{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Select glaciers within region shapes and basic attributes: Level 2 files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "L0 and L1 files are too large: they are selected roughly by lon lat boxes and a buffer, so there is a lot of duplication. Here we select glaciers by region by overlaying the glaciers representative points with the region outlines. This reduces the size of the files and make them ready for regional processing.\n",
    "\n",
    "In addition, we compute glacier area and store the CenLon, CenLat attributes for later use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import utils \n",
    "import geopandas as gpd\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import shutil\n",
    "import glob\n",
    "import os\n",
    "import subprocess\n",
    "import tarfile\n",
    "import shapely.geometry as shpg\n",
    "import progressbar\n",
    "import matplotlib.pyplot as plt\n",
    "import warnings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def xy_coord(geom):\n",
    "    x, y = geom.xy\n",
    "    return x[0], y[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# go down from rgi7_scripts/workflow/preprocessing\n",
    "data_dir = '../../../rgi7_data/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "reg_file = os.path.join(data_dir, '00_rgi70_regions', '00_rgi70_O1Regions')\n",
    "reg_f = gpd.read_file(reg_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "../../../rgi7_data//l1_with_interiors/RGI01/RGI01.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI01/\n",
      "RGI01/RGI01.dbf\n",
      "RGI01/RGI01.shp\n",
      "RGI01/RGI01.prj\n",
      "RGI01/RGI01.cpg\n",
      "RGI01/RGI01.shx\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI01.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI01'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI02/RGI02.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI02/\n",
      "RGI02/RGI02.dbf\n",
      "RGI02/RGI02.shp\n",
      "RGI02/RGI02.prj\n",
      "RGI02/RGI02.cpg\n",
      "RGI02/RGI02.shx\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI02.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI02'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI03/RGI03.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI03/\n",
      "RGI03/RGI03.dbf\n",
      "RGI03/RGI03.shp\n",
      "RGI03/RGI03.cpg\n",
      "RGI03/RGI03.prj\n",
      "RGI03/RGI03.shx\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI03.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI03'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI04/RGI04.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI04/\n",
      "RGI04/RGI04.cpg\n",
      "RGI04/RGI04.prj\n",
      "RGI04/RGI04.shx\n",
      "RGI04/RGI04.shp\n",
      "RGI04/RGI04.dbf\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI04.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI04'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI05/RGI05.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI05/\n",
      "RGI05/RGI05.shp\n",
      "RGI05/RGI05.dbf\n",
      "RGI05/RGI05.cpg\n",
      "RGI05/RGI05.prj\n",
      "RGI05/RGI05.shx\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI05.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI05'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI06/RGI06.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI06/\n",
      "RGI06/RGI06.shp\n",
      "RGI06/RGI06.dbf\n",
      "RGI06/RGI06.shx\n",
      "RGI06/RGI06.prj\n",
      "RGI06/RGI06.cpg\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI06.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI06'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI07/RGI07.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI07/\n",
      "RGI07/RGI07.cpg\n",
      "RGI07/RGI07.prj\n",
      "RGI07/RGI07.shx\n",
      "RGI07/RGI07.dbf\n",
      "RGI07/RGI07.shp\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI07.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI07'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI08/RGI08.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI08/\n",
      "RGI08/RGI08.shp\n",
      "RGI08/RGI08.dbf\n",
      "RGI08/RGI08.shx\n",
      "RGI08/RGI08.cpg\n",
      "RGI08/RGI08.prj\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI08.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI08'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI09/RGI09.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI09/\n",
      "RGI09/RGI09.dbf\n",
      "RGI09/RGI09.shp\n",
      "RGI09/RGI09.shx\n",
      "RGI09/RGI09.prj\n",
      "RGI09/RGI09.cpg\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI09.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI09'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI10/RGI10.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI10/\n",
      "RGI10/RGI10.shp\n",
      "RGI10/RGI10.dbf\n",
      "RGI10/RGI10.shx\n",
      "RGI10/RGI10.prj\n",
      "RGI10/RGI10.cpg\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI10.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI10'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI11/RGI11.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI11/\n",
      "RGI11/RGI11.dbf\n",
      "RGI11/RGI11.shp\n",
      "RGI11/RGI11.shx\n",
      "RGI11/RGI11.prj\n",
      "RGI11/RGI11.cpg\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI11.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI11'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI12/RGI12.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI12/\n",
      "RGI12/RGI12.prj\n",
      "RGI12/RGI12.cpg\n",
      "RGI12/RGI12.shx\n",
      "RGI12/RGI12.dbf\n",
      "RGI12/RGI12.shp\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI12.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI12'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI13/RGI13.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI13/\n",
      "RGI13/RGI13.prj\n",
      "RGI13/RGI13.cpg\n",
      "RGI13/RGI13.shx\n",
      "RGI13/RGI13.shp\n",
      "RGI13/RGI13.dbf\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI13.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI13'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI14/RGI14.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI14/\n",
      "RGI14/RGI14.cpg\n",
      "RGI14/RGI14.prj\n",
      "RGI14/RGI14.shx\n",
      "RGI14/RGI14.dbf\n",
      "RGI14/RGI14.shp\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI14.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI14'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI15/RGI15.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI15/\n",
      "RGI15/RGI15.shx\n",
      "RGI15/RGI15.cpg\n",
      "RGI15/RGI15.prj\n",
      "RGI15/RGI15.dbf\n",
      "RGI15/RGI15.shp\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI15.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI15'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI16/RGI16.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI16/\n",
      "RGI16/RGI16.shp\n",
      "RGI16/RGI16.dbf\n",
      "RGI16/RGI16.shx\n",
      "RGI16/RGI16.cpg\n",
      "RGI16/RGI16.prj\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI16.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI16'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI17/RGI17.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI17/\n",
      "RGI17/RGI17.shp\n",
      "RGI17/RGI17.dbf\n",
      "RGI17/RGI17.shx\n",
      "RGI17/RGI17.prj\n",
      "RGI17/RGI17.cpg\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI17.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI17'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI18/RGI18.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI18/\n",
      "RGI18/RGI18.shp\n",
      "RGI18/RGI18.dbf\n",
      "RGI18/RGI18.shx\n",
      "RGI18/RGI18.cpg\n",
      "RGI18/RGI18.prj\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI18.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI18'], returncode=0)\n",
      "../../../rgi7_data//l1_with_interiors/RGI19/RGI19.shp\n",
      "Reading...\n",
      "Pointing...\n",
      "Area-ing...\n",
      "Writing...\n",
      "Taring...\n",
      "RGI19/\n",
      "RGI19/RGI19.prj\n",
      "RGI19/RGI19.cpg\n",
      "RGI19/RGI19.shx\n",
      "RGI19/RGI19.dbf\n",
      "RGI19/RGI19.shp\n",
      "CompletedProcess(args=['tar', '-zcvf', '../../../rgi7_data//l2_sel_reg_tars/RGI19.tar.gz', '-C', '../../../rgi7_data//l2_sel_reg', 'RGI19'], returncode=0)\n"
     ]
    }
   ],
   "source": [
    "idir = utils.mkdir(data_dir + '/l1_with_interiors')\n",
    "odir = utils.mkdir(data_dir + '/l2_sel_reg')\n",
    "odir_tar = utils.mkdir(data_dir + '/l2_sel_reg_tars')\n",
    "\n",
    "for reg in range(1, 20):\n",
    "#for reg in [4, 19]:\n",
    "\n",
    "    fp = f'{idir}/RGI{reg:02d}/RGI{reg:02d}.shp'\n",
    "    print(fp)\n",
    "    print('Reading...')\n",
    "    shp = gpd.read_file(fp)\n",
    "    \n",
    "    print('Pointing...')\n",
    "    \n",
    "    rp = shp.representative_point()\n",
    "    \n",
    "    coordinates = np.array(list(rp.apply(xy_coord)))\n",
    "    shp['CenLon'] = coordinates[:, 0]\n",
    "    shp['CenLat'] = coordinates[:, 1]\n",
    "    \n",
    "    rp = rp.to_frame('geometry')\n",
    "    rp['orig_index'] = shp.index\n",
    "    \n",
    "    intersect = gpd.overlay(rp, reg_f.loc[reg_f.o1region == f'{reg:02d}'], how='intersection')\n",
    "    \n",
    "    odf = shp.loc[intersect['orig_index']]\n",
    "    \n",
    "    print('Area-ing...')\n",
    "    odf['area'] = odf.to_crs({'proj':'cea'}).area\n",
    "    \n",
    "    dd = utils.mkdir(f'{odir}/RGI{reg:02d}/', reset=True)\n",
    "    \n",
    "    print('Writing...')\n",
    "    odf.to_file(dd + f'RGI{reg:02d}.shp')\n",
    "    \n",
    "    print('Taring...')\n",
    "    print(subprocess.run(['tar', '-zcvf', f'{odir_tar}/RGI{reg:02d}.tar.gz', '-C', odir, f'RGI{reg:02d}']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done!\n"
     ]
    }
   ],
   "source": [
    "print('Done!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.10.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}