{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "827e9aae-17df-4df6-9441-d9a6486d4b84",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import rastereasy"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8903a54f-410d-487e-b9ab-bc455cc9259d",
   "metadata": {},
   "source": [
    "# Dempster-Shafer fusion for 2 hypotheses\n",
    "\n",
    "It is possible to fuse belief masses issued from images for two hypotheses A and B.\n",
    "For this, we assume that each image is composed of 3 spectral bands : \n",
    "- 1st band : belief mass $m(A)$ for hypothesis $A$\n",
    "- 2nd band : belief mass $m(B)$ for hypothesis $B$\n",
    "- 3rd band : belief mass $m(A \\cup B)$ for hypothesis $A \\cup B$ (uncertainty)\n",
    "\n",
    "For consistent mass functions, the sum along bands must be $1$ for each pixel of each source"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed1ea3b8-0564-4280-8c37-0b45a4570986",
   "metadata": {},
   "source": [
    "## 1) Create 5 toy images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "b488fb18-aed9-4ddd-a22d-a0795587b4b8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1000, 1000, 3)\n",
      "[[1. 1. 1. ... 1. 1. 1.]\n",
      " [1. 1. 1. ... 1. 1. 1.]\n",
      " [1. 1. 1. ... 1. 1. 1.]\n",
      " ...\n",
      " [1. 1. 1. ... 1. 1. 1.]\n",
      " [1. 1. 1. ... 1. 1. 1.]\n",
      " [1. 1. 1. ... 1. 1. 1.]]\n"
     ]
    }
   ],
   "source": [
    "# 1. Read image (just to have meta informations)\n",
    "name_im='./data/demo/sentinel.tif'\n",
    "image=rastereasy.Geoimage(name_im)\n",
    "# 2. Create a 3 band image consistent with mass functions (i.e. positive and  sum to 1)\n",
    "import numpy as np\n",
    "raw = np.random.rand(image.shape[0], image.shape[1], 3)\n",
    "# 3. Normalize  mass functions such that m(A) + m(B) + m(A∪B) = 1 in each pixel\n",
    "mass_functions = raw / np.sum(raw, axis=2, keepdims=True)\n",
    "print(mass_functions.shape)\n",
    "# 4. Upload the table \n",
    "mass_function1 = image.upload_image(mass_functions,channel_first=False)\n",
    "# 5 verify that the sum is 1 along bands\n",
    "print(mass_function1.sum(axis='band'))\n",
    "\n",
    "## Repeat this operation 5 times to create 5 sources\n",
    "raw = np.random.rand(image.shape[0], image.shape[1], 3)\n",
    "mass_functions = raw / np.sum(raw, axis=2, keepdims=True)\n",
    "mass_function2 = image.upload_image(mass_functions,channel_first=False)\n",
    "\n",
    "raw = np.random.rand(image.shape[0], image.shape[1], 3)\n",
    "mass_functions = raw / np.sum(raw, axis=2, keepdims=True)\n",
    "mass_function3 = image.upload_image(mass_functions,channel_first=False)\n",
    "\n",
    "raw = np.random.rand(image.shape[0], image.shape[1], 3)\n",
    "mass_functions = raw / np.sum(raw, axis=2, keepdims=True)\n",
    "mass_function4 = image.upload_image(mass_functions,channel_first=False)\n",
    "\n",
    "raw = np.random.rand(image.shape[0], image.shape[1], 3)\n",
    "mass_functions = raw / np.sum(raw, axis=2, keepdims=True)\n",
    "mass_function5 = image.upload_image(mass_functions,channel_first=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "288729da-9320-4917-aeb9-4c389f2fe16b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[0.12571802, 0.25086056, 0.18328983, ..., 0.09160784,\n",
       "         0.16420958, 0.22534334],\n",
       "        [0.22449783, 0.05369949, 0.13427468, ..., 0.48883442,\n",
       "         0.55197949, 0.61125815],\n",
       "        [0.41689285, 0.04157854, 0.50268653, ..., 0.07765427,\n",
       "         0.41123555, 0.33504538],\n",
       "        ...,\n",
       "        [0.05546719, 0.75829898, 0.50423343, ..., 0.38172759,\n",
       "         0.25669968, 0.37609214],\n",
       "        [0.64042402, 0.29924155, 0.1708952 , ..., 0.2388334 ,\n",
       "         0.48679624, 0.20806528],\n",
       "        [0.23922514, 0.53796906, 0.44941302, ..., 0.35755458,\n",
       "         0.25504246, 0.446345  ]],\n",
       "\n",
       "       [[0.4546271 , 0.10578535, 0.49255176, ..., 0.472167  ,\n",
       "         0.59570854, 0.38831502],\n",
       "        [0.27936   , 0.69792962, 0.194582  , ..., 0.18611645,\n",
       "         0.36901982, 0.23794597],\n",
       "        [0.27687108, 0.46032741, 0.0916974 , ..., 0.60872449,\n",
       "         0.07072767, 0.32069137],\n",
       "        ...,\n",
       "        [0.44288423, 0.12847884, 0.15907349, ..., 0.12155445,\n",
       "         0.46674052, 0.56966528],\n",
       "        [0.34087475, 0.69525966, 0.70591245, ..., 0.30560047,\n",
       "         0.32621901, 0.73158923],\n",
       "        [0.71419186, 0.05437057, 0.1414912 , ..., 0.61960515,\n",
       "         0.35577265, 0.40328638]],\n",
       "\n",
       "       [[0.41965488, 0.64335409, 0.32415841, ..., 0.43622516,\n",
       "         0.24008188, 0.38634163],\n",
       "        [0.49614217, 0.24837089, 0.67114331, ..., 0.32504913,\n",
       "         0.07900069, 0.15079588],\n",
       "        [0.30623606, 0.49809405, 0.40561608, ..., 0.31362124,\n",
       "         0.51803679, 0.34426325],\n",
       "        ...,\n",
       "        [0.50164858, 0.11322218, 0.33669308, ..., 0.49671796,\n",
       "         0.27655981, 0.05424259],\n",
       "        [0.01870123, 0.00549879, 0.12319235, ..., 0.45556613,\n",
       "         0.18698475, 0.06034549],\n",
       "        [0.046583  , 0.40766037, 0.40909578, ..., 0.02284027,\n",
       "         0.38918489, 0.15036862]]], shape=(3, 1000, 1000))"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mass_function1.image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "61797809-5aef-45e7-b0df-6ef1a81c61ab",
   "metadata": {},
   "source": [
    "## 2) Perform the fusion"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12c32490-36a4-4bc6-8516-986e59c82d11",
   "metadata": {},
   "source": [
    "### 1) With function `fuse_dempster_shafer_2` of class `Geoimage`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "62a3af07-8dc5-4ca4-a54e-56259c68ea85",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Help on method fuse_dempster_shafer_2 in module rastereasy.rastereasy:\n",
      "\n",
      "fuse_dempster_shafer_2(*images) method of rastereasy.rastereasy.Geoimage instance\n",
      "    Fuse the 3 band image (associated with mass functions) from multiple\n",
      "    sources using Dempster-Shafer theory with two hypotheses: A and B.\n",
      "    \n",
      "    Parameters\n",
      "    ----------\n",
      "    *images : Geoimage\n",
      "        Each input is a 3-band Geoimage.\n",
      "    \n",
      "        - Band 1: mass function m(A)\n",
      "    \n",
      "        - Band 2: mass function m(B)\n",
      "    \n",
      "        - Band 3: mass function m(A ∪ B)\n",
      "    \n",
      "    Returns\n",
      "    -------\n",
      "    Geoimage\n",
      "        A new Geoimage with 3 bands containing the fused mass functions:\n",
      "        m(A), m(B), and m(A ∪ B).\n",
      "    Geoimage\n",
      "        A new Geoimage with 1 band containing the conflict values.\n",
      "    \n",
      "    Examples\n",
      "    --------\n",
      "    >>> fused, conflict = im1.fuse_dempster_shafer_2(im2)\n",
      "    >>> fused, conflict = im1.fuse_dempster_shafer_2(im1, im2, im3, im4)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "help(mass_function1.fuse_dempster_shafer_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7b7cce2c-402b-44ff-9e7e-e19ca84ec83b",
   "metadata": {},
   "outputs": [],
   "source": [
    "fusion, conflict = mass_function2.fuse_dempster_shafer_2(mass_function5,mass_function4,mass_function1,mass_function3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03eea926-6037-405f-9d27-d43b74cab5e9",
   "metadata": {},
   "source": [
    "### 2) With function `fuse_dempster_shafer_2hypotheses` of class `InferenceTools`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "ae530d16-5d8c-482e-90fd-4b863275c7a9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Help on function fuse_dempster_shafer_2hypotheses in module rastereasy.rastereasy:\n",
      "\n",
      "fuse_dempster_shafer_2hypotheses(*images)\n",
      "    Fuse mass functions from multiple sources using Dempster-Shafer theory\n",
      "    with two hypotheses: A and B.\n",
      "    \n",
      "    Parameters\n",
      "    ----------\n",
      "    *images : Geoimage\n",
      "        Each input is a 3-band Geoimage.\n",
      "    \n",
      "        - Band 1: mass function m(A)\n",
      "    \n",
      "        - Band 2: mass function m(B)\n",
      "    \n",
      "        - Band 3: mass function m(A ∪ B)\n",
      "    \n",
      "    Returns\n",
      "    -------\n",
      "    Geoimage\n",
      "        A new Geoimage with 3 bands containing the fused mass functions:\n",
      "        m(A), m(B), and m(A ∪ B).\n",
      "    Geoimage\n",
      "        A new Geoimage with 1 band containing the conflict values.\n",
      "    \n",
      "    Examples\n",
      "    --------\n",
      "    >>> fused, conflict = fuse_dempster_shafer_2hypotheses(im1, im2, im3)\n",
      "    >>> fused, conflict = fuse_dempster_shafer_2hypotheses(im1, im2, im3, im4)\n",
      "    >>> fused, conflict = fuse_dempster_shafer_2hypotheses(im1, im2)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "help(rastereasy.InferenceTools.fuse_dempster_shafer_2hypotheses)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "34f12558-e609-4e5d-a738-1c5a62ce6876",
   "metadata": {},
   "outputs": [],
   "source": [
    "fusion, conflict = rastereasy.InferenceTools.fuse_dempster_shafer_2hypotheses(\n",
    "    mass_function5,\n",
    "    mass_function4,\n",
    "    mass_function3,\n",
    "    mass_function2,\n",
    "    mass_function1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97823c02-919f-4681-ab70-e3e1fbcc8ce2",
   "metadata": {},
   "source": [
    "Verify that the fused masses sum to $1$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "eb63c253-061d-4e48-ac4c-f5ce52331595",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1., 1., 1., ..., 1., 1., 1.],\n",
       "       [1., 1., 1., ..., 1., 1., 1.],\n",
       "       [1., 1., 1., ..., 1., 1., 1.],\n",
       "       ...,\n",
       "       [1., 1., 1., ..., 1., 1., 1.],\n",
       "       [1., 1., 1., ..., 1., 1., 1.],\n",
       "       [1., 1., 1., ..., 1., 1., 1.]], shape=(1000, 1000))"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fusion.sum(axis='band')"
   ]
  }
 ],
 "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": "3.10.19"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}