{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Misc notes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Barry's data is mostly from  from 3 deg to 8 deg of eccentricity; most cells around 5 deg eccentricity. There is a nonlinear relationships between retinal eccentricity (in $mm$) and visual angle (in $deg$) in macaque, see see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC50193/. A polynomial fit is given by\n",
    "\\begin{eqnarray}\n",
    "A = a + b E + c E^2\n",
    "\\end{eqnarray}\n",
    "with, $a = 0.1$, $b = 4.21$, and $c = 0.038$. Here A is eccentricity in $deg$, and E is eccentricity in $mm$. Using this equation, the cells are situated between $0.7$mm to $1.8$mm, mostly around $1.2$mm.\n",
    "\n",
    "Taking the derivative of the quation, one gets the magnification factor\n",
    "\\begin{eqnarray}\n",
    "\\frac{dA}{dE} = b + 2cE \\qquad\\text{ in deg/mm}\n",
    "\\end{eqnarray}\n",
    "Evaluating this at $E=1.2\\,[0.7,\\,1.8]$mm, the range of manification factors is $4.30\\,[4.26,\\,4.35]$deg/mm. Inverting yields \n",
    "\\begin{eqnarray}\n",
    "    s = 0.232 [0.235,0.230] \\text{mm/deg}\n",
    "\\end{eqnarray}\n",
    "\n",
    "The field of view is 256 px and 4.6degree across. Therefore\n",
    "* 256px are 4.6deg $\\to$ 56px per 1deg\n",
    "* 256px are $s*4.6\\text{deg}=1.07[1.06, 1.08]$mm $\\to$ 48px are 0.2mm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import cv2 as cv2\n",
    "import pylab as pl\n",
    "from matplotlib import cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the movie to show in comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "frame_raw = np.ones((3,256,256,4500))\n",
    "\n",
    "vidcap = cv2.VideoCapture('../../1x10_256.mpg')\n",
    "success, image = vidcap.read()\n",
    "counter = 0\n",
    "success = True\n",
    "while success and (counter < 4500):   # 4500 to only process first 2.5 min *60sec *30fps in the mpg -> 30sec ret video\n",
    "    success, image = vidcap.read()\n",
    "    \n",
    "    #Raw RGB values\n",
    "    im_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n",
    "\n",
    "    red   = np.double(im_rgb[:,:,0])\n",
    "    green = np.double(im_rgb[:,:,1])\n",
    "    blue  = np.double(im_rgb[:,:,2])\n",
    "    \n",
    "    #Save luminance frame\n",
    "    frame_raw[:,:,:,counter] = np.array([red/256., green/256., blue/256.])\n",
    "    counter += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Parse the npy file to show model prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_frames = np.load(\"./Gon-19829503-frames.npy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "startframe = 600 #1sec before video begins at 150fps\n",
    "\n",
    "for idx in range(startframe,4500):\n",
    "    \n",
    "\n",
    "    cell_frame = filtered_frames[:,:,idx]\n",
    "    cell_frame[230:240,180:(180+48)] = 10  # 0.2 mm\n",
    "\n",
    "    fig, axes = pl.subplots(nrows=1, ncols=2, figsize=(7, 4))\n",
    "    \n",
    "    r = np.reshape(frame_raw[0,:,:,idx],(256, 256))\n",
    "    g = np.reshape(frame_raw[1,:,:,idx],(256, 256))\n",
    "    b = np.reshape(frame_raw[2,:,:,idx],(256, 256))\n",
    "\n",
    "    img = np.zeros((256,256,3))\n",
    "    img[:,:,0] = np.array(r, dtype = np.float)\n",
    "    img[:,:,1] = np.array(g, dtype = np.float)\n",
    "    img[:,:,2] = np.array(b, dtype = np.float)\n",
    "\n",
    "    img[230:240,180:(180+56),0] = 1 # 1 deg scale bar\n",
    "    img[230:240,180:(180+56),1] = 1\n",
    "    img[230:240,180:(180+56),2] = 1\n",
    "    \n",
    "    axes[0].imshow(img)\n",
    "    axes[0].set_title(\"Video\")\n",
    "    axes[0].axes.xaxis.set_ticklabels([])\n",
    "    axes[0].axes.yaxis.set_ticklabels([])\n",
    "    axes[0].text(x=182,y=220, s = \"1 deg\", fontsize=15, color = 'w')\n",
    "\n",
    "    axes[1].imshow(cell_frame, vmin=0, vmax=1, cmap=cm.Greys_r)\n",
    "    axes[1].set_title(\"PC-Gon-cell\")\n",
    "    axes[1].axes.xaxis.set_ticklabels([])\n",
    "    axes[1].axes.yaxis.set_ticklabels([])\n",
    "    axes[1].text(x=20,y=30, s = \"time = \"+str(np.floor((idx-startframe)/150*1000)/1000)+\"s\", fontsize=15,color = 'w')\n",
    "    axes[1].text(x=170,y=220, s = \"0.2 mm\", fontsize=15, color = 'w')\n",
    "\n",
    "    fig.tight_layout()\n",
    "\n",
    "    pl.savefig(\"./video_frames/file%04d.png\" % (idx-startframe))\n",
    "    pl.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Make into movie:\n",
    "\n",
    "ffmpeg -r 60 -f image2 -s 504x288 -i ./video_frames/file%04d.png -vcodec libx264 -crf 5  -pix_fmt yuv420p simulation.mp4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}