{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gradio as gr\n",
    "import numpy as np\n",
    "from os import environ\n",
    "from PIL import Image as PImage\n",
    "from torchvision import transforms as T\n",
    "from transformers import MaskFormerForInstanceSegmentation, MaskFormerImageProcessor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ade_mean=[0.485, 0.456, 0.406]\n",
    "ade_std=[0.229, 0.224, 0.225]\n",
    "\n",
    "palette = [\n",
    "    [120, 120, 120], [4, 200, 4], [4, 4, 250], [6, 230, 230],\n",
    "    [80, 50, 50], [120, 120, 80], [140, 140, 140], [204, 5, 255]\n",
    "]\n",
    "\n",
    "model_id = f\"thiagohersan/maskformer-satellite-trees\"\n",
    "\n",
    "vegetation_labels = [\"vegetation\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize_instance_seg_mask(img_in, mask, id2label, included_labels):\n",
    "    img_out = np.zeros((mask.shape[0], mask.shape[1], 3))\n",
    "    image_total_pixels = mask.shape[0] * mask.shape[1]\n",
    "    label_ids = np.unique(mask)\n",
    "\n",
    "    id2color = {id: palette[id] for id in label_ids}\n",
    "    id2count = {id: 0 for id in label_ids}\n",
    "\n",
    "    for i in range(img_out.shape[0]):\n",
    "      for j in range(img_out.shape[1]):\n",
    "        img_out[i, j, :] = id2color[mask[i, j]]\n",
    "        id2count[mask[i, j]] = id2count[mask[i, j]] + 1\n",
    "\n",
    "    image_res = (0.5 * img_in + 0.5 * img_out).astype(np.uint8)\n",
    "\n",
    "    dataframe = [[\n",
    "        f\"{id2label[id]}\",\n",
    "        f\"{(100 * id2count[id] / image_total_pixels):.2f} %\",\n",
    "        f\"{np.sqrt(id2count[id] / image_total_pixels):.2f} m\"\n",
    "        ] for id in label_ids if id2label[id] in included_labels]\n",
    "\n",
    "    if len(dataframe) < 1:\n",
    "        dataframe = [[\n",
    "            f\"\",\n",
    "            f\"{(0):.2f} %\",\n",
    "            f\"{(0):.2f} m\"\n",
    "        ]]\n",
    "\n",
    "    return image_res, dataframe\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# preprocessor = MaskFormerImageProcessor.from_pretrained(model_id)\n",
    "preprocessor = MaskFormerImageProcessor(\n",
    "    do_resize=False,\n",
    "    do_normalize=False,\n",
    "    do_rescale=False,\n",
    "    ignore_index=255,\n",
    "    do_reduce_labels=False\n",
    ")\n",
    "\n",
    "hf_token = environ.get('HFTOKEN') or True\n",
    "model = MaskFormerForInstanceSegmentation.from_pretrained(model_id, use_auth_token=hf_token)\n",
    "\n",
    "test_transform = T.Compose([\n",
    "    T.ToTensor(),\n",
    "    T.Normalize(mean=ade_mean, std=ade_std)\n",
    "])\n",
    "\n",
    "with PImage.open(\"../color-filter-calculator/assets/Artshack_screen.jpg\") as img:\n",
    "    img_size = (img.height, img.width)\n",
    "    norm_image = test_transform(np.array(img))\n",
    "    inputs = preprocessor(images=norm_image, return_tensors=\"pt\")\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "outputs = model(**inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "results = preprocessor.post_process_semantic_segmentation(outputs=outputs, target_sizes=[img_size])[0]\n",
    "mask_img, dataframe = visualize_instance_seg_mask(np.array(img), results.numpy(), model.config.id2label, vegetation_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataframe"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.8.15 ('gradio2023')",
   "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": "3.9.17"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "311e94dbd43374307e33a15d3b7324b15a4f7b1d7ecfe8226f18075b87b9fae7"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}