img\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_indexes_batch\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0morig_image_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/.pyenv/versions/3.7.2/envs/devpost/lib/python3.7/site-packages/mxnet/ndarray/register.py\u001b[0m in \u001b[0;36mindex_copy\u001b[0;34m(old_tensor, index_vector, new_tensor, out, name, **kwargs)\u001b[0m\n",
      "\u001b[0;32m~/.pyenv/versions/3.7.2/envs/devpost/lib/python3.7/site-packages/mxnet/_ctypes/ndarray.py\u001b[0m in \u001b[0;36m_imperative_invoke\u001b[0;34m(handle, ndargs, keys, vals, out, is_np_op, output_is_list)\u001b[0m\n\u001b[1;32m     89\u001b[0m         \u001b[0mc_str_array\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkeys\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     90\u001b[0m         \u001b[0mc_str_array\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mvals\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 91\u001b[0;31m         ctypes.byref(out_stypes)))\n\u001b[0m\u001b[1;32m     92\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m     \u001b[0mcreate_ndarray_fn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_global_var\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_np_ndarray_cls\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mis_np_op\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0m_global_var\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_ndarray_cls\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.pyenv/versions/3.7.2/envs/devpost/lib/python3.7/site-packages/mxnet/base.py\u001b[0m in \u001b[0;36mcheck_call\u001b[0;34m(ret)\u001b[0m\n\u001b[1;32m    244\u001b[0m     \"\"\"\n\u001b[1;32m    245\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mret\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 246\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mget_last_ffi_error\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    247\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    248\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mMXNetError\u001b[0m: Traceback (most recent call last):\n  File \"../src/operator/contrib/./index_copy-inl.h\", line 64\nMXNetError: Check failed: in_attrs->at(1).ndim() == 1 (3 vs. 1) : "
     ]
    }
   ],
   "source": [
    "# mx.nd.contrib.index_copy(template_img, new_indexes, orig_image_batch)\n",
    "# mx.nd.contrib.index_copy(template_img, new_indexes_batch, orig_image_batch)\n",
    "#mx.nd.contrib.index_copy(template_img_batch, new_indexes_batch, orig_image_batch)\n",
    "# best guess at the api, would be\n",
    "# * template img is not batched, but it is three columns?\n",
    "# * \n",
    "mx.nd.contrib.index_copy(template_img, new_indexes_batch, orig_image_batch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f668a555",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a301b27d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "729293a6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "892900ff",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b9dbf486",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c2d48a9e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1978caeb",
   "metadata": {},
   "outputs": [],
   "source": [
    "orig_indexes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c8eda68e",
   "metadata": {},
   "outputs": [],
   "source": [
    "in_data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "31c9f3b3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ddd947df",
   "metadata": {},
   "outputs": [],
   "source": [
    "in_data[0,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d96f0ca8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c0585a5f",
   "metadata": {},
   "outputs": [],
   "source": [
    "orig_indexes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7625abf7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d3548a6a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "66a8b455",
   "metadata": {},
   "outputs": [],
   "source": [
    "a = np.arange(50).reshape(10, 5) # Array to be indexed\n",
    "# b = 9 7 5 3 1  which means start with row 9 and step down to \n",
    "b = np.arange(9, -1, -2) # Indexing array. Start at 9, step down by 2 each time until 1\n",
    "print('a = ', a)\n",
    "print('b (indexing) = ', b)\n",
    "result = a[b, np.arange(a.shape[1])]\n",
    "print('result = ', result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e8345de1",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[0] = in_data[0,new_indexes]\n",
    "print('out_data is ', out_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2282d0cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[0,new_indexes] = in_data\n",
    "print('out_data is ', out_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "03eb008b",
   "metadata": {},
   "outputs": [],
   "source": [
    "a = nd.arange(50).reshape(10, 5) # Array to be indexed\n",
    "b = nd.arange(9, -1, -2) # Indexing array\n",
    "print('a = ', a)\n",
    "print('b (indexing) = ', b)\n",
    "result = a[b, nd.arange(a.shape[1])]\n",
    "print('result = ', result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0d3dece0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3f66da2d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "85546e1d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0f034c1a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "42d477f1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e8fd3834",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ebe000c5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1fb27374",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b438688e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93bb4234",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[new_indexes]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "14fc5f5b",
   "metadata": {},
   "outputs": [],
   "source": [
    "new_indexes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc9e7f10",
   "metadata": {},
   "outputs": [],
   "source": [
    "new_indexes[:,0], new_indexes[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4903b505",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[ 0, new_indexes[:,0]: , new_indexes[:,1]: ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "816e8044",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[0, 0-2:, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e90a1764",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[0, np.array([0,1,2]):, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8760f361",
   "metadata": {},
   "outputs": [],
   "source": [
    "# x = mx.nd.zeros((5,3))\n",
    "# t = mx.nd.array([[1,2,3],[4,5,6],[7,8,9]])\n",
    "# index = mx.nd.array([0,4,2])\n",
    "\n",
    "mx.nd.contrib.index_copy(out_data[0], new_indexes, in_data[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d473751e",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3805692c",
   "metadata": {},
   "outputs": [],
   "source": [
    "in_data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1f9cf38c",
   "metadata": {},
   "outputs": [],
   "source": [
    "new_indexes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e6a14d21",
   "metadata": {},
   "outputs": [],
   "source": [
    "out_data[0,new_indexes[0,0],0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ed7b6b04",
   "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}