{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from datasets import load_dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatasetDict({\n",
       "    train: Dataset({\n",
       "        features: ['patient_id', 'series_id', 'frame_id', 'image', 'mask'],\n",
       "        num_rows: 70291\n",
       "    })\n",
       "})"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = load_dataset('ziq/RSNA-ATD2023')\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = data['train'].train_test_split(test_size=0.2)\n",
    "labels = [\"background\", \"liver\", \"spleen\", \"right_kidney\", \"left_kidney\", \"bowel\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(56232, 14059)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train, test = data['train'], data['test']\n",
    "len(train), len(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'patient_id': 62360,\n",
       " 'series_id': 51136,\n",
       " 'frame_id': 631,\n",
       " 'image': <PIL.PngImagePlugin.PngImageFile image mode=L size=512x512>,\n",
       " 'mask': <PIL.PngImagePlugin.PngImageFile image mode=L size=512x512>}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ids = 3\n",
    "image, mask = train[ids]['image'], train[ids]['mask']\n",
    "\n",
    "mask = np.array(mask)\n",
    "liver, spleen, right_kidney, left_kidney, bowel = [(mask == i,1,0)[0] * i for i in range(1, len(labels))]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.colors import ListedColormap, BoundaryNorm\n",
    "\n",
    "colors = ['#02020e', '#520e6d', '#c13a50', '#f57d15', '#fac62c', '#f4f88e'] # inferno\n",
    "bounds = range(0, len(colors) + 1)\n",
    "\n",
    "# Define the boundaries for each class in the colormap\n",
    "cmap, norm = ListedColormap(colors), BoundaryNorm(bounds, len(colors))\n",
    "\n",
    "# Plot the segmentation mask with the custom colormap\n",
    "def plot_mask(mask, alpha=1.0):\n",
    "    _, ax = plt.subplots()\n",
    "    cax = ax.imshow(mask, cmap=cmap, norm=norm, alpha=alpha)\n",
    "    cbar = plt.colorbar(cax, cmap=cmap, norm=norm, boundaries=bounds, ticks=bounds)\n",
    "    cbar.set_ticks([])\n",
    "    _labels = [\"\"] + labels\n",
    "    for i in range(1, len(_labels)):\n",
    "        cbar.ax.text(2, -0.5 + i, _labels[i], ha='left', color=colors[i - 1], fontsize=8)\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "    \n",
    "def plot(image, mask, cmap='gray'):\n",
    "    plt.imshow(image * np.where(mask > 0,1,0), cmap=cmap)\n",
    "    plt.axis('off')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(image, mask)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_mask(mask)"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}