Simple Notebook to process svs images

OpenSlide_Tiles.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import openslide\n",
    "import itertools\n",
    "from fastai2.basics import *\n",
    "from fastprogress import progress_bar"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Enter path\n",
    "path = Path('~/WAMRI/nuclei/').expanduser()\n",
    "data_path = path/'data/test_svs'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Path('/data2/ashaw/WAMRI/nuclei/data/test_svs/P1C054g07.svs'), 'aperio')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fn = data_path.ls()[0]\n",
    "fn, openslide.OpenSlide.detect_format(str(fn))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4, ((74256, 75067), (18564, 18766), (4641, 4691), (2320, 2345)))"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "slide = openslide.OpenSlide(str(fn))\n",
    "slide.level_count, slide.level_dimensions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "slide.properties"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "out_dir = path/'data/processed'\n",
    "out_dir.mkdir(parents=True, exist_ok=True)\n",
    "lvl = 0 # level 0 is the max zoom\n",
    "sz = (2048,2048) # tile resolution\n",
    "overlap = 0 # overlap between tiles\n",
    "dims = slide.level_dimensions[lvl] # overall size at zoom level"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dimx = range(0,dims[0],sz[0]-overlap)\n",
    "dimy = range(0,dims[1],sz[1]-overlap)\n",
    "regions = list(itertools.product(dimx,dimy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_region(coord):\n",
    "    x,y = coord\n",
    "    out_fn = out_dir/f'{fn.stem}_lvl{lvl}_{x}x{y}.jpg'\n",
    "    if out_fn.exists(): return\n",
    "    slide = openslide.OpenSlide(str(fn))\n",
    "    reg = slide.read_region((x,y), lvl, size=sz)\n",
    "    reg.convert('RGB').save(out_fn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(#1369) [None,None,None,None,None,None,None,None,None,None...]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_workers=16\n",
    "parallel(process_region, regions, n_workers=n_workers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(regions), len(out_dir.ls())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}