{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3a49ee24",
   "metadata": {},
   "source": [
    "# Tutorial 3: Customisation Details\n",
    "In this notebook we will expand on tutorial 2 by showing you exactly what form the transformations and IQMs need to be in.\n",
    "\n",
    "Let's start by importing the package."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b1162501",
   "metadata": {},
   "outputs": [],
   "source": [
    "import IQM_Vis"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4786d159",
   "metadata": {},
   "source": [
    "## Metrics\n",
    "A metric is defined as a callable object e.g. a function.\n",
    "\n",
    "Custom metrics need to take arguments:\n",
    " - im_ref: the reference image (numpy array - float in the range 0-1)\n",
    " - im_comp: the comparison/distorted image (numpy array - float in the range 0-1)\n",
    " - **kwargs: peel off any extra key word arguments that might be called to the function\n",
    " \n",
    "Metrics need to return a single value (float) with the score\n",
    "\n",
    "We can make our own mean squared error function:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "263197ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "def custom_MAE_function(im_ref, im_comp, **kwargs):\n",
    "    L1 = np.abs(im_ref - im_comp)\n",
    "    return L1.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0da93aa",
   "metadata": {},
   "source": [
    "We can also define metrics as class objects. This is useful if you need to store attributes. You will still need to class to be callable like the function above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "47f9225d",
   "metadata": {},
   "outputs": [],
   "source": [
    "class custom_MAE_class:\n",
    "    def __init__(self, att=0):\n",
    "        self.att = att\n",
    "        \n",
    "    def __call__(self, im_ref, im_comp, **kwargs):\n",
    "        L1 = np.abs(im_ref - im_comp)\n",
    "        return L1.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a46b232",
   "metadata": {},
   "source": [
    "Extra key word arguments can be included if the metric needs to take them. Make sure they have a default value and you can add them as parameters to the IQM-Vis UI"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "602d1b12",
   "metadata": {},
   "outputs": [],
   "source": [
    "def dummy_args(im_ref, im_comp, param1=0, **kwargs):\n",
    "    # now we can use param here\n",
    "    score = custom_MAE_function(im_ref, im_comp)\n",
    "    return score + param1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7f216509",
   "metadata": {},
   "source": [
    "Now we can add them to the dictionary to pass to the IQM-Vis UI as shown in tutorial 2. Note that the class metric need to initialised."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8dd95119",
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics = {'MAE function': custom_MAE_function,\n",
    "          'MAE class': custom_MAE_class(),\n",
    "          'dummy args': dummy_args}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8938ef3",
   "metadata": {},
   "source": [
    "We also need to tell IQM-Vis about param1 to be used by our dummy_args function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "ac5b1655",
   "metadata": {},
   "outputs": [],
   "source": [
    "params = {'param1': {'min':-1.0, 'max':1.0, 'init_value': 0}}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "45fbae7b",
   "metadata": {},
   "source": [
    "## Transformations\n",
    "Transforms/distortions are callable with the following arguments:\n",
    " - image: image to be transformed (numpy array - float in the range 0-1)\n",
    " - param: value of the parameter of the transformation\n",
    "\n",
    "Transforms need to return a numpy array of the same shape as the image input and be within the bounds 0-1.\n",
    "\n",
    "For example we can make our own brightness adjuster, note the parameter 'value' can be called anything:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cce17a36",
   "metadata": {},
   "outputs": [],
   "source": [
    "def custom_brightness(image, value=0):\n",
    "    return np.clip(image + value, 0, 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4487ad14",
   "metadata": {},
   "source": [
    "Then we put in a dictionary with the parameter range to tell the UI later"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "00ebecb1",
   "metadata": {},
   "outputs": [],
   "source": [
    "transformations = {'brightness':{'min':-1.0, 'max':1.0, 'function':custom_brightness}}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cee8b094",
   "metadata": {},
   "source": [
    "## Making the UI\n",
    "Now make the UI as shown in tutorial 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ec3b343d",
   "metadata": {},
   "outputs": [],
   "source": [
    "images = ['/home/matt/datasets/kodak/kodim01.png', \n",
    "          '/home/matt/datasets/kodak/kodim02.png']\n",
    "\n",
    "IQM_Vis.make_UI(transformations=transformations,\n",
    "                image_list=images,\n",
    "                metrics=metrics,\n",
    "                metric_params=params)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cdd47d68",
   "metadata": {},
   "source": [
    "![](./resources/custom_UI_2.png)"
   ]
  }
 ],
 "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}