lidar_workflow_applied_SCB.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "797f557e-435f-49cd-8002-2135758b8c67",
   "metadata": {},
   "source": [
    "# Lidar Processing Workflow\n",
    "**Cara Piske, Graduate Program of Hydrologic Sciences, 2022; Advisor: Dr. Adrian Harpold**<br>\n",
    "This code processes raw lidar point clouds in order to calculate snow depth using PDAL. <br>\n",
    "Lidar data were provided by the Airborne Snow Observatory (ASO), the National Center for Airborne Laser Mapping (NCALM), and Watershed Sciences Inc. (WSI). <br>\n",
    "\n",
    "The goal of this project is to process snow depth to the one-meter spatial scale while maintaining conservative under-canopy estimates. Therefore, little interpolation occurs under-canopy. We follow these protocols in order to obtain a 1-m rasterized product (as opposed to the 3-m rasterized product provided by ASO on the NSIDC data portal). NCALM and WSI flights were obtained through OpenTopography."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "01e7346b-ed91-48fe-912a-4dc9d7b20be4",
   "metadata": {},
   "source": [
    "Start by importing necessary files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "ef19ecac-1150-469a-9a3f-3eb6f8b1c531",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import necessary files\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt\n",
    "import json # where we will save the json files to run a pipeline\n",
    "import os # for file management\n",
    "import subprocess # allows us to run command line commands\n",
    "import pdal # lidar processing package\n",
    "import rasterio\n",
    "import pyproj\n",
    "from osgeo import gdal\n",
    "\n",
    "import time\n",
    "# packages to copy files\n",
    "from pathlib import Path\n",
    "import shutil\n",
    "import glob\n",
    "# packages to extract wkt from polygon\n",
    "import shapefile\n",
    "import pygeoif\n",
    "# for parallel processing\n",
    "import concurrent.futures\n",
    "from multiprocessing import Pool\n",
    "import csv\n",
    "\n",
    "import re\n",
    "import seaborn as sns\n",
    "import scipy\n",
    "#import seaborn as sns\n",
    "from scipy.interpolate import interp1d\n",
    "from scipy.stats import kde\n",
    "from scipy.stats import gaussian_kde\n",
    "# from sklearn import linear_model\n",
    "# from sklearn.metrics import mean_squared_error, r2_score\n",
    "# See lidar_functions.py\n",
    "import lidar_functions\n",
    "pdal_pipeline = 'C:\\\\Users\\cpiske\\.conda\\envs\\lidar\\Lib\\site-packages\\pdal\\pipeline.py'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afec494a-5f52-4fc5-8ced-a5886f938892",
   "metadata": {},
   "source": [
    "Note that many functions are dependent on specific directory structures. See README"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "11eb2762-4c68-4a8e-a2e0-3c4434606bbf",
   "metadata": {},
   "outputs": [],
   "source": [
    "json_base_path = 'piske_processing/PDAL_workflow/JSON/' # set so that we can redefine json across operating systems"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1933dbf5-603d-4eac-a4df-8ec90e6278d4",
   "metadata": {},
   "source": [
    "Applied"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e7a6c326-633d-4550-b011-0f012e2b2825",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'G:\\\\'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Set current working directory\n",
    "#path = '/Volumes/Piske_lidar'\n",
    "path = '/'\n",
    "os.chdir(path)\n",
    "os.getcwd() # print to ensure we're in the right directory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "87f30faa-fe09-4f19-b736-214b2aad11c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "json_base_path = 'piske_processing/PDAL_workflow/JSON/' # set so that we can redefine json across operating systems"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "600a92ec-7986-41ff-99d1-28daea0535a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define all json files\n",
    "filterMergeRasterize_json = json_base_path + 'filterMergeRasterize.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c9edcfd-2013-4858-90a5-d9ca09c35507",
   "metadata": {},
   "outputs": [],
   "source": [
    "# All json paths\n",
    "DEM_json = json_base_path+'DEM_from_las.json' \n",
    "HAG_json = json_base_path + 'HAG_dem.json'\n",
    "clip_json = json_base_path +'clip_to_polygon.json'\n",
    "extract_las_atPoint = json_base_path +'extract_las_atPoint.json'\n",
    "extract_las_atPolygon = json_base_path +'extract_las_atPolygon.json'\n",
    "correct_merge_rasterize_json = json_base_path + 'correct_merge_rasterize.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "cda00195-88f1-48ea-a2fe-addbc08c7eef",
   "metadata": {},
   "outputs": [],
   "source": [
    "writers_bounds = '([730235.96, 738826.45], [4364741.66, 4372273.16])'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c371e470-c770-4b4d-93a4-52d882661c02",
   "metadata": {},
   "source": [
    "# Pre-Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "019afb1c-f0f0-4e99-8072-cdabfaee5be6",
   "metadata": {},
   "source": [
    "## Info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "caff1251-3a50-4fec-9e91-4837b39954f5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CompletedProcess(args=['pdal', 'info', 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160417/NAD83_NAD83_epoch2010/ASO_SCB_20160417_730234_4365782.laz', '--metadata'], returncode=0)"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# define input file\n",
    "input_lid = 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160417/NAD83_NAD83_epoch2010/ASO_SCB_20160417_730234_4365782.laz'\n",
    "pdal_info_cmd = ['pdal','info',input_lid] # general info\n",
    "pdal_metadata_cmd = ['pdal','info',input_lid,'--metadata'] # full file metadata, including details crs\n",
    "subprocess.run(pdal_metadata_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "47ecd2f6-067f-4c0a-bba7-86442f1d6c7f",
   "metadata": {},
   "outputs": [],
   "source": [
    "pdal_info_results = subprocess.run(pdal_metadata_cmd, stdout = subprocess.PIPE) # stout (standard out), PIPE indicates that a new pipe to the child should be created\n",
    "pdal_info_dict = json.loads(pdal_info_results.stdout.decode()) # create dict with metadata info\n",
    "#pdal_info_dict # print results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "25fb2bc9-743a-480a-9ce6-7838c8ca8be6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# pdal_info_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2bd7cd5d-0ac8-419c-9d36-fa0589d65e04",
   "metadata": {},
   "source": [
    "Check extents (in order to create bounds for rasterization below)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fcfa978a-0e82-4d69-81f3-fcc1a59e258c",
   "metadata": {},
   "outputs": [],
   "source": [
    "lidar_folder = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20220321/las/'\n",
    "# lidar_folder = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/NAD83_NAD83_epoch2010/'\n",
    "onlyfiles = [f for f in os.listdir(lidar_folder) if os.path.isfile(os.path.join(lidar_folder, f))]\n",
    "full_paths = [lidar_folder +f for f in onlyfiles]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e0990454-2c2d-4b1b-906a-0561933b7954",
   "metadata": {},
   "outputs": [],
   "source": [
    "# NCALM_2014_ext = lidar_functions.check_flight_extent(full_paths)\n",
    "# ASO_SCB_20160326_ext = lidar_functions.check_flight_extent(full_paths)\n",
    "# ASO_SCB_20160417_ext = lidar_functions.check_flight_extent(full_paths)\n",
    "# ASO_SCB_20160518_ext = lidar_functions.check_flight_extent(full_paths)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8516d352-8076-4ecd-89bd-2ca235b8ba93",
   "metadata": {},
   "outputs": [],
   "source": [
    "# print(NCALM_2014_ext)\n",
    "# print(ASO_SCB_20160326_ext)\n",
    "# print(ASO_SCB_20160417_ext)\n",
    "print(ASO_SCB_20160518_ext)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ff1066ad-9023-43c1-a6a9-199681166dec",
   "metadata": {},
   "source": [
    "05/18/2016 has the most limited extent, so we'll use those bounds for rasterization."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4751d94d-d53d-40ce-a8d4-d62adad47b08",
   "metadata": {},
   "source": [
    "## Retile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ea4032e7-3646-4f0e-9d5e-acc98c7ea589",
   "metadata": {},
   "outputs": [],
   "source": [
    "lidar_folder = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/NAD83_NAD83_epoch2010/'\n",
    "retile_folder = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/retile_uo/'\n",
    "onlyfiles = [f for f in os.listdir(lidar_folder) if os.path.isfile(os.path.join(lidar_folder, f))]\n",
    "for files in onlyfiles:\n",
    "    full_path = os.path.join(lidar_folder, files)\n",
    "    output_path = retile_folder+'#' + files\n",
    "    retile_command = ['pdal', 'tile', full_path, output_path, '--length=1000','--buffer=50']\n",
    "    subprocess.run(retile_command)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26292a6b-d7e1-4671-8a2c-81b27432469c",
   "metadata": {},
   "source": [
    "## Rename"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fd725789-60cb-4aa6-b79f-dfca7e1ff3f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "lidar_folder = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20220321/laz/'\n",
    "tic = time.perf_counter()\n",
    "if __name__ == '__main__':\n",
    "    with concurrent.futures.ProcessPoolExecutor(max_workers=10) as executor:\n",
    "        onlyfiles = [f for f in os.listdir(lidar_folder) if os.path.isfile(os.path.join(lidar_folder, f))]\n",
    "        full_path = [os.path.join(lidar_folder, s) for s in onlyfiles]\n",
    "        executor.map(lidar_functions.rename_llx_lly, full_path) \n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aa1b2a67-2212-4cd4-94e1-41f66634a893",
   "metadata": {},
   "outputs": [],
   "source": [
    "lidar_folder = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20220321/laz/'\n",
    "tic = time.perf_counter()\n",
    "if __name__ == '__main__':\n",
    "    with concurrent.futures.ProcessPoolExecutor(max_workers=10) as executor:\n",
    "        onlyfiles = [f for f in os.listdir(lidar_folder) if os.path.isfile(os.path.join(lidar_folder, f))]\n",
    "        full_path = [os.path.join(lidar_folder, s) for s in onlyfiles]\n",
    "        executor.map(lidar_functions.add_str_to_filename, full_path) \n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a97d2d1-e28e-4bed-a4ba-b0363d4434a4",
   "metadata": {},
   "source": [
    "## Save Tile Boundaries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cbace8bd-7da0-4913-9960-c65f1eca84cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# one folder\n",
    "tic = time.perf_counter()\n",
    "lidar_folder = 'MRB/Merced_lidar/ASO/ASO_MRB_20210429/laz/'\n",
    "output_folder = 'MRB/Merced_lidar/ASO/ASO_MRB_20210429/tindex/original/'\n",
    "if __name__ == '__main__':\n",
    "    with concurrent.futures.ProcessPoolExecutor(max_workers=3) as executor:\n",
    "        onlyfiles = [f for f in os.listdir(lidar_folder) if os.path.isfile(os.path.join(lidar_folder, f))]\n",
    "        full_path = [os.path.join(lidar_folder, s) for s in onlyfiles]\n",
    "        # change depending on directory formats\n",
    "        #output_path = [os.path.join(os.path.basename(lidar_folder),'tindex/tiles/' + s[:-3] + 'sqlite') for s in onlyfiles]\n",
    "        output_path = [output_folder + s[:-3] + 'sqlite' for s in onlyfiles]\n",
    "        executor.map(lidar_functions.create_tindex, full_path, output_path) #running 10 times\n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a59151d2-faa6-445f-b84a-86ecb4fa60fe",
   "metadata": {},
   "source": [
    "## Copy Files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4528ffe9-6116-491e-967c-7fbfb1fdbade",
   "metadata": {},
   "outputs": [],
   "source": [
    "tic = time.perf_counter()\n",
    "lidar_folder = 'MRB/Merced_lidar/ASO/ASO_MRB_20210429/laz/'\n",
    "output_path = 'MRB/Merced_lidar/ASO/ASO_MRB_20210429/ICB_tiles/'\n",
    "onlyfiles = [f for f in os.listdir(lidar_folder) if os.path.isfile(os.path.join(lidar_folder, f))]\n",
    "full_paths = [os.path.join(lidar_folder, s) for s in onlyfiles] \n",
    "output_paths = np.repeat(output_path, len(full_paths))\n",
    "if __name__ == '__main__':\n",
    "    with concurrent.futures.ProcessPoolExecutor(max_workers=3) as executor:\n",
    "        executor.map(lidar_functions.copy_lid_by_ext_ICB, full_paths, output_paths)\n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6df8c3d4-0865-494a-915e-3e0a3d4051ed",
   "metadata": {},
   "source": [
    "**Fix Metadata**\n",
    "\n",
    "Due to an issue with earlier vertical datum conversions (using VDATUM version before update in 05/2022), some files were repreojected into the incorrect CRS. Others were simply assigned the wrong metadata. We'll deal with this issue within the rasterization pipelines. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a8934cc-5009-45a1-9b5b-852df2c3a908",
   "metadata": {},
   "source": [
    "# Create DEM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c05e3051-45c4-4ab1-8076-68fe36d1e098",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "DEM_json = json_base_path+'DEM_from_las.json' # define path to json files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "cd7b5912-904f-4064-9a78-32ac0ad932aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/NAD83_NAD83_epoch2010/'\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/DEM/SCB/NCALM_2014_BE.tif'\n",
    "onlyfiles = [f for f in os.listdir(input_path) if os.path.isfile(os.path.join(input_path, f))]\n",
    "input_list = [input_path + s for s in onlyfiles]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "0850ec42-97c2-4944-88b4-256f8d4dbe51",
   "metadata": {},
   "outputs": [],
   "source": [
    "filename_dict = {}\n",
    "tags = ['']*len(input_list)\n",
    "filenames = ['']*len(input_list)\n",
    "for i in range(len(input_list)):\n",
    "    filename_dict['filename_'+str(i)] = {'filename':input_list[i], 'tag':'A_'+str(i)}\n",
    "    tags[i] = 'A_'+str(i)\n",
    "    filenames[i] = filename_dict[list(filename_dict)[i]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "6f364afe-c052-49b4-a645-25e33b360247",
   "metadata": {},
   "outputs": [],
   "source": [
    "#writers_bounds = '(['+ str(ASO_SCB_20160518_ext[0])+','+ str(ASO_SCB_20160518_ext[1])+'],['+str(ASO_SCB_20160518_ext[2])+','+str(ASO_SCB_20160518_ext[3])+'])'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "b114ffa3-9aef-4a9c-91cb-0528a5a27d9b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# filter out the ground points of the tiles\n",
    "filter_range = {\"type\":\"filters.range\",\n",
    "                \"limits\":\"Classification[2:2]\"}\n",
    "# merge all las files or stages\n",
    "filter_merge = {\"type\":\"filters.merge\",\n",
    "               \"tag\": \"merged\",\n",
    "               \"inputs\": tags}\n",
    "# write merged las to raster\n",
    "writers_gdal= {\"type\": \"writers.gdal\",\n",
    "               'output_type': 'mean', \n",
    "              'resolution': '1.0',\n",
    "               'override_srs':'EPSG:6339',\n",
    "               #'bounds': writers_bounds,\n",
    "              'radius': '0.7',\n",
    "               'window_size':3,# we want more of a wall-to-wall product here so we use a secondary algorithm to increase calculation distance\n",
    "               'filename': output_tif}\n",
    "# Append each stage to a list prior to saving to json \n",
    "pipeline_list = filenames.copy()\n",
    "pipeline_list.append(filter_merge)\n",
    "pipeline_list.append(filter_range)\n",
    "pipeline_list.append(writers_gdal)\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "# save to json\n",
    "with open(DEM_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "0c1324e2-3971-44fc-a2f8-10f60ab76ab0",
   "metadata": {},
   "outputs": [],
   "source": [
    "tic = time.perf_counter()\n",
    "pdal_cmd = ['pdal','pipeline', DEM_json]\n",
    "subprocess.run(pdal_cmd)\n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "bf0d2cf8-94ad-49b5-ad98-938347a60cce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "32.35950028999999"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(toc-tic)/60"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b3bb028f-69b8-4ebd-8c38-0e66b294858a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "13399307-5a71-472d-9039-262c703e83e7",
   "metadata": {},
   "source": [
    "## Heigh Above Ground"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "cf501c4d-8952-462c-9c9a-138e0b7e52eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define json path \n",
    "HAG_json = json_base_path + 'HAG_dem.json'\n",
    "# convert all z values to the height above ground \n",
    "target_dem = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/DEM/NCALM_2014_BE.tif'\n",
    "filter_hag = {\"type\":\"filters.hag_dem\",\n",
    "              \"raster\":target_dem, # full file path of target DEM (.tif)\n",
    "              \"zero_ground\":\"false\"} # Do not assign 0 to ground classified points\n",
    "filter_ferry = {\"type\":\"filters.ferry\",\n",
    "                \"dimensions\":\"HeightAboveGround=>Z\"} # replace all Z dimensions with HAG instead of elevation\n",
    "filter_range = {\"type\":\"filters.range\",\n",
    "                \"limits\":\"Z[-0.2:70]\"} # apply a noise filter\n",
    "pipeline_list = [filter_hag, filter_ferry,filter_range]\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "with open(HAG_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4ab54586-461a-4e36-813e-ac22e597df44",
   "metadata": {},
   "source": [
    "### **Define Flight and Date"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "id": "8a018bdf-b543-4c57-8868-f1077947eadb",
   "metadata": {},
   "outputs": [],
   "source": [
    "flight = 'NCALM'\n",
    "date = '20080210'\n",
    "base_name = flight+'_SCB_'+date"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "id": "8be62742-e0d1-4f5b-b0ff-0632b0286250",
   "metadata": {},
   "outputs": [],
   "source": [
    "# parallel processing\n",
    "# time = 1.1 min\n",
    "tic = time.perf_counter()\n",
    "input_path = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/NAD83_NAD83_epoch2010/'\n",
    "output_path = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/HAG/'\n",
    "if __name__ == '__main__':\n",
    "    with concurrent.futures.ProcessPoolExecutor(max_workers=3) as executor:\n",
    "        onlyfiles = [f for f in os.listdir(input_path) if os.path.isfile(os.path.join(input_path, f))]\n",
    "        full_path = [input_path + s for s in onlyfiles]\n",
    "        output_path_full = [output_path  + s for s in onlyfiles]\n",
    "        executor.map(lidar_functions.HAG_dem, full_path, output_path_full)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a22bb40-5dd2-4a68-986c-582474a5cd67",
   "metadata": {},
   "source": [
    "RASTERIZE TEST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ab2dae61-1ed9-44df-9705-93a6a238626b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# parallel processing\n",
    "# time = 1.1 min\n",
    "tic = time.perf_counter()\n",
    "input_path = 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/HAG/'\n",
    "output_path = 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/rasterize_test_count/'\n",
    "if __name__ == '__main__':\n",
    "    with concurrent.futures.ProcessPoolExecutor(max_workers=3) as executor:\n",
    "        onlyfiles = [f for f in os.listdir(input_path) if os.path.isfile(os.path.join(input_path, f))]\n",
    "        full_path = [input_path + s for s in onlyfiles]\n",
    "        output_path_full = [output_path  + s for s in onlyfiles]\n",
    "        executor.map(lidar_functions.rasterize_count, full_path, output_path_full)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7178fbeb-cbe4-439e-aab9-57e2ec3084ef",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "7ad2a53a-72cd-4fe7-8c9f-fe68102cdbc5",
   "metadata": {},
   "source": [
    "# Vertical Bias Correction"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "926a5b5b-e820-473b-8016-c6844b1d40a0",
   "metadata": {},
   "source": [
    "## Move Files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "f5818571-a968-43f1-9514-6ef591da593a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# defining source and destination\n",
    "# paths\n",
    "\n",
    "src = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/HAG/*.la*'\n",
    "trg = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/hwy89_vertical_bias/target_lid'\n",
    "file_paths = glob.glob(src)\n",
    "\n",
    "for files in file_paths:\n",
    "    filename = os.path.basename(files)\n",
    "    if int(files[-18:-12]) > 737000 and int(files[-11:-4]) > 4370000:\n",
    "    # copying the files to the\n",
    "    # destination directory\n",
    "        shutil.copy2(files, trg)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8451906-bed0-43fb-be02-7a1f242dc0e0",
   "metadata": {},
   "source": [
    "### Merge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "78b38fda-766e-48b1-847f-94ae5a1c585c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# unnecessary - use function below\n",
    "# input_path = 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/hwy89_vertical_bias/target_lid/*.la*' # define path of input files\n",
    "# input_fname = glob.glob(input_path) # save to list\n",
    "# output_fname = 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/hwy89_vertical_bias/target_lid/ASO_SCB_20160326_hwy89_merge.'+input_fname[0][-3:]# set output filename\n",
    "# pdal_merge_command = input_fname\n",
    "# pdal_merge_command.insert(len(pdal_merge_command),output_fname) # insert output file to list\n",
    "# #pdal_merge_command.insert(0,'-f')\n",
    "# pdal_merge_command.insert(0,'merge')\n",
    "# pdal_merge_command.insert(0,'pdal')\n",
    "# subprocess.run(pdal_merge_command)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebdd7829-317e-42c3-8f2d-9a882d93c793",
   "metadata": {},
   "source": [
    "### Crop"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f35d800a-8876-447f-86a7-89e8c4997fc8",
   "metadata": {},
   "source": [
    "#### ASO 2016\n",
    "The 2016 data uses points overlapping a hwy89 polygon for vertical bias correction because it should have no snow. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "id": "d916c772-c2e7-4aba-82d8-997cd8ba09ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "# name JSON file\n",
    "clip_json = json_base_path +'clip_to_geometries.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "id": "5863300c-fc3f-4fc5-a875-b38f72ae50f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "path_to_shapefile = 'SCB/supporting_files/bounding_box/hwy89_poly_epsg6339.shp'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "id": "69d954e8-927c-460e-8c87-38577deb244c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# extract wkt from hwy 89 polygon\n",
    "hwy89 = shapefile.Reader(path_to_shapefile)\n",
    "geom=[]\n",
    "for s in hwy89.shapes():\n",
    "    geom.append(pygeoif.geometry.as_shape(s)) \n",
    "poly_base = pygeoif.MultiPolygon(geom)\n",
    "target_poly = poly_base.wkt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "id": "039a00c7-1565-4dca-9471-21c7b787ee37",
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_las_atPolygon(filenames, tags, target_poly, output_txt):\n",
    "    extract_las_atPolygon_json = json_base_path +'extract_las_atPolygon.json'\n",
    "    # merge all las files or stages\n",
    "    filter_merge = {\"type\":\"filters.merge\",\n",
    "                    \"tag\": \"merged\",\n",
    "                    \"inputs\": tags}\n",
    "    # crop\n",
    "    filter_crop = {'type':'filters.crop',\n",
    "                   'polygon':target_poly,\n",
    "                   'distance':1,\n",
    "                   'inputs':'merged',\n",
    "                   'tag': 'cropped'}\n",
    "    # write merged las to raster\n",
    "    writers_gdal = {\"type\":\"writers.text\",\n",
    "                    \"format\":\"csv\",\n",
    "                    \"order\":\"Z\",\n",
    "                    'keep_unspecified':False,\n",
    "                    'filename':output_txt}\n",
    "    # Append each stage to a list prior to saving to json \n",
    "    pipeline_list = filenames.copy()\n",
    "    pipeline_list.append(filter_merge)\n",
    "    pipeline_list.append(filter_crop)\n",
    "    pipeline_list.append(writers_gdal)\n",
    "\n",
    "    pipeline_dict = {'pipeline' : pipeline_list}\n",
    "    # save to json\n",
    "    with open(extract_las_atPolygon_json, 'w') as out:\n",
    "        json.dump(pipeline_dict, out, indent=4)\n",
    "    pdal_cmd = ['pdal','pipeline', extract_las_atPolygon_json]\n",
    "    subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "id": "20baaa64-9e5d-4746-8dac-cb73e50daa59",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/hwy89_vertical_bias/target_lid/'\n",
    "output_txt = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/hwy89_vertical_bias/clipped/'+base_name+'_hwy89_clip.txt'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)\n",
    "\n",
    "extract_las_atPolygon(filenames, tags, target_poly, output_txt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "71402413-e728-42b6-91cb-a000371da216",
   "metadata": {},
   "outputs": [],
   "source": [
    "# unnecessary\n",
    "# input_las = 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/hwy89_vertical_bias/target_lid/ASO_SCB_20160326_hwy89_merge.laz'\n",
    "# output_txt = 'SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/hwy89_vertical_bias/clipped/ASO_SCB_20160326_hwy89_clip.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "9d09b842-a504-4bab-8c81-8af8d202f3c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#unnecessary\n",
    "# pdal_commands = ['pdal', 'translate', input_las, output_txt, '--json', clip_json]\n",
    "# subprocess.run(pdal_commands)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5974506f-0990-414e-98db-ad79a9eb29e4",
   "metadata": {},
   "source": [
    "#### NCALM 2008\n",
    "The 2008 data uses snow-on field data extracted at points for vertical bias correction. <br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c7551b5d-a8a1-40eb-84f5-25e940002bc3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# name JSON file\n",
    "extract_las_atPoint_json = json_base_path +'extract_las_atPoint.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "56467508-027c-4fda-b448-bcbed112e845",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Read in ground_truthed data\n",
    "def csv_to_list(input_csv):\n",
    "    src = open(input_csv)\n",
    "    csvreader = csv.reader(src)\n",
    "    header = next(csvreader)\n",
    "    output_list = []\n",
    "    for row in csvreader:\n",
    "        output_list.append(row)\n",
    "    src.close()\n",
    "    return(header, output_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2fee3fb6-292f-4727-9bb2-f691722e7e72",
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_las_atPoint(filenames, tags, target_point, output_txt,distance):\n",
    "    # merge all las files or stages\n",
    "    filter_merge = {\"type\":\"filters.merge\",\n",
    "                    \"tag\": \"merged\",\n",
    "                    \"inputs\": tags}\n",
    "    # crop\n",
    "    filter_crop = {'type':'filters.crop',\n",
    "                   'point':target_point,\n",
    "                   'distance':distance,\n",
    "                   'inputs':'merged',\n",
    "                   'tag': 'cropped'}\n",
    "    # crop\n",
    "    filter_range = {'type':'filters.range',\n",
    "                   'limits':'Classification[2:2]',\n",
    "                   'inputs':'cropped',\n",
    "                   'tag': 'range'}\n",
    "    \n",
    "    # write merged las to raster\n",
    "    writers_gdal = {\"type\":\"writers.text\",\n",
    "                    \"format\":\"csv\",\n",
    "                    \"order\":\"Z\",\n",
    "                    \"write_header\":False,\n",
    "                    'keep_unspecified':False,\n",
    "                    'filename':output_txt}\n",
    "    # Append each stage to a list prior to saving to json \n",
    "    pipeline_list = filenames.copy()\n",
    "    pipeline_list.append(filter_merge)\n",
    "    pipeline_list.append(filter_crop)\n",
    "    pipeline_list.append(filter_range)\n",
    "    pipeline_list.append(writers_gdal)\n",
    "\n",
    "    pipeline_dict = {'pipeline' : pipeline_list}\n",
    "    # save to json\n",
    "    with open(extract_las_atPoint_json, 'w') as out:\n",
    "        json.dump(pipeline_dict, out, indent=4)\n",
    "    pdal_cmd = ['pdal','pipeline', extract_las_atPoint_json]\n",
    "    subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df4742ce-3bb6-45ad-ae59-855fa60c362d",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20080210/HAG/'\n",
    "output_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20080210/vertical_bias_correction/'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "dc0026e0-8c6f-44fe-b339-aa695bba261e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# extract data from field data\n",
    "SNOTEL_src = \"SCB/supporting_files/combined_field_snow/Harpold_data_paper/snowdepth_filtered.csv\"\n",
    "snotel_header, snotel_data = csv_to_list(SNOTEL_src)\n",
    "hunt_src = 'SCB/supporting_files/combined_field_snow/Huntingon_2008_snow/Hungington_2008_snow.csv'\n",
    "hunt_header, hunt_data = csv_to_list(hunt_src)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9efb5d7d-a0d8-4c35-bb52-b7723ad71e0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# for i in range(len(hunt_data)):\n",
    "#     for j in range(1,6):\n",
    "#         output_txt = output_path+'huntington_'+str(i)+'_dist'+str(j)+'.csv'\n",
    "#         target_point = 'POINT('+hunt_data[i][7]+' '+hunt_data[i][8]+')'\n",
    "#         extract_las_atPoint(filenames, tags, target_point, output_txt,j)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "63dde43b-606f-4d94-a50b-fc7135ae6f93",
   "metadata": {},
   "outputs": [],
   "source": [
    "# time = 4 min\n",
    "for i in range(len(snotel_data)):\n",
    "    for j in range(1,6):\n",
    "        output_txt = output_path+'snotel_'+str(i)+'_dist'+str(j)+'.csv'\n",
    "        target_point = 'POINT('+snotel_data[i][5]+' '+snotel_data[i][6]+')'\n",
    "        extract_las_atPoint(filenames, tags, target_point, output_txt,j)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3bbc45e7-d895-43f8-9aab-af78f1328c3a",
   "metadata": {},
   "source": [
    "## Calculate Stats"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8b75ae5-8299-468b-a120-67249d868a86",
   "metadata": {},
   "source": [
    "Applied\n",
    "Using 2014 NCALM Flight"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dca5384c-26c6-41bf-9688-1295e44054f6",
   "metadata": {},
   "source": [
    "### ASO 2016"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "id": "5d2617ee-a413-4c38-b404-5d8a6f0276d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "hwy89_arr = np.loadtxt('SCB/Sagehen_lidar/'+flight+'/'+base_name+'/hwy89_vertical_bias/clipped/'+base_name+'_hwy89_clip.txt',skiprows=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "2960560b-0428-448b-9030-e199946193f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ASO_20160326_hwy89_arr = np.loadtxt('SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/hwy89_vertical_bias/clipped/ASO_SCB_20160326_hwy89_clip.txt',skiprows=1)\n",
    "# ASO_20160417_hwy89_arr = np.loadtxt('SCB/Sagehen_lidar/ASO/ASO_SCB_20160417/hwy89_vertical_bias/clipped/ASO_SCB_20160417_hwy89_clip.csv',skiprows=1)\n",
    "# ASO_20160518_hwy89_arr = np.loadtxt('SCB/Sagehen_lidar/ASO/ASO_SCB_20160518/hwy89_vertical_bias/clipped/ASO_SCB_20160518_hwy89_clip.csv',skiprows=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "bbe0cd43-447f-40c8-9bd1-73711bfbe4cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ASO_20160326_hwy89_stats = lidar_functions.calculate_vertical_bias_over_snowOff('SCB/Sagehen_lidar/ASO/ASO_SCB_20160326/hwy89_vertical_bias/clipped/ASO_SCB_20160326_hwy89_clip.laz')\n",
    "# ASO_20160417_hwy89_stats = lidar_functions.calculate_vertical_bias_over_snowOff('SCB/Sagehen_lidar/ASO/ASO_SCB_20160417/hwy89_vertical_bias/clipped/ASO_SCB_20160417_hwy89_clip.laz')\n",
    "# ASO_20160518_hwy89_stats = lidar_functions.calculate_vertical_bias_over_snowOff('SCB/Sagehen_lidar/ASO/ASO_SCB_20160518/hwy89_vertical_bias/clipped/ASO_SCB_20160518_hwy89_clip.las')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c724eab8-e366-4c29-a3d0-33d69b753df8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# NCALM_2014_hwy89_stats = lidar_functions.calculate_vertical_bias('SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/hwy89_vertical_bias/clipped/NCALM_SCB_2014_hwy89_clip.las')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "id": "a4523ec8-1eb6-427b-ad3a-9f152fbf902f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.45\n"
     ]
    }
   ],
   "source": [
    "print(np.nanpercentile((hwy89_arr),50))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "37e85b56-2fbf-4b1d-a18d-f0c36bd90ab7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.21\n",
      "0.29\n",
      "0.37\n"
     ]
    }
   ],
   "source": [
    "print(np.nanpercentile((ASO_20160326_hwy89_arr),10))\n",
    "print(np.nanpercentile((ASO_20160326_hwy89_arr),50))\n",
    "print(np.nanpercentile((ASO_20160326_hwy89_arr),90))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "aeaaa728-27ac-4ded-974b-f9d88fa65fcd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.27\n",
      "0.35\n"
     ]
    }
   ],
   "source": [
    "print(np.nanpercentile((ASO_20160417_hwy89_arr),10))\n",
    "print(np.nanpercentile((ASO_20160417_hwy89_arr),50))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4bd95dde-4dd2-4cef-a705-921f1403e148",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.37\n",
      "0.44\n"
     ]
    }
   ],
   "source": [
    "print(np.nanpercentile((ASO_20160518_hwy89_arr),10))\n",
    "print(np.nanpercentile((ASO_20160518_hwy89_arr),50))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "2c00b8cd-d1e1-4f4d-baa0-cb7d58b4c083",
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.hist(ASO_20160326_hwy89_arr-0.29, bins=40)\n",
    "# plt.axvline(0, color='k')\n",
    "# #plt.xlim([-.2, 0.75])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6670984d-73d8-462e-b036-31e6d3dd2015",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(ASO_20160326_hwy89_stats)\n",
    "print(ASO_20160417_hwy89_stats)\n",
    "print(ASO_20160518_hwy89_stats)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f456a359-3004-44c3-b3a5-e2cea8e8d92f",
   "metadata": {},
   "source": [
    "### **NCALM 2008**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "3a31fa74-0489-4f75-9ddc-2801b3bfa425",
   "metadata": {},
   "outputs": [],
   "source": [
    "# input_las_txt = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20080210/SNOTEL_vertical_bias/clipped/snotel_541.csv'\n",
    "input_las_txt = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20080210/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b32fe203-316a-4976-81e5-6d41d5e8730b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# hag_arr = np.loadtxt(input_las_txt,skiprows=1)\n",
    "# lowest_10th_per = np.nanpercentile(hag_arr, 10)\n",
    "# mean_hag = np.nanmean(hag_arr)\n",
    "# median_hag = np.nanmedian(hag_arr)\n",
    "# stats_539 = [\"lowest_10th\",lowest_10th_per, \"mean\",mean_hag, \"median\", median_hag]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "2d0b2900-2d7f-41d6-a065-3cdaeb8e8bed",
   "metadata": {},
   "outputs": [],
   "source": [
    "NCALM_2008_vbc_dict = {}\n",
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_20080210/vertical_bias_correction/ground_pts/'\n",
    "onlyfiles = [f for f in os.listdir(input_path) if os.path.isfile(os.path.join(input_path, f))] # make a list of all filenames in directory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "cf21b891-2b16-4b8b-b0a5-209888375c3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "NCALM_2008_vbc_dict = {}\n",
    "all_files = glob.glob('SCB/Sagehen_lidar/NCALM/NCALM_SCB_20080210/vertical_bias_correction/ground_pts/*')\n",
    "for files in onlyfiles:\n",
    "    txt_path = input_path + files\n",
    "    txt_arr = np.loadtxt(txt_path)\n",
    "    NCALM_2008_vbc_dict[files[:-4]] = txt_arr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "03e1b4d9-54b2-4c4c-9a71-8fc47f317a83",
   "metadata": {},
   "outputs": [],
   "source": [
    "NCALM_2008_error_dict = {}\n",
    "keys_list = list(NCALM_2008_vbc_dict.keys())\n",
    "for i in range(len(keys_list)):\n",
    "    index = int(re.findall(r'\\d+',  keys_list[i])[0])\n",
    "    if keys_list[i][0:3]=='hun':\n",
    "        NCALM_2008_error_dict[keys_list[i]] = float(hunt_data[index][4]) - NCALM_2008_vbc_dict[keys_list[i]]\n",
    "    elif keys_list[i][0:3]=='sno':\n",
    "        NCALM_2008_error_dict[keys_list[i]] = float(snotel_data[index][4])/100 - NCALM_2008_vbc_dict[keys_list[i]]\n",
    "    else:\n",
    "        print('error')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "26b17132-f15b-49ff-ba08-3d15090f8cd0",
   "metadata": {},
   "outputs": [],
   "source": [
    "NCALM_2008_vbc_dict[keys_list[19]] = np.array([0.8])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "8c6911fc-8501-4942-98d8-e90a2235983c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'snotel_2_dist5'"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "keys_list[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "4b6dc373-2439-4cb9-98ac-0d1532af3d99",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x720 with 50 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "NCALM_2008_error_plot, axs = plt.subplots(5, 10, figsize =(30,10));\n",
    "#NCALM_2008_error_plot.suptitle('NCALM 2008 Lidar Depth Distributions Around Ground Truth') # add title to entire figure\n",
    "\n",
    "k = 0\n",
    "for i in range(10):\n",
    "    for j in range(5):\n",
    "        keys_list = list(NCALM_2008_vbc_dict.keys())\n",
    "        index = int(re.findall(r'\\d+',  keys_list[k])[0])\n",
    "        true_val = float(hunt_data[index][4])\n",
    "        g = sns.histplot(ax = axs[j,i], data = NCALM_2008_vbc_dict[keys_list[k]])\n",
    "        med = np.median(NCALM_2008_vbc_dict[keys_list[k]])\n",
    "        g.axvline(true_val, color='red')\n",
    "        g.axvline(med, color = 'black')\n",
    "        y_lab = str(j+1)+'m'\n",
    "        x_lab = \"Point \" +str(i)\n",
    "        axs[j,i].set(yticklabels = [], title=' ',xlabel = '',ylabel ='')\n",
    "        axs[0,0].set_ylabel('R = 5m', fontsize=16); axs[1,0].set_ylabel('R = 4m',fontsize=16);axs[2,0].set_ylabel('R = 3m',fontsize=16);axs[3,0].set_ylabel('R = 2m',fontsize=16);axs[4,0].set_ylabel('R = 1m',fontsize=16)\n",
    "        k = k+1\n",
    "#NCALM_2008_error_plot.savefig('SCB/stats/figures/vertical_bias_correction_2008.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "11b62a20-03a4-4e5a-be51-1e36ea255b67",
   "metadata": {},
   "outputs": [],
   "source": [
    "NCALM_2008_error_plot, axs = plt.subplots(5, 11, figsize =(30,10));\n",
    "#NCALM_2008_error_plot.suptitle('NCALM 2008 Lidar Depth Distributions Around Ground Truth') # add title to entire figure\n",
    "\n",
    "k = 0\n",
    "for i in range(11):\n",
    "    for j in range(5):\n",
    "        keys_list = list(NCALM_2008_error_dict.keys())\n",
    "        index = int(re.findall(r'\\d+',  keys_list[k])[0])\n",
    "        #true_val = float(hunt_data[index][4])\n",
    "        g = sns.histplot(ax = axs[j,i], data = NCALM_2008_error_dict[keys_list[k]])\n",
    "        med = np.median(NCALM_2008_error_dict[keys_list[k]])\n",
    "        # g.axvline(true_val, color='red')\n",
    "        g.axvline(med, color = 'green')\n",
    "        y_lab = str(j+1)+'m'\n",
    "        x_lab = \"Point \" +str(i)\n",
    "        axs[j,i].set(yticklabels = [], title=' ',xlabel = '',ylabel ='')\n",
    "        axs[0,0].set_ylabel('Radius = 1m'); axs[1,0].set_ylabel('Radius = 2m');axs[2,0].set_ylabel('Radius = 3m');axs[3,0].set_ylabel('Radius = 4m');axs[4,0].set_ylabel('Radius = 5m')\n",
    "        k = k+1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "24815514-66ef-4dc9-9fc3-0e2506db0bff",
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_error(dist):\n",
    "    error_med = []\n",
    "    keys_list = list(NCALM_2008_error_dict.keys())\n",
    "    for k in range(int(len(NCALM_2008_error_dict))):\n",
    "        if int(re.findall(r'\\d+',  keys_list[k])[1])==dist:\n",
    "            error_med.extend(NCALM_2008_error_dict[keys_list[k]])\n",
    "    return(error_med)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "id": "26b9cadc-b444-4478-850f-216c1bbadc44",
   "metadata": {},
   "outputs": [],
   "source": [
    "error_med_2 = calculate_error(2)\n",
    "error_med_3 = calculate_error(3)\n",
    "error_med_4 = calculate_error(4)\n",
    "error_med_5 = calculate_error(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "id": "f5acebaa-af2b-4fb8-a063-2407bf1c3083",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Radius = 2m, median error =  -0.07620000000000005 stdev =  0.3939059976952006\n",
      "Radius = 3m, median error =  -0.04640000000000022 stdev =  0.38573160057604067\n",
      "Radius = 4m, median error =  -0.03958333299999994 stdev =  0.37924952640387316\n",
      "Radius = 5m, median error =  -0.029200000000000115 stdev =  0.3819159572093684\n"
     ]
    }
   ],
   "source": [
    "print(\"Radius = 2m, median error = \",np.median(error_med_2),'stdev = ',np.std(error_med_2))\n",
    "print(\"Radius = 3m, median error = \",np.median(error_med_3),'stdev = ',np.std(error_med_3))\n",
    "print(\"Radius = 4m, median error = \",np.median(error_med_4),'stdev = ',np.std(error_med_4))\n",
    "print(\"Radius = 5m, median error = \",np.median(error_med_5),'stdev = ',np.std(error_med_5))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c8d404e-911a-431b-82b4-a3c3c84d70b4",
   "metadata": {},
   "source": [
    "hold off on corrections for now"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "91e5bcbb-8274-4847-b129-d1c92bb3cb58",
   "metadata": {},
   "source": [
    "## Correct and Rasterize\n",
    "In this case we know that the May flight has the most limited extent. One way to check for this would be to use a similar code to "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "a60dcefd-7390-4041-844c-da5c58049da4",
   "metadata": {},
   "outputs": [],
   "source": [
    "#correct_merge_rasterize_json = json_base_path + 'correct_merge_rasterize_Zlim_3.json'\n",
    "correct_merge_rasterize_json = json_base_path + 'correct_merge_rasterize.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "id": "ea3c9439-5269-4964-94e3-40a61bb02147",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/HAG/'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)\n",
    "output_tif = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/corrected_tif/'+base_name+'_vbc_5.tif'\n",
    "target_assign = 'Z-'+str(0.03) #str(np.nanpercentile((hwy89_arr),50))\n",
    "zlim = 'Z[-0.3:5]'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62d93f1b-301e-4406-b58f-c8dbcf62cf2a",
   "metadata": {},
   "source": [
    "Create the reader stages of the pipeline. Each file is read as an individual reader stage here, and we cheat here by copying the formatting of json files and creating a dictionary with the values as the correctly formatted inputs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "28fec394-70ca-404f-9fb5-67b32e934ed7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# # [730235.96, 738826.45, 4364741.66, 4372273.16, 8590.48999999999, 7531.5]\n",
    "# writers_bounds = '(['+ str(ASO_SCB_20160518_ext[0])+','+ str(ASO_SCB_20160518_ext[1])+'],['+str(ASO_SCB_20160518_ext[2])+','+str(ASO_SCB_20160518_ext[3])+'])'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "id": "822e3cdd-58ac-48a0-b748-663ea519090d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def correct_and_rasterize(filenames, tags, target_assign,zlim,writers_bounds,output_tif):# merge all las files or stages\n",
    "    correct_merge_rasterize_json = json_base_path + 'correct_merge_rasterize.json'\n",
    "    filter_merge = {\"type\":\"filters.merge\",\n",
    "                   \"tag\": \"merged\",\n",
    "                   \"inputs\": tags}\n",
    "    # filter out the ground points of the tiles\n",
    "    filter_assign = {'type': 'filters.assign',\n",
    "                     'value':\"Z=\"+target_assign,\n",
    "                     'inputs':'merged',\n",
    "                     'tag':'corrected'}\n",
    "    # filter out the ground points of the tiles\n",
    "    # filter_range = {'type': 'filters.range',\n",
    "    #                  'limits':\"Z[0.15:5]\",\n",
    "    #                  'inputs':'corrected',\n",
    "    #                  'tag':'ranged'}\n",
    "    filter_range = {'type': 'filters.range',\n",
    "                     'limits':zlim,\n",
    "                     'inputs':'corrected',\n",
    "                     'tag':'ranged'}\n",
    "    # write merged las to raster\n",
    "    writers_gdal= {\"type\": \"writers.gdal\",\n",
    "                   'output_type': 'mean',\n",
    "                  'resolution': '1.0',\n",
    "                  'radius': '0.7',\n",
    "                   'override_srs':'EPSG:6339',\n",
    "                   'bounds': writers_bounds,\n",
    "                   'inputs': 'ranged',\n",
    "                   'filename':output_tif}\n",
    "    # Append each stage to a list prior to saving to json \n",
    "    pipeline_list = filenames.copy()\n",
    "    pipeline_list.append(filter_merge)\n",
    "    pipeline_list.append(filter_assign)\n",
    "    pipeline_list.append(filter_range)\n",
    "    pipeline_list.append(writers_gdal)\n",
    "\n",
    "    pipeline_dict = {'pipeline' : pipeline_list}\n",
    "    # save to json\n",
    "    with open(correct_merge_rasterize_json, 'w') as out:\n",
    "        json.dump(pipeline_dict, out, indent=4)\n",
    "        # time = ~3 min\n",
    "    pdal_cmd = ['pdal','pipeline', correct_merge_rasterize_json]\n",
    "    subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "id": "4b31e155-87fd-406c-925d-bfdedd9fe3e7",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/HAG/'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)\n",
    "output_tif = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/corrected_tif/'+base_name+'_vbc_5.tif'\n",
    "target_assign = 'Z-'+str(0.03)\n",
    "zlim = 'Z[-0.3:5]'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "id": "ff10f019-778c-489e-b5b3-8286b8bd0dfd",
   "metadata": {},
   "outputs": [],
   "source": [
    "correct_and_rasterize(filenames, tags, target_assign, zlim, writers_bounds, output_tif)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "id": "1a46bda4-7c28-4259-8047-af8c9ece72f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/HAG/'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)\n",
    "output_tif = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/corrected_tif/'+base_name+'_vbc_3.tif'\n",
    "target_assign = 'Z-'+str(0.03)#str(np.nanpercentile((hwy89_arr),50))\n",
    "zlim = 'Z[-0.3:3]'\n",
    "correct_and_rasterize(filenames, tags, target_assign, zlim, writers_bounds, output_tif)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "dd087e75-5505-4400-8d42-40872a90fb12",
   "metadata": {},
   "outputs": [],
   "source": [
    "correct_and_rasterize(filenames, tags, target_assign, zlim, writers_bounds, output_tif)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "id": "bbdca192-f85b-424d-952a-203dac20b7f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/HAG/'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)\n",
    "output_tif = 'SCB/Sagehen_lidar/'+flight+'/'+base_name+'/corrected_tif/'+base_name+'_vbc.tif'\n",
    "target_assign = 'Z-'+str(0.03)#+str(np.nanpercentile((hwy89_arr),50))\n",
    "zlim = 'Z[-0.3:70]'\n",
    "correct_and_rasterize(filenames, tags, target_assign, zlim, writers_bounds, output_tif)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "b37aec8c-4285-4fee-81c0-108c5db1a819",
   "metadata": {},
   "outputs": [],
   "source": [
    "correct_and_rasterize(filenames, tags, target_assign, zlim, writers_bounds, output_tif)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6732e61a-1663-4cb3-be01-cba704dbadd3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "bb980f62-9d55-40a9-99cc-1136f519df41",
   "metadata": {},
   "outputs": [],
   "source": [
    "# # time = ~3 min\n",
    "# pdal_cmd = ['pdal','pipeline', correct_merge_rasterize_json]\n",
    "# subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b466837-460a-4f8b-9202-f81a226fa651",
   "metadata": {},
   "source": [
    "## Snow-Off Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aa814a07-3c70-4bc8-b6d3-3b35f3336c30",
   "metadata": {},
   "source": [
    "### Vegetation Strata"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7cba995-877a-4a96-a828-784ec57aade0",
   "metadata": {},
   "source": [
    "**Combined Pipeline Method**\n",
    "We'll use the same bounds as above"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "106d03d8-7a0f-4f32-9f9d-6848384eca22",
   "metadata": {},
   "outputs": [],
   "source": [
    "# This function allows us to change the json pipeline depending on the files\n",
    "# the final output of a subprocess that uses this pipeline will be one raster with 1x1m pixels, each containing the number of returns (counts) of the specified lidar height strata\n",
    "# e.g. if the range_limit is Z[-0.15:0.15), each pixel will give us the number of returns between -0.15 and 0.15 in elevation (HAG in this case)\n",
    "# The goal of this function is to return a json pipeline with specific Z limits\n",
    "# input: filenames - see: lidar_functions.create_command_template\n",
    "# input: target_assign - if you need to perform a vertical bias correction on the snow-off flight, insert a string with the correction function (e.g. 'Z+0.05' if you need to add 0.05 to each lidar point), if no correction is necessary\n",
    "# insert \"Z+0\" [str]\n",
    "# input: writers_bounds - '([minx, maxx], [miny, maxy])';  e.g.'([730235.96, 738826.45], [4364741.66, 4372273.16])' [str]\n",
    "# input: range_limits: Z limits for output file\n",
    "def create_vegetation_heigh_strata(filenames, tags, target_assign, writers_bounds, range_limits, output_tif): \n",
    "    # merge all las files or stages\n",
    "    filter_merge = {\"type\":\"filters.merge\",\n",
    "                   \"tag\": \"merged\",\n",
    "                   \"inputs\": tags}\n",
    "    # bias correct files\n",
    "    filter_assign = {'type': 'filters.assign',\n",
    "                     'value':\"Z=\"+target_assign,\n",
    "                     'inputs':'merged',\n",
    "                     'tag':'corrected'}\n",
    "    # filter out the ground points of the tiles\n",
    "    filter_range = {\"type\":\"filters.range\",\n",
    "                    \"limits\":range_limits,\n",
    "                   'inputs':'corrected',\n",
    "                   'tag':'filtered'}\n",
    "    # write merged las to raster\n",
    "    writers_gdal= {\"type\": \"writers.gdal\",\n",
    "                   'output_type': 'count',\n",
    "                  'resolution': '1.0',\n",
    "                  'radius': '0.7',\n",
    "                   'bounds': writers_bounds,\n",
    "                   'filename':output_tif}\n",
    "    # Append each stage to a list prior to saving to json \n",
    "    pipeline_list = filenames.copy()\n",
    "    pipeline_list.append(filter_merge)\n",
    "    pipeline_list.append(filter_assign)\n",
    "    pipeline_list.append(filter_range)\n",
    "    pipeline_list.append(writers_gdal)\n",
    "\n",
    "    pipeline_dict = {'pipeline' : pipeline_list}\n",
    "    return(pipeline_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "af094e83-88cd-46f4-801d-7f31c2bbbd05",
   "metadata": {},
   "outputs": [],
   "source": [
    "writers_bounds = '(['+ str(ASO_SCB_20160518_ext[0])+','+ str(ASO_SCB_20160518_ext[1])+'],['+str(ASO_SCB_20160518_ext[2])+','+str(ASO_SCB_20160518_ext[3])+'])'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06a4bd3a-451f-4066-b6f2-4fe6bbfebc13",
   "metadata": {},
   "outputs": [],
   "source": [
    "# '([730235.96,738826.45],[4364741.66,4372273.16])'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0ee62682-e968-4798-93b2-3151092624d9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c7625457-52e8-4ee8-93c2-29f219c43bca",
   "metadata": {},
   "source": [
    "### Vegetation Strata"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52fec196-b0ae-4cd3-95ea-1f7e2f368558",
   "metadata": {},
   "source": [
    "**Combined Pipeline Method**\n",
    "Vegetation Height Strata:<br>\n",
    "[-0.15:0.15), used to determine open sites<br>\n",
    "[0.15:1.5), if there are returns in this region, we can look into weather there is short vegetation or understory here<br>\n",
    "[1.5:3), we remove all pixels with returns here, classified as \"low branches or significant shrub/grass/ground-veg\"<br>\n",
    "[3:), this is our threshold for \"tall\" vegetation<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f42c572c-a917-4b43-bf87-575537ef0427",
   "metadata": {},
   "source": [
    "Constant inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "d0e39095-c393-48d9-b76e-7a986f06b7eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "# this remains the same for all heigh strata\n",
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/HAG/'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)\n",
    "target_assign = \"Z+0.08\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b39e087c-2054-4c61-8d10-73a735264156",
   "metadata": {},
   "source": [
    "Variable inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c69a57ef-861f-400e-8958-445fb3200909",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "e8996656-7d05-47a0-a523-569fce88891c",
   "metadata": {},
   "source": [
    "[-0.15:0.15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "5f04f5bd-3c99-45fb-8c0b-9eebf7c28d4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# specify name of the json pipeline file\n",
    "filterMergeRasterize_json = json_base_path+'filterMergeRasterize_neg0pt15_0pt15.json'\n",
    "# specify output tif file path\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/canopy_metrics/veg_height_strata/vegStrata_neg0pt15_0pt1.tif'\n",
    "# heigh strata range limits\n",
    "range_limits = 'Z[-0.15:0.15)'\n",
    "# create pipline\n",
    "pipeline_dict = create_vegetation_heigh_strata(filenames, tags, target_assign, writers_bounds, range_limits, output_tif)\n",
    "# save to json\n",
    "with open(filterMergeRasterize_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "29b3eb18-55a0-4e85-be82-e9cc2b169a7a",
   "metadata": {},
   "outputs": [],
   "source": [
    "tic = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "e7ef168d-a7ff-4de0-9453-039025456434",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CompletedProcess(args=['pdal', 'pipeline', 'piske_processing/PDAL_workflow/JSON/filterMergeRasterize_neg0pt15_0pt15.json'], returncode=0)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pdal_cmd = ['pdal','pipeline', filterMergeRasterize_json]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "d7b0af17-b3f0-41fb-893b-6a49242130f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "bf8bedaa-a7fa-43c1-aa7e-84f8316dfb73",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "16.56809745"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(toc-tic)/60"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94fb9866-b4df-4611-80b9-4d99ab207dc9",
   "metadata": {},
   "source": [
    "[0.15:1.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "43e4db18-72fd-49f6-a556-765ef21865ff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CompletedProcess(args=['pdal', 'pipeline', 'piske_processing/PDAL_workflow/JSON/filterMergeRasterize_0pt15_1pt5.json'], returncode=0)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# specify name of the json pipeline file\n",
    "filterMergeRasterize_json = json_base_path+'filterMergeRasterize_0pt15_1pt5.json'\n",
    "# specify output tif file path\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/canopy_metrics/veg_height_strata/vegStrata_0pt15_1pt5.tif'\n",
    "# heigh strata range limits\n",
    "range_limits = 'Z[0.15:1.5)'\n",
    "# create pipline\n",
    "pipeline_dict = create_vegetation_heigh_strata(filenames, tags, target_assign, writers_bounds, range_limits, output_tif)\n",
    "# save to json\n",
    "with open(filterMergeRasterize_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n",
    "pdal_cmd = ['pdal','pipeline', filterMergeRasterize_json]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b03b9c9c-9c3b-453c-b7cf-00dfbe5767b3",
   "metadata": {},
   "source": [
    "[1.5:3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "6ae2bfd5-e9a2-4222-869d-5a82551a408a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CompletedProcess(args=['pdal', 'pipeline', 'piske_processing/PDAL_workflow/JSON/filterMergeRasterize_1pt5_3.json'], returncode=0)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# specify name of the json pipeline file\n",
    "filterMergeRasterize_json = json_base_path+'filterMergeRasterize_1pt5_3.json'\n",
    "# specify output tif file path\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/canopy_metrics/veg_height_strata/vegStrata_1pt5_3.tif'\n",
    "# heigh strata range limits\n",
    "range_limits = 'Z[1.5:3)'\n",
    "# create pipline\n",
    "pipeline_dict = create_vegetation_heigh_strata(filenames, tags, target_assign, writers_bounds, range_limits, output_tif)\n",
    "# save to json\n",
    "with open(filterMergeRasterize_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n",
    "pdal_cmd = ['pdal','pipeline', filterMergeRasterize_json]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec7ba3e0-0944-4aa3-810e-913c6da37d4e",
   "metadata": {},
   "source": [
    "[3:)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "3ecfea3a-f28c-4e33-9a1b-53b1c76bc2f7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CompletedProcess(args=['pdal', 'pipeline', 'piske_processing/PDAL_workflow/JSON/filterMergeRasterize_3.json'], returncode=0)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# specify name of the json pipeline file\n",
    "# this remains the same for all heigh strata\n",
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/HAG/'\n",
    "filenames, tags = lidar_functions.create_command_template(input_path)\n",
    "target_assign = \"Z+0.08\"\n",
    "\n",
    "filterMergeRasterize_json = json_base_path+'filterMergeRasterize_3.json'\n",
    "# specify output tif file path\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/canopy_metrics/veg_height_strata/vegStrata_3.tif'\n",
    "# heigh strata range limits\n",
    "range_limits = 'Z[3:)'\n",
    "# create pipline\n",
    "pipeline_dict = create_vegetation_heigh_strata(filenames, tags, target_assign, writers_bounds, range_limits, output_tif)\n",
    "# save to json\n",
    "with open(filterMergeRasterize_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n",
    "pdal_cmd = ['pdal','pipeline', filterMergeRasterize_json]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c86f8358-6e91-4889-b3bc-e1300bea86f0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a6eee961-41c0-47e1-b6b3-f7a492816548",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c4a0bf40-98c0-423a-8b6a-a1a587782508",
   "metadata": {},
   "source": [
    "### Create CHM"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "747e3961-714c-49fd-bbd5-022c07a0b8d6",
   "metadata": {},
   "source": [
    "Note that there are two CHMs for these data. One is based off all points, another is limited to possible short/understory vegetation (<1.5m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "a3107c9e-e6db-46b8-96f6-d10b3a5d49ab",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "CHM_json_LT1pt5 = json_base_path+'CHM_from_las_LT1pt5.json' # define path to json files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8b1a911e-a30c-47d0-b961-4d5b8818a145",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/HAG/'\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/CHM/NCALM_2014_CHM_LT1pt5.tif'\n",
    "onlyfiles = [f for f in os.listdir(input_path) if os.path.isfile(os.path.join(input_path, f))]\n",
    "input_list = [input_path + s for s in onlyfiles]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ae874a5e-2129-4105-98da-fae42d42026a",
   "metadata": {},
   "outputs": [],
   "source": [
    "filename_dict = {}\n",
    "tags = ['']*len(input_list)\n",
    "filenames = ['']*len(input_list)\n",
    "for i in range(len(input_list)):\n",
    "    filename_dict['filename_'+str(i)] = {'filename':input_list[i], 'tag':'A_'+str(i)}\n",
    "    tags[i] = 'A_'+str(i)\n",
    "    filenames[i] = filename_dict[list(filename_dict)[i]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93c5a436-3444-4b72-8fd0-bb837b12087d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge all las files or stages\n",
    "filter_merge = {\"type\":\"filters.merge\",\n",
    "               \"tag\": \"merged\",\n",
    "               \"inputs\": tags}\n",
    "filter_return = {'type': 'filters.returns',\n",
    "                'groups': 'first',\n",
    "                'inputs':'merged',\n",
    "                'tag':'filtered'}\n",
    "filter_range = {'type': 'filters.range',\n",
    "                'limits': 'Z[:1.5)',\n",
    "                'inputs':'merged',\n",
    "                'tag':'ranged'}\n",
    "# write merged las to raster\n",
    "writers_gdal= {\"type\": \"writers.gdal\",\n",
    "               'output_type': 'max', \n",
    "              'resolution': '1.0',\n",
    "               'bounds': writers_bounds,\n",
    "               'inputs':'ranged',\n",
    "              'radius': '0.7',\n",
    "               'filename': output_tif}\n",
    "# Append each stage to a list prior to saving to json \n",
    "pipeline_list = filenames.copy()\n",
    "pipeline_list.append(filter_merge)\n",
    "pipeline_list.append(filter_range)\n",
    "pipeline_list.append(writers_gdal)\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "# save to json\n",
    "with open(CHM_json_LT1pt5, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "762b779c-8d65-4b9d-ab1b-240eb55fcbeb",
   "metadata": {},
   "outputs": [],
   "source": [
    "tic = time.perf_counter()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7c666f24-f4ec-419d-b056-b8e3c3557ab3",
   "metadata": {},
   "outputs": [],
   "source": [
    "pdal_cmd = ['pdal','pipeline', CHM_json_LT1pt5]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "872d681c-e0ab-4278-863d-04927bb403e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5fa43794-5cd5-4afd-8a32-a06882b7d96a",
   "metadata": {},
   "outputs": [],
   "source": [
    "(toc-tic)/60"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b2a21a4-d5cf-481b-a971-ce58e982ffbb",
   "metadata": {},
   "source": [
    "All canopy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5a34d001-e117-4740-9a6f-ad6c811adedf",
   "metadata": {},
   "outputs": [],
   "source": [
    "CHM_json = json_base_path+'CHM_from_las.json' # define path to json files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c1451584-cc2a-48c3-94d1-5fa6a53122f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/HAG/'\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/canopy_metrics/veg_structure/CHM/NCALM_2014_CHM.tif'\n",
    "onlyfiles = [f for f in os.listdir(input_path) if os.path.isfile(os.path.join(input_path, f))]\n",
    "input_list = [input_path + s for s in onlyfiles]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "7053f8b6-9b61-458b-8ed7-cd8650a4aa4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "filename_dict = {}\n",
    "tags = ['']*len(input_list)\n",
    "filenames = ['']*len(input_list)\n",
    "for i in range(len(input_list)):\n",
    "    filename_dict['filename_'+str(i)] = {'filename':input_list[i], 'tag':'A_'+str(i)}\n",
    "    tags[i] = 'A_'+str(i)\n",
    "    filenames[i] = filename_dict[list(filename_dict)[i]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "779ae682-29ae-4bf7-94ba-b9b85ed4cf1f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge all las files or stages\n",
    "filter_merge = {\"type\":\"filters.merge\",\n",
    "               \"tag\": \"merged\",\n",
    "               \"inputs\": tags}\n",
    "filter_range = {'type': 'filters.range',\n",
    "                'limits': 'Z[-0.15:70]',\n",
    "                'inputs':'merged',\n",
    "                'tag':'ranged'}\n",
    "# write merged las to raster\n",
    "writers_gdal= {\"type\": \"writers.gdal\",\n",
    "               'output_type': 'max', \n",
    "              'resolution': '1.0',\n",
    "               'bounds': writers_bounds,\n",
    "               'inputs':'ranged',\n",
    "              'radius': '0.7',\n",
    "               'filename': output_tif}\n",
    "# Append each stage to a list prior to saving to json \n",
    "pipeline_list = filenames.copy()\n",
    "pipeline_list.append(filter_merge)\n",
    "pipeline_list.append(filter_range)\n",
    "pipeline_list.append(writers_gdal)\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "# save to json\n",
    "with open(CHM_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7c31561e-d32d-470c-91b3-45dfdc084076",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CompletedProcess(args=['pdal', 'pipeline', 'piske_processing/PDAL_workflow/JSON/CHM_from_las.json'], returncode=0)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pdal_cmd = ['pdal','pipeline', CHM_json]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8cf4ef0-37c4-4b64-bd85-1b7a448426e5",
   "metadata": {},
   "source": [
    "### Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9b46a254-f10a-4643-9078-a15a67ac7d50",
   "metadata": {},
   "outputs": [],
   "source": [
    "litree_json = json_base_path+'litree.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c73688b4-dc87-440a-8e8d-4eccd64e3889",
   "metadata": {},
   "outputs": [],
   "source": [
    "readers_las = {'type':\"readers.las\",\n",
    "              \"filename\":'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/NAD83_NAD83_epoch2010/NCALM_SCB_2014_733000_4365000.las'}\n",
    "filter_ferry = {\"type\":\"filters.ferry\", \n",
    "                \"dimensions\":\"Z => HeightAboveGround\"}\n",
    "filter_sort = {\"type\":\"filters.sort\", \n",
    "               \"dimension\":\"HeightAboveGround\", \n",
    "               \"order\":\"DESC\"}\n",
    "filter_litree = {\"type\":\"filters.litree\"}\n",
    "writers_gdal= {\"type\": \"writers.gdal\",\n",
    "               'output_type': 'mean', \n",
    "              'resolution': '1.0',\n",
    "               'dimension':'ClusterID',\n",
    "              'radius': '0.7',\n",
    "               'filename': 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/canopy_metrics/veg_structure/cluster/test.tif'}\n",
    "pipeline_list = [readers_las, filter_ferry, filter_sort, filter_litree, writers_gdal]\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "# save to json\n",
    "with open(litree_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e231b35b-d820-4c9a-88f8-c1021e563d6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b304de03-0837-4a24-8847-82d093bd05b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "pdal_cmd = ['pdal','pipeline', litree_json]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6fe69ddb-f958-430a-82a4-3c6005ea36fd",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "385081df-634d-493f-8772-cdd0831c436e",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "cluster_json = json_base_path+'cluster_from_las.json' # define path to json files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0d98de70-acbf-4676-b0cf-5b55a69b04e5",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_path = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/HAG/'\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/canopy_metrics/veg_structure/NCALM_SCB_2014_cluster_test.tif'\n",
    "onlyfiles = [f for f in os.listdir(input_path) if os.path.isfile(os.path.join(input_path, f))]\n",
    "input_list = [input_path + s for s in onlyfiles]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "23f18635-a4ec-4667-bcbc-bb904a6558bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "filename_dict = {}\n",
    "tags = ['']*len(input_list)\n",
    "filenames = ['']*len(input_list)\n",
    "for i in range(len(input_list)):\n",
    "    filename_dict['filename_'+str(i)] = {'filename':input_list[i], 'tag':'A_'+str(i)}\n",
    "    tags[i] = 'A_'+str(i)\n",
    "    filenames[i] = filename_dict[list(filename_dict)[i]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "774e20ed-6499-428e-bcf5-9ca0af3fead6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge all las files or stages\n",
    "filter_merge = {\"type\":\"filters.merge\",\n",
    "               \"tag\": \"merged\",\n",
    "               \"inputs\": tags}\n",
    "filter_cluster = {'type': 'filters.cluster',\n",
    "                'inputs':'merged',\n",
    "                'tag':'cluster'}\n",
    "# write merged las to raster\n",
    "writers_gdal= {\"type\": \"writers.gdal\",\n",
    "               'output_type': 'mean', \n",
    "              'resolution': '1.0',\n",
    "               'bounds': writers_bounds,\n",
    "               'dimension':'ClusterID',\n",
    "              'radius': '0.7',\n",
    "               'filename': output_tif}\n",
    "# Append each stage to a list prior to saving to json \n",
    "pipeline_list = filenames.copy()\n",
    "pipeline_list.append(filter_merge)\n",
    "pipeline_list.append(filter_cluster)\n",
    "pipeline_list.append(writers_gdal)\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "# save to json\n",
    "with open(cluster_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "db5b57d2-da70-452c-9893-5e3da6579dbd",
   "metadata": {},
   "outputs": [],
   "source": [
    "pdal_cmd = ['pdal','pipeline', cluster_json]\n",
    "subprocess.run(pdal_cmd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b89c9765-949d-4373-906c-15bd91f6bb06",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "efe8463e-2c0e-4d0d-b2a7-1b132a8e1f4e",
   "metadata": {},
   "source": [
    "**Move to Raster workflow**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a658118-ad67-41a7-97e2-7d64913b8d80",
   "metadata": {},
   "source": [
    "# ------------------------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f221a885-025a-44c5-a762-a9fdafc5a11f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "706a4c5e-29e3-4dce-93b0-3639e61925d6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7c740714-5881-4763-bcb2-b2387741f477",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "abbd18c8-82e0-4a86-b878-9a53710554f2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "26833f00-aa67-426b-8e20-c1c3eff46862",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "56715656-b594-4285-9b2a-5a875aaabd01",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5842dd22-6cf8-4215-9c3a-d548908e818f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "33e66bb1-b42b-445d-af25-de4b376c9652",
   "metadata": {},
   "source": [
    "# Extras"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a0db123a-b3ee-4d86-b524-b6ead9a35931",
   "metadata": {},
   "source": [
    "## Rasterize\n",
    "Basic outline for rasterizing single file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c7e682f2-1de4-4c85-9f71-d3c524a4bfae",
   "metadata": {},
   "outputs": [],
   "source": [
    "rasterize_json = json_base_path+'rasterize_mean.json' # define path to json files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "53913992-ea01-43bc-8148-2d02d0dc635c",
   "metadata": {},
   "outputs": [],
   "source": [
    "readers_las = {\"type\":\"readers.las\"}\n",
    "# write merged las to raster\n",
    "writers_gdal= {\"type\": \"writers.gdal\",\n",
    "               'output_type': 'mean', \n",
    "              'resolution': '1.0',\n",
    "               'radius': '0.7',\n",
    "               'window_size':3}\n",
    "# Append each stage to a list prior to saving to json \n",
    "pipeline_list = [readers_las, writers_gdal]\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "# save to json\n",
    "with open(rasterize_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "b164a1ab-0e58-40cb-bb6b-5a8b11536603",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_lid = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/NAD83_NAD83_epoch2010/NCALM_SCB_2014_731000_4365000.las'\n",
    "output_tif = 'SCB/Sagehen_lidar/NCALM/NCALM_SCB_2014/vdatum_test/NCALM_SCB_2014_731000_4365000_oldVDATUM_noSRSoverride.tif'\n",
    "lidar_functions.rasterize_mean(input_lid,output_tif)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1bf738b4-4eae-41dc-9cc8-898b177ac071",
   "metadata": {},
   "source": [
    "## File format conversions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "be10eab7-1dc5-4561-9ff8-dbabb55abf57",
   "metadata": {},
   "source": [
    "### Convert .las to .txt\n",
    "see: https://pdal.io/stages/writers.text.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "486c745d-b984-41bc-a04a-b54dcfbbaf18",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "069497d0-a99b-4952-bfa2-3af3d3c0ef5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# set up json file commands\n",
    "output_txt = '/Volumes/cpiske/lidar_processing/python_scripts/PDAL/test_las/mcc_part_b_tile_004_000Test.asc'\n",
    "output_json = 'lidar_processing/python_scripts/PDAL_workflow/JSON/las_to_txt.json'\n",
    "\n",
    "# create a pipeline and save to a json file \n",
    "\n",
    "filter_dict = {'type':'readers.las',\n",
    "               'override_srs': \"EPSG:4326\",\n",
    "              'filename': input_las} # we are reading in a las file\n",
    "rasterize_dict = {'type':'writers.las',\n",
    "'format':'geojson',\n",
    "'order':'X,Y,Z',\n",
    "'keep_unspecified':'false',\n",
    "'filename':output_txt}\n",
    "\n",
    "\n",
    "pipeline_list = [filter_dict, rasterize_dict]\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "with open(output_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9f570972-2f18-40d1-b4d2-06f0ed5ec7bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "json_path = 'lidar_processing/python_scripts/PDAL/JSON/las_to_txt.json'\n",
    "pdal_commands = ['pdal', 'pipeline', json_path]\n",
    "subprocess.run(pdal_commands)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "142a3a30-1ae6-481b-96ac-f95a1f6f4ddc",
   "metadata": {},
   "source": [
    "### .laz to .las"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "efe33f9e-84de-49ea-a538-0cd37b7f7877",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# # set up json file commands\n",
    "# input_laz = 'lidar_processing/python_scripts/PDAL/test_las/ASO_USCAMB20180425f1a1_180425_1_dem_filter.laz'\n",
    "# output_las = 'lidar_processing/python_scripts/PDAL/test_las/ASO_USCAMB20180425f1a1_180425_1_dem_filter.las'\n",
    "# output_json = 'lidar_processing/python_scripts/PDAL/JSON/laz_to_las.json'\n",
    "\n",
    "# # create a pipeline and save to a json file \n",
    "\n",
    "# filter_dict = {'type':'readers.las',\n",
    "#                'filename': input_las} # we are reading in a las file\n",
    "# translate_dict = {'type':'writers.las',\n",
    "#                   \"a_srs\": \"EPSG:4326\",\n",
    "#                   'filename':output_las}\n",
    "\n",
    "\n",
    "# pipeline_list = [filter_dict, translate_dict]\n",
    "# pipeline_dict = {'pipeline' : pipeline_list}\n",
    "# with open(output_json, 'w') as out:\n",
    "#     json.dump(pipeline_dict, out, indent=4)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bbb37c88-c5d9-4859-91ab-5c35e53bd514",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "efe40151-4f3c-4fc8-acc0-ff2e092d788b",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# json_path = 'lidar_processing/python_scripts/PDAL/JSON/laz_to_las.json'\n",
    "# pdal_commands = ['pdal', 'pipeline', json_path]\n",
    "# subprocess.run(pdal_commands)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea292cbb-156e-4be0-9fa8-62560bd55feb",
   "metadata": {
    "toc-hr-collapsed": true
   },
   "source": [
    "# Raster Caluclations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "66b90302-4ff2-461e-84b4-28f68e90575f",
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in [0,1,2,3,4,5]:\n",
    "    apr_elev = 'lidar_processing/PDAL_testFiles_tutorials/test_las/SCB/rasterize/ASO_20160417/mcc_part_b_tile_004_00'+str(i)+'.tif'\n",
    "    may_elev = 'lidar_processing/PDAL_testFiles_tutorials/test_las/SCB/rasterize/ASO_20160518/mcc_part_b_tile_004_00'+str(i)+'.tif'\n",
    "    output = 'lidar_processing/PDAL_testFiles_tutorials/test_las/SCB/raster_subtract/mcc_part_b_tile_004_00'+str(i)+'.tif'\n",
    "    raster_sub = ['gdal_calc.py', '-a', apr_elev, '-b', may_elev, '--calc=\"a - b\"', '--outfile', output]\n",
    "    subprocess.run(raster_sub)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ef95aabd-bd70-47a0-be17-993b89088961",
   "metadata": {},
   "source": [
    "# Theo Code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "784a85fc-2834-4d0d-81a9-19f790b08678",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_las = 'SCB/kost_lidar_data/ASO_2016/2016_05_18/WGS84_G1762_to_NAD83_NAVD88/mcc_part_b_tile_004_000.las'\n",
    "\n",
    "z_min = 0.15\n",
    "z_max = 2\n",
    "z_range = 'Z[' + str(z_min) + ':' + str(z_max) + ']'\n",
    "output_raster = 'lidar_processing/python_scripts/PDAL/test_file/mcc_part_b_tile_004_000.tif'\n",
    "output_json = 'lidar_processing/python_scripts/PDAL/JSON/las_to_tif.json'\n",
    "resolution = 0.01\n",
    "\n",
    "filter_dict = {'type':'filters.range', 'limits':z_range}\n",
    "rasterize_dict = {'filename':output_raster,\n",
    "'gdaldriver':'GTiff',\n",
    "'output_type':'count',\n",
    "'resolution':resolution,\n",
    "'type': 'writers.gdal'}\n",
    "\n",
    "\n",
    "pipeline_list = [input_las, filter_dict, rasterize_dict]\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "with open(output_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1929f364-00c7-482e-8d5f-537652eed3cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_las = 'SCB/kost_lidar_data/ASO_2016/2016_05_18/WGS84_G1762_to_NAD83_NAVD88/mcc_part_b_tile_004_000.las'\n",
    "\n",
    "# z_min = 0.15\n",
    "# z_max = 2\n",
    "# z_range = 'Z[' + str(z_min) + ':' + str(z_max) + ']'\n",
    "output_raster = 'lidar_processing/python_scripts/PDAL/test_file/mcc_part_b_tile_004_000.asc'\n",
    "output_json = 'lidar_processing/python_scripts/PDAL/JSON/las_to_asc.json'\n",
    "resolution = 0.01\n",
    "\n",
    "filter_dict = {'type':'filters.range', 'limits':z_range}\n",
    "rasterize_dict = {'filename':output_raster,\n",
    "'gdaldriver':'XYZ',\n",
    "'output_type':'count',\n",
    "'resolution':resolution,\n",
    "'type': 'writers.gdal'}\n",
    "\n",
    "\n",
    "pipeline_list = [input_las, filter_dict, rasterize_dict]\n",
    "pipeline_dict = {'pipeline' : pipeline_list}\n",
    "with open(output_json, 'w') as out:\n",
    "    json.dump(pipeline_dict, out, indent=4)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a4f6338-edab-4c85-8941-b05d63584dd8",
   "metadata": {},
   "source": [
    "sudo docker run -v path_to_laz_folder:/input 0b pdal info /input/test.laz\n",
    "\n",
    "where path_to_laz_folder is the path to the LAS/LAZ file (you just need the folder path, not the file path).\n",
    "\n",
    ":/input is the new folder that will be created in your Docker container that will hold your point cloud.\n",
    "\n",
    "0b is just the image id of pdal\n",
    "\n",
    "/input/test.laz is the path to the point cloud in the Docker container.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a941783e-828f-4226-a25a-b772a51d12d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "path_to_laz_folder = 'lidar_processing/python_scripts/PDAL/test_las'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4aee8f9b-8d85-490a-a3cd-22d4b06bb646",
   "metadata": {},
   "outputs": [],
   "source": [
    "sudo docker run -v path_to_laz_folder:/input 0b pdal info /input/test.laz\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ac352849-3336-4cd7-b7d4-2f0c2a2c79fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import subprocess\n",
    "import pdal\n",
    "\n",
    "def assemblePipeline(input_las, list_of_dicts):\n",
    "    pipeline_list = [input_las]\n",
    "    pipeline_list.extend(list_of_dicts)\n",
    "    pipeline_dict = {'pipeline' : pipeline_list}\n",
    "    return pipeline_dict\n",
    "\n",
    "def makeHeightFilter(height, buffer):\n",
    "    z_min = height - buffer/2\n",
    "    z_max = height + buffer/2\n",
    "    z_range = 'Z[' + str(z_min) + ':' + str(z_max) + ']'\n",
    "    heightDict = {'type':'filters.range', 'limits':z_range}\n",
    "    return heightDict\n",
    "\n",
    "def makeRasterizeFilter(output_raster, resolution, epsg):\n",
    "    rasterize_dict = {'filename':output_raster,\n",
    "                      'gdaldriver':'GTiff',\n",
    "                      'output_type':'count',\n",
    "                      'resolution':resolution,\n",
    "                      'override_srs' : epsg,\n",
    "                      'type': 'writers.gdal'}\n",
    "    return rasterize_dict\n",
    "\n",
    "def convertTifForPIL(input_raster, output_raster, epsg):\n",
    "    ''' GDAL bindings are an alien concept to me, so I gave up and used\n",
    "        subprocess.'''\n",
    "    commands = ['gdal_translate', input_raster, output_raster, '-ot', 'Byte', '-a_srs', epsg]\n",
    "    subprocess.run(commands)\n",
    "\n",
    "\n",
    "def buildHeightSlice(input_las, height, buffer, output_raster, resolution, epsg, json_path=None):\n",
    "    filter_dict = makeHeightFilter(height, buffer)\n",
    "    rasterize_dict = makeRasterizeFilter(output_raster, resolution, epsg)\n",
    "    filter_list = [filter_dict, rasterize_dict]\n",
    "    pipeline_dict = assemblePipeline(input_las, filter_list)\n",
    "\n",
    "    if json_path is not None:\n",
    "        with open(json_path, 'w') as out:\n",
    "            json.dump(pipeline_dict, out, indent=4)\n",
    "        pdal_commands = ['pdal', 'pipeline', json_path]\n",
    "        subprocess.run(pdal_commands)\n",
    "    else:\n",
    "        pdal_commands = json.dumps(pipeline_dict)\n",
    "        pipeline = pdal.Pipeline(pdal_commands)\n",
    "        pipeline.execute()\n",
    "\n",
    "input_las = '/Users/theo/data/las/TLS_0244_20180612_01_v003_30m_clip_height_norm.las'\n",
    "height = 1.37\n",
    "buffer = 0.05\n",
    "z_min = height - buffer/2\n",
    "z_max = height + buffer/2\n",
    "z_range = 'Z[' + str(z_min) + ':' + str(z_max) + ']'\n",
    "temp_raster = '/Users/theo/Pictures/almost_cool.tif'\n",
    "final_raster = '/Users/theo/Pictures/cool.tif'\n",
    "resolution = 0.01\n",
    "epsg = 'EPSG:3310'\n",
    "\n",
    "buildHeightSlice(input_las, height, buffer, temp_raster, resolution, epsg)\n",
    "convertTifForPIL(temp_raster, final_raster, epsg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6d188910-a9a7-4c6c-b164-8a163c1d5445",
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import subprocess\n",
    "import pdal\n",
    "import argparse\n",
    "\n",
    "# Create flags for the user to utilize.\n",
    "parser = argparse.ArgumentParser(description=\"Generate JSON pipeline to generate DTM from a point cloud.\")\n",
    "      \n",
    "required = parser.add_argument_group('Required arguments')\n",
    "required.add_argument('-crs', '--coordinate_system', required=True, action='store', help=\"EPSG code.\")\n",
    "required.add_argument('-i', '--infile', required=True, action='store', help=\"Input path to point cloud\")\n",
    "required.add_argument('-o', '--outfile', required=True, action='store', help=\"Output path.\")\n",
    "args = parser.parse_args()\n",
    "\n",
    "def generateJSON(infile, list_of_dicts):\n",
    "    pipeline_list = [infile]\n",
    "    pipeline_list.extend(list_of_dicts)\n",
    "    pipeline_dict = {'pipeline': pipeline_list}\n",
    "    with open(\"pipeline.json\", 'w') as out:\n",
    "        json.dump(pipeline_dict, out, indent=4)\n",
    "\n",
    "def generateDTM(epsg, infile, outfile):\n",
    "    reproject_dict = {\"type\": \"filters.reprojection\",\n",
    "                      \"out_srs\": \"EPSG:{}\".format(epsg)}\n",
    "    reclassify_zero_dict = {\"type\": \"filters.assign\",\n",
    "                       \"assignment\": \"Classification[:]=0\"}\n",
    "    elm_dict = {\"type\": \"filters.elm\"}\n",
    "    outlier_dict = {\"type\": \"filters.outlier\"}\n",
    "    smrf_dict = {\"type\": \"filters.smrf\", \"ignore\": \"Classification[7:7]\",\n",
    "                 \"slope\": 0.2, \"window\": 16, \"threshold\": 0.45, \"scalar\": 1.2}\n",
    "    range_dict = {\"type\":\"filters.range\", \"limits\":\"Classification[2:2]\"}\n",
    "    output_dict = {\"filename\": outfile, \"gdaldriver\": \"GTiff\", \"output_type\": \"all\", \"resolution\": 0.01, \"type\": \"writers.gdal\"}\n",
    "    list_of_dicts = list([reproject_dict, reclassify_zero_dict, elm_dict, outlier_dict, smrf_dict, range_dict, output_dict])\n",
    "    generateJSON(infile, list_of_dicts)\n",
    "    pdal_cmds = ['pdal', 'pipeline', 'pipeline.json']\n",
    "    subprocess.run(pdal_cmds)\n",
    "    \n",
    "generateDTM(args.coordinate_system, args.infile, args.outfile)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0efcfb16-ecfc-4ce2-8a5e-6bf2974c4510",
   "metadata": {},
   "source": [
    "# More helpful things"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84929bbc-a6c6-49ad-bfc5-9789fe5dd577",
   "metadata": {},
   "source": [
    "### Get stats of a dataset\n",
    "see: https://www.spatialised.net/lidar-qa-with-pdal-part-1/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fa7713f5-fbfd-412c-8488-04489fa868db",
   "metadata": {},
   "outputs": [],
   "source": [
    "# name JSON file\n",
    "\n",
    "stats_json = 'lidar_processing/python_scripts/PDAL_workflow/JSON/stats.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9cc3f4a2-adf9-4d2e-a62e-a2a538437396",
   "metadata": {},
   "outputs": [],
   "source": [
    "reader_las = {\"type\":\"readers.las\",\n",
    "              \"filename\": input_las_stats}\n",
    "filter_stats = {\"type\":\"filters.stats\",\n",
    "                \"dimensions\":\"Z\",\n",
    "                \"global\":\"Z\",\n",
    "                \"advanced\":\"true\"}\n",
    "pipeline_list = [reader_las, filter_stats]\n",
    "#pipeline_dict = {reader_las, filter_stats}\n",
    "# with open(stats_json, 'w') as out:\n",
    "#     json.dump(pipeline_dict, out, indent=4)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8ff5cfca-34a5-471d-8b40-bef477ef6438",
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "94a58e53-92c7-47c2-b75b-f9c9466b9e2d",
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline = pdal.Pipeline(json.dumps(pipeline_list))\n",
    "pipeline.execute()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "722c7605-47f2-4698-bec0-f2aa20a26f4a",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "json.loads(pipeline.metadata)[\"metadata\"][\"filters.stats\"][\"statistic\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c67b19a-0d9f-4b10-ad08-ccf9593fa659",
   "metadata": {},
   "source": [
    "## Navigating folders/files"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67984138-2046-46f9-9187-26121ac09dca",
   "metadata": {},
   "source": [
    "**create list of files/folders with a wildcard (*)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b8b23dd1-01e1-46d4-81e7-b9e5be2df38b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ex. list all files in folder4 that end in .laz = folder1/folder2/folder3/folder4/*.laz\n",
    "# ex. list all folders named folder3 in folder 1 = foler1/*/folder3\n",
    "# ex. list all list all contents in folder2 = folder2/* - note just folder 2, no subdirectories\n",
    "glob_cmd = 'path'\n",
    "glob_exe = glob.glob(file_glob_cmd)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "124af7f3-340e-4a3d-8b74-fbff724635f2",
   "metadata": {},
   "source": [
    "**create a list with all directories/subdirectories on a path**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "316d4080-2f26-48b8-b908-992955d0200f",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_folders = [x[0] for x in os.walk('path')]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1ddee80f-1c7f-46e1-a8e8-2cee82ac6f02",
   "metadata": {},
   "source": [
    "**create a list with all directories/subdirectories on a path with specific folders**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2fb7ef3c-3534-4566-9a4d-63ce9bcd074d",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_folders = [x[0] for x in os.walk('path')]\n",
    "index_pos_list = [ i for i in range(len(all_folders)) if all_folders[i][-6:] == 'retile' ] # must change this to meet requirements\n",
    "full_list = [all_folders[i] for i in index_pos_list]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "507fb6c4-a80d-4b97-8533-1852b3a8e992",
   "metadata": {},
   "source": [
    "**get name of the directory just above one listed**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ea8cc0c9-4dcb-44e7-a72f-8586c6828636",
   "metadata": {},
   "outputs": [],
   "source": [
    "subdirname = os.path.basename(os.path.dirname('path'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "141720ac-a112-4891-b727-3e4999ed4298",
   "metadata": {},
   "source": [
    "**create a list with only filenames**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5412acaf-9f08-430c-8ce2-ca0027d4d129",
   "metadata": {},
   "outputs": [],
   "source": [
    "onlyfiles = [f for f in os.listdir('path') if os.path.isfile(os.path.join('path', f))]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d80f4b2-ebb0-4eb2-9a39-58d554be97b0",
   "metadata": {},
   "source": [
    "**create a list with full file paths**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dfc96133-4087-4aca-a90b-302487a803e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "full_path = ['path' + '/' + s for s in onlyfiles]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5b8a6064-f348-405d-bf7d-5e8c8f7bfd81",
   "metadata": {},
   "source": [
    "## Parallel Processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "215beb4e-46bb-4271-b452-6849510f3a36",
   "metadata": {},
   "outputs": [],
   "source": [
    "# time = 142.5\n",
    "tic = time.perf_counter()\n",
    "list(map(function, args));\n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0a4ca6cc-9fb5-48ce-8146-b41150dd9674",
   "metadata": {},
   "outputs": [],
   "source": [
    "# time = 53 s\n",
    "tic = time.perf_counter()\n",
    "if __name__ == \"__main__\":\n",
    "    pool = Pool(3)\n",
    "    pool.map(function, arg)\n",
    "    pool.close()\n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "920e3ba6-7d5f-4a00-b563-ee2a53074b87",
   "metadata": {},
   "outputs": [],
   "source": [
    "# time = very fast? .06s\n",
    "tic = time.perf_counter()\n",
    "if __name__ == \"__main__\":\n",
    "    executor = concurrent.futures.ThreadPoolExecutor(max_workers=3)\n",
    "    executor.map(function, arg)\n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bb4c6986-5a20-472a-a5bd-4a854161ff30",
   "metadata": {},
   "outputs": [],
   "source": [
    "# time = 0.22\n",
    "tic = time.perf_counter()\n",
    "if __name__ == \"__main__\":\n",
    "    executor = concurrent.futures.ProcessPoolExecutor(max_workers=3)\n",
    "    executor.map(function, arg)\n",
    "toc = time.perf_counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d1f97306-b0c1-4b9d-8b1e-5bd95d1732c6",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}