flows_solver.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "using JuMP, Cbc, PyPlot, Images, FileIO, Statistics"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Flows Solver\n",
    "\n",
    "_Last updated: 29 Jul 19_\n",
    "\n",
    "Solver for mobile game Flows. Currently written and checked using Julia 1.0. If you see a problem tell me on GitHub or message me somehow.\n",
    "\n",
    "For those who are unaware, Flows free is a puzzle game where you are given a grid with coloured dots, and you are trying to draw a path connecting the same coloured dots together while not crossing lines and filling all the grid. \n",
    "\n",
    "<img src=\"test1.png\" width=\"20%\">\n",
    "\n",
    "If you have the screenshot of the game itself, can directly feed filepath into `flows_solver` and get the answer.\n",
    "\n",
    "If you have a similar game which is not Flows, you can enter the problem directly into `find_answer` in a matrix form, like \n",
    "```\n",
    "[1 0 0 0 1,\n",
    " 2 0 0 0 3,\n",
    " 0 0 4 0 4,\n",
    " 0 0 0 0 0,\n",
    " 2 5 0 5 3]\n",
    "```\n",
    "\n",
    "The program below is for solving the puzzle and drawing a completed solution. Run all code blocks below then play around at the bottom of page. \n",
    "\n",
    "The solver tries to find solution which fills the entire board. If your solution comes out weird, the problem given might have no solution that fills the entire board (source: experimenting).\n",
    "\n",
    "Due to a bit of code which involves random number generator, code sometimes break. Happens when youre unlucky enough. Run again and it should be fine (Reason: randomising the place where the grid lines are being detected and sometimes it randomises into bad places).\n",
    "\n",
    "Code can also take a bit long (some of the larger grids can take up to half an hour). Not sure if it's the solver (Gurobi takes a bit shorter? But it's not free so I didn't use it here) or that there's something with the code that takes a long time, or nothing can be done. Let me know if you can fix it or something."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Functions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The function below takes in the problem as a 2-D matrix form (generated from other functions below) and defines the problem mathematically and solves it. This problem is a integer linear program, which is formulated as shown below.\n",
    "\n",
    "### Variables\n",
    "\n",
    "Suppose we have a problem set which is of size $p \\times q$ and there are $c$ different colours.\n",
    "\n",
    "There are two main sets of variables - one for edges, and one for nodes (which are the boxes in the grid). \n",
    "\n",
    "A particular node can be specified by two indices $1 \\leq i \\leq p$ and $1 \\leq j \\leq q$ which tells which row and column the node is in. The variables for the nodes are set to be $a_{ijk}$, where $a_{ijk} = 1$ if node at row $i$ and column $j$ is of colour $k$ (where $1 \\leq k \\leq c$ is the number to represent one of the colours) and is 0 otherwise. They are only allowed to be either 0 or 1. \n",
    "\n",
    "For the edges, we divide them into two sets - edges which runs horizontally, and edges which runs vertically. For the edges which runs horizontally, we assign them two indices $1 \\leq i \\leq p$ and $1 \\leq j \\leq q$ such that the vertex connected to the left of the edge has the same indices as the edge, i.e. an edge connecting nodes (1, 2) and (1, 3) will be given index (1, 2). Variables used for horizontal edges are assigned as $h_{ijk}$, where $h_{ijk} = 1$ if edge ($i$, $j$) is painted colour $k$, and is 0 otherwise. Vertical edges can be defined in a similar manner but instead using variables $v_{ijk}$. Again, both $h_{ijk}$ and $v_{ijk}$ can only be either 0 or 1.\n",
    "\n",
    "### Constraints\n",
    "\n",
    "For notation purposes, we let $[1,x]$ represent a range of number, i.e. $ \\{ 1, 2, ..., x \\}$. \n",
    "\n",
    "The list of constraints are as follows.\n",
    "\n",
    "- An edge can only be painted at most one colour (they can also not be painted as well).\n",
    "\n",
    "$$ \\sum_{k} v_{ijk} \\leq 1 \\quad \\forall i \\in [1,p], \\forall j \\in [1,q], \\forall k \\in [1,c]$$\n",
    "\n",
    "$$ \\sum_{k} h_{ijk} \\leq 1 \\quad \\forall i \\in [1,p], \\forall j \\in [1,q], \\forall k \\in [1,c]$$\n",
    "\n",
    "- A node must be painted exactly one colour (as the game specifies that all of the boxes have a line dragged through).\n",
    "\n",
    "$$ \\sum_{k} a_{ijk} = 1 \\quad \\forall i \\in [1,p], \\forall j \\in [1,q], \\forall k \\in [1,c]$$\n",
    "\n",
    "- An edge can only be painted a certain colour if the two nodes connected to that edge are both that colour (however if two nodes adjacent to each other are of the same colour they might not have an edge connected between it).\n",
    "\n",
    "$$ 2 v_{ijk} \\leq a_{ijk} + a_{(i+1)jk} \\quad \\forall i \\in [1,p], \\forall j \\in [1,q], \\forall k \\in [1,c] $$\n",
    "\n",
    "$$ 2 h_{ijk} \\leq a_{ijk} + a_{i(j+1)k} \\quad \\forall i \\in [1,p], \\forall j \\in [1,q], \\forall k \\in [1,c] $$\n",
    "\n",
    "- If a node has no fixed colour, when it is assigned a colour (i.e. a line is drawn through the node), two of the edges connected to it must be painted that colour.\n",
    "\n",
    "$$ 2 * a_{ijk} = h_{ijk} + h_{(i-1)jk} + v_{ijk} + v_{i(j-1)k} \\quad \\forall i \\in [1,p], \\forall j \\in [1,q], \\forall k \\in [1,c] $$\n",
    "\n",
    "- If a node has a fixed colour (one of the given starting points), then the node has a specified colour. Additionally, only one edge leaving the node can be painted that colour. If the starting node is painted colour $k'$, then the constraints are given as\n",
    "\n",
    "$$ a_{ijk'} = 1,  \\quad h_{ijk'} + h_{(i-1)jk'} + v_{ijk'} + v_{i(j-1)k'} = 1, \\quad \\forall i \\in [1,p], \\forall j \\in [1,q] $$\n",
    "\n",
    "Note that in the last three constraints there may be cases where the index is out of range (i.e. when dealing with the edge of the puzzle grid). This can be fixed by adding \"ghost\" edges and \"ghost\" nodes which are all set to zero, but will also be a part of the summation so the computer stays happy.\n",
    "\n",
    "### Objective\n",
    "\n",
    "This is chosen arbitrarily as this problem only requires a solution, so any feasible answer will be fine. Additionally, there is no suitable measurement of which answer is better than another (as often there will only be one answer to a problem anyway). In case that there are more than one answer in a case, we use the objective as to maximise $\\sum_{i, j, k} a_{ijk}$ so that we ensure as much of the grid gets filled - ideally all of it gets filled (I'm 90% sure it doesn't actually matter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "find_answer (generic function with 1 method)"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# input: array with all the given dots (each number represents the colours) and number of colours\n",
    "# output: solved flows problem\n",
    "function find_answer(array)\n",
    "    \n",
    "    m = Model(with_optimizer(Cbc.Optimizer, logLevel=0)) \n",
    "    xsize, ysize = size(array)\n",
    "    n = Int8(findmax(array)[1])\n",
    "    \n",
    "    # variable to control the colours of each blocks (and constraints)\n",
    "    @variable(m, ans[0:xsize+1,0:ysize+1,1:n], Bin)\n",
    "    @constraint(m, [i=0:xsize+1,k=1:n], ans[i,0,k] == 0)\n",
    "    @constraint(m, [i=0:xsize+1,k=1:n], ans[i,ysize+1,k] == 0)\n",
    "    @constraint(m, [j=0:ysize+1,k=1:n], ans[0,j,k] == 0)\n",
    "    @constraint(m, [j=0:ysize+1,k=1:n], ans[xsize+1,j,k] == 0)\n",
    "    \n",
    "    # only one colour in each square\n",
    "    @constraint(m, [i=1:xsize, j=1:ysize], sum(ans[i,j,:]) == 1)\n",
    "    \n",
    "    # variable for edges between each nodes (one variable for edge between two points and which colour)\n",
    "    @variable(m, hedges[0:xsize,0:ysize+1,1:n], Bin)\n",
    "    @variable(m, vedges[0:xsize+1,0:ysize,1:n], Bin)\n",
    "    @constraint(m, [j=0:ysize+1,k=1:n], hedges[0,j,k] == 0)\n",
    "    @constraint(m, [j=0:ysize+1,k=1:n], hedges[xsize,j,k] == 0)\n",
    "    @constraint(m, [i=0:xsize+1,k=1:n], vedges[i,0,k] == 0)\n",
    "    @constraint(m, [i=0:xsize+1,k=1:n], vedges[i,ysize,k] == 0)\n",
    "    \n",
    "    # each edges can only be there if colours are the same\n",
    "    @constraint(m, [i=0:xsize,j=0:ysize+1,k=1:n], ans[i,j,k] + ans[i+1,j,k] >= 2*hedges[i,j,k])\n",
    "    @constraint(m, [i=0:xsize+1,j=0:ysize,k=1:n], ans[i,j,k] + ans[i,j+1,k] >= 2*vedges[i,j,k])\n",
    "    \n",
    "    # at most one of the edges is filled for each colour\n",
    "    @constraint(m, [i=1:xsize-1,j=1:ysize], sum(hedges[i,j,:]) <= 1)\n",
    "    @constraint(m, [i=1:xsize,j=1:ysize-1], sum(vedges[i,j,:]) <= 1)\n",
    "    \n",
    "    # fix the given points in place\n",
    "    # and fix the number of edges through each of the points\n",
    "    for i in 1:xsize\n",
    "        for j in 1:ysize\n",
    "            if array[i,j] != 0\n",
    "                co = array[i,j]\n",
    "                @constraint(m, ans[i,j,co] == 1)\n",
    "                @constraint(m, hedges[i-1,j,co] + hedges[i,j,co] + vedges[i,j-1,co] + vedges[i,j,co] == 1)\n",
    "            else\n",
    "                @constraint(m, [k=1:n], hedges[i-1,j,k]+hedges[i,j,k]+vedges[i,j-1,k]+vedges[i,j,k] == 2*ans[i,j,k])\n",
    "            end\n",
    "        end\n",
    "    end\n",
    "    \n",
    "    @objective(m, Max, sum(ans))\n",
    "    optimize!(m)\n",
    "    \n",
    "    ans_array = JuMP.value.(ans)\n",
    "    hedges_ans = JuMP.value.(hedges)\n",
    "    vedges_ans = JuMP.value.(vedges)\n",
    "    final_answer = zeros(Int8, xsize, ysize)\n",
    "    for x in 1:xsize\n",
    "        for y in 1:ysize\n",
    "            for c in 1:n\n",
    "                if ans_array[x,y,c] == 1\n",
    "                    final_answer[x,y] = c\n",
    "                end\n",
    "            end\n",
    "        end\n",
    "    end\n",
    "    \n",
    "    return final_answer, hedges_ans, vedges_ans\n",
    "    \n",
    "end"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The function below are to detect the board from a screenshot of the game using image processing packages. It detects the grid lines to find size of the puzzle. It returns the coordinates of the rows and columns of game board.\n",
    "\n",
    "Note that they will sometimes run into a problem, which may require another run of the same code (as it uses a random number generator in there - future fix?)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "detect_board (generic function with 2 methods)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "function detect_board(img,thr=0.1)\n",
    "    \n",
    "    grad_y, grad_x, mag, orient = imedge(img)\n",
    "    sy, sx = size(img)\n",
    "    \n",
    "    edges_mat = Array(channelview(mag));\n",
    "    \n",
    "    # randomly scan one of the columns, and find where the edges are\n",
    "    # the boundaries should have edges at a constant place each time, so select those places\n",
    "    # scan a couple of times\n",
    "    row_loop = 0\n",
    "    rows = Set(1:sy)\n",
    "    while row_loop < 5\n",
    "        \n",
    "        test_col = rand(floor(Int,sx*0.2):floor(Int,sx*0.8))\n",
    "        # if overall edge magnitude here too high then we know that we are on top of a line, so change sample\n",
    "        if mean(edges_mat[:,test_col]) >= 0.2\n",
    "            continue\n",
    "        end\n",
    "        \n",
    "        pot_rows = []\n",
    "        rising = false\n",
    "        for i in 1:sy\n",
    "            if edges_mat[i,test_col] > thr && !rising && ((length(pot_rows) == 0) || (i - last(pot_rows) > 4))\n",
    "                push!(pot_rows,i)\n",
    "                rising = true\n",
    "            elseif edges_mat[i,test_col] <= thr && rising\n",
    "                rising = false\n",
    "            end\n",
    "        end\n",
    "        rows = intersect(rows,pot_rows)\n",
    "        \n",
    "        row_loop += 1\n",
    "    end\n",
    "    rows = sort(collect(rows))\n",
    "    \n",
    "    # detect the position of columns given we know the rows\n",
    "    cols = Set(1:sx)\n",
    "    for i in 1:length(rows)-1\n",
    "        test_row = floor(Int, (rows[i] + rows[i+1])/2)\n",
    "        pot_cols = []\n",
    "        rising = false\n",
    "        for i in 1:sx\n",
    "            if edges_mat[test_row,i] > thr && !rising && ((length(pot_cols) == 0) || (i - last(pot_cols) > 4))\n",
    "                push!(pot_cols,i)\n",
    "                rising = true\n",
    "            elseif edges_mat[test_row,i] <= thr && rising\n",
    "                rising = false\n",
    "            end\n",
    "        end\n",
    "        cols = intersect(cols,pot_cols)\n",
    "    end\n",
    "    cols = sort(collect(cols))\n",
    "    \n",
    "    return rows, cols\n",
    "    \n",
    "end"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This function converts problem into a matrix form, given grid locations from `detect_board`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "get_problem (generic function with 1 method)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Input: two colours in XYZ (closely related to human perception colours)\n",
    "# Output: similarity index for colours\n",
    "function colour_similar(c1, c2, thr=0.05)\n",
    "    return sqrt((c1.x - c2.x)^2 + (c1.y - c2.y)^2 + (c1.z - c2.z)^2) < 0.05\n",
    "end\n",
    "\n",
    "# Input: image and colour similar threshold (arbitrary)\n",
    "# Output: array of the problem, colours of each pair of starting points, and size of problem\n",
    "function get_problem(img)\n",
    "    \n",
    "    rows, cols = detect_board(img)\n",
    "    xsize = length(rows)-1\n",
    "    ysize = length(cols)-1\n",
    "    problem = zeros(Int, (xsize,ysize))\n",
    "    colours = []\n",
    "    colour_count = 0\n",
    "    \n",
    "    for x in 1:xsize\n",
    "        for y in 1:ysize\n",
    "            xco = floor(Int, (rows[x] + rows[x+1])/2)\n",
    "            yco = floor(Int, (cols[y] + cols[y+1])/2)\n",
    "            c = img[xco,yco]\n",
    "            if convert(HSV,c).v > 0.2  # not part of black background\n",
    "                c = convert(XYZ,c)\n",
    "                new_colour = true\n",
    "                for i in 1:colour_count\n",
    "                    if colour_similar(c, colours[i])\n",
    "                        new_colour = false\n",
    "                        problem[x,y] = i\n",
    "                    end\n",
    "                end\n",
    "                if new_colour\n",
    "                    push!(colours, c)\n",
    "                    colour_count += 1\n",
    "                    problem[x,y] = colour_count\n",
    "                end\n",
    "            end\n",
    "        end\n",
    "    end\n",
    "    \n",
    "    convert(Array{ColorTypes.RGB4{FixedPointNumbers.Normed{UInt8,8}},1},colours)\n",
    "    \n",
    "    return problem, colours, rows, cols\n",
    "    \n",
    "end"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This function takes the solution from `find_answer` and draws it out on top of a given image."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "draw_solution (generic function with 3 methods)"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "function draw_solution(img, img_filename, rows, cols, board_ans, hedges_ans, vedges_ans, colours)\n",
    "    \n",
    "    fig = figure()\n",
    "    ax = fig.add_subplot(111)\n",
    "    \n",
    "    imgplot = ax.imshow(imread(img_filename))\n",
    "    thickness = (size(img)[2]) * 0.07 / (length(cols)-1)\n",
    "    \n",
    "    xcos = [floor(Int, (rows[i+1] + rows[i])/2) for i in 1:length(rows)-1]\n",
    "    ycos = [floor(Int, (cols[i+1] + cols[i])/2) for i in 1:length(cols)-1]\n",
    "    \n",
    "    for x in 1:length(xcos)\n",
    "        for y in 1:length(ycos)\n",
    "            ci = board_ans[x,y]\n",
    "            c = convert(RGB, colours[ci])\n",
    "            if (x+1 <= length(xcos)) && hedges_ans[x,y,ci] == 1\n",
    "                ax.plot([ycos[y],ycos[y]],[xcos[x],xcos[x+1]],linewidth=thickness,color=(c.r,c.g,c.b))\n",
    "            end\n",
    "            if (y+1 <= length(ycos)) && vedges_ans[x,y,ci] == 1\n",
    "                ax.plot([ycos[y],ycos[y+1]],[xcos[x],xcos[x]],linewidth=thickness,color=(c.r,c.g,c.b))\n",
    "            end\n",
    "        end\n",
    "    end\n",
    "    \n",
    "    return fig\n",
    "    \n",
    "end"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is the function which is a combination of everything above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "flows_solver (generic function with 1 method)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Input: filename\n",
    "# Output: the solution\n",
    "function flows_solver(img_filename)\n",
    "    \n",
    "    img = load(img_filename)\n",
    "    \n",
    "    prob, colours, rows, cols = get_problem(img)\n",
    "    \n",
    "    board_ans, h_ans, v_ans = find_answer(prob)\n",
    "    \n",
    "    imgplot = draw_solution(img, img_filename, rows, cols, board_ans, h_ans, v_ans, colours)\n",
    "    \n",
    "    return imgplot\n",
    "    \n",
    "end"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Demo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "Figure(PyObject <Figure size 640x480 with 1 Axes>)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "PyObject <matplotlib.image.AxesImage object at 0x150f5e470>"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img_filename = \"test4.png\"\n",
    "imshow(imread(img_filename), interpolation=\"nearest\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "Figure(PyObject <Figure size 640x480 with 1 Axes>)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1721.324444 seconds (1.76 M allocations: 189.018 MiB, 0.02% gc time)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "Figure(PyObject <Figure size 640x480 with 1 Axes>)"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "@time flows_solver(img_filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  0.055511 seconds (172.60 k allocations: 6.863 MiB)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "6×5 Array{Int8,2}:\n",
       " 1  1  1  1  1\n",
       " 2  3  3  3  3\n",
       " 2  3  4  4  4\n",
       " 2  3  3  3  3\n",
       " 2  5  5  5  3\n",
       " 6  6  6  6  6"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "@time board_ans, h_ans, v_ans = find_answer([\n",
    "        1 0 0 0 1;\n",
    "        2 0 0 0 3;\n",
    "        0 0 4 0 4;\n",
    "        0 0 0 0 0;\n",
    "        2 5 0 5 3;\n",
    "        6 0 0 0 6\n",
    "        ])\n",
    "\n",
    "board_ans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3-dimensional DenseAxisArray{Float64,3,...} with index sets:\n",
       "    Dimension 1, 0:6\n",
       "    Dimension 2, 0:6\n",
       "    Dimension 3, 1:6\n",
       "And data, a 7×7×6 Array{Float64,3}:\n",
       "[:, :, 1] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 2] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  1.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  1.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  1.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 3] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  1.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  1.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  1.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 4] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 5] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 6] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0  0.0"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "h_ans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3-dimensional DenseAxisArray{Float64,3,...} with index sets:\n",
       "    Dimension 1, 0:7\n",
       "    Dimension 2, 0:5\n",
       "    Dimension 3, 1:6\n",
       "And data, a 8×6×6 Array{Float64,3}:\n",
       "[:, :, 1] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  1.0  1.0  1.0  1.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 2] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 3] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  1.0  1.0  1.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  1.0  1.0  1.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 4] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  1.0  1.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 5] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  1.0  1.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       "\n",
       "[:, :, 6] =\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0\n",
       " 0.0  1.0  1.0  1.0  1.0  0.0\n",
       " 0.0  0.0  0.0  0.0  0.0  0.0"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "v_ans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Julia 1.0.4",
   "language": "julia",
   "name": "julia-1.0"
  },
  "language_info": {
   "file_extension": ".jl",
   "mimetype": "application/julia",
   "name": "julia",
   "version": "1.0.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}