{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "Augmentor has warpped some commonly used basic operations so that you can use to build the augmentation\n",
    "pipeline. The `internal` column is what you'd probably do outside of `Augmentor`."
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "| Category             | internal                | Augmentor          |\n",
    "| ---                  | ---                     | ---                |\n",
    "| Conversion           | `T.(img)`               | `ConvertEltype(T)` |\n",
    "| Information Layout   | `ImageCore.channelview` | `SplitChannels`    |\n",
    "| Information Layout   | `ImageCore.colorview`   | `CombineChannels`  |\n",
    "| Information Layout   | `Base.permutedims`      | `PermuteDims`      |\n",
    "| Information Layout   | `Base.reshape`          | `Reshape`          |"
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "It is not uncommon that machine learning frameworks require the data in a specific form and layout.\n",
    "For example many deep learning frameworks expect the colorchannel of the images to be encoded in\n",
    "the third dimension of a 4-dimensional array. Augmentor allows to convert from (and to) these\n",
    "different layouts using special operations that are mainly useful in the beginning or end of a\n",
    "augmentation pipeline."
   ],
   "metadata": {}
  },
  {
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "true"
     },
     "metadata": {},
     "execution_count": 1
    }
   ],
   "cell_type": "code",
   "source": [
    "using Augmentor\n",
    "using ImageCore\n",
    "\n",
    "# 300×400 Matrix{RGB{N0f8}, 2} => 300×400×3 Array{Float32, 3}\n",
    "img = testpattern(RGB, ratio=0.5)\n",
    "img_in = augment(img, SplitChannels() |> PermuteDims(2, 3, 1) |> ConvertEltype(Float32))\n",
    "\n",
    "# 300×400×3 Array{Float32, 3} => 300×400 Matrix{RGB{N0f8}, 2}\n",
    "img_out = augment(img_in, ConvertEltype(N0f8) |> PermuteDims(3, 1, 2) |> CombineChannels(RGB))\n",
    "\n",
    "img_out == img"
   ],
   "metadata": {},
   "execution_count": 1
  },
  {
   "cell_type": "markdown",
   "source": [
    "## References"
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "---\n",
    "\n",
    "*This notebook was generated using [Literate.jl](https://github.com/fredrikekre/Literate.jl).*"
   ],
   "metadata": {}
  }
 ],
 "nbformat_minor": 3,
 "metadata": {
  "language_info": {
   "file_extension": ".jl",
   "mimetype": "application/julia",
   "name": "julia",
   "version": "1.7.3"
  },
  "kernelspec": {
   "name": "julia-1.7",
   "display_name": "Julia 1.7.3",
   "language": "julia"
  }
 },
 "nbformat": 4
}