Skip to content
Snippets Groups Projects
Commit 40e88264 authored by Nick Bearson's avatar Nick Bearson
Browse files

reset make_GLM_grids.py to Eric's original example script (#12)

parent 1e1de04c
No related branches found
No related tags found
1 merge request!52020 02 10 beta1 high priority
Hide whitespace changes
Inline Side-by-side
gridded_glm/libexec/_make_glm_grids.py
+ 248
− 189
View file @ 40e88264
#!/usr/bin/env python3
# Based on https://github.com/deeplycloudy/glmtools/blob/master/examples/grid/make_GLM_grids.py
parse_desc = """Grid GLM flash data.
The start and end times can be specified
# This example was provided by GLMTools
# https://github.com/deeplycloudy/glmtools/blob/master/examples/grid/make_GLM_grids.py
import argparse
parse_desc = """Grid GLM flash data. The start and end times can be specified
independently, or if not provided they will be inferred from the filenames.
Grid spacing is regular in latitude and longitude with the grid box
......@@ -14,84 +13,129 @@ standard GOES imagery naming convention. This behavior can be fully controlled
by adjusting the -o argument.
"""
import numpy as np
from datetime import datetime
import os
from functools import partial
from lmatools.grid.make_grids import write_cf_netcdf_latlon, write_cf_netcdf_noproj, write_cf_netcdf_fixedgrid
from lmatools.grid.make_grids import dlonlat_at_grid_center, grid_h5flashfiles
from glmtools.grid.make_grids import grid_GLM_flashes
from glmtools.io.glm import parse_glm_filename
from lmatools.grid.fixed import get_GOESR_grid, get_GOESR_coordsys
import logging
log = logging.getLogger(__name__)
output_help = """Specify the output path and filename using a configurable path
template. -o ./{dataset_name} (the default) will generate files in the current
directory using the standard GOES imagery naming convention, including a .nc
extension. Any intermediate directories will be created as needed. All allowed
names in the template are listed in the docs for
glmtools.io.imagery.write_goes_imagery. For example: this script can be used to
process multiple days and that are written to a standardized directory
structure by specifying a path like so: -o
{start_time:%%Y/%%b/%%d}/{dataset_name}"""
def create_parser():
import argparse
parser = argparse.ArgumentParser(description=parse_desc)
parser.add_argument('-v', '--verbose', dest='verbosity', action="count", default=0,
help='each occurrence increases verbosity 1 level through ERROR-WARNING-INFO-DEBUG (default INFO)')
parser.add_argument('-l', '--log', dest="log_fn", default=None,
help="specify the log filename")
parser.add_argument('-o', '--output-dir', metavar='output directory',
default=os.getcwd())
parser.add_argument('--ctr-lat', metavar='latitude',
type=float, help='center latitude')
parser.add_argument('--ctr-lon', metavar='longitude',
type=float, help='center longitude')
parser.add_argument(dest='filenames',metavar='filename', nargs='*')
parser.add_argument('-o', '--output_path',
metavar='filename template including path',
required=False, dest='outdir', action='store',
default='./{dataset_name}', help=output_help)
parser.add_argument('--ctr_lat', metavar='latitude', required=False,
dest='ctr_lat', action='store', type=float,
help='center latitude')
parser.add_argument('--ctr_lon', metavar='longitude', required=False,
dest='ctr_lon', action='store', type=float,
help='center longitude')
parser.add_argument('--start', metavar='yyyy-mm-ddThh:mm:ss',
dest='start', action='store',
help='UTC start time, e.g., 2017-07-04T08:00:00')
parser.add_argument('--end', metavar='yyyy-mm-ddThh:mm:ss',
dest='end', action='store',
help='UTC end time, e.g., 2017-07-04T09:00:00')
parser.add_argument('--dx', metavar='km',
default=10.0, type=float,
dest='dx', action='store', default=10.0, type=float,
help='approximate east-west grid spacing')
parser.add_argument('--dy', metavar='km',
default=10.0, type=float,
dest='dy', action='store', default=10.0, type=float,
help='approximate north-south grid spacing')
parser.add_argument('--dt', metavar='seconds',
default=60.0, type=float,
dest='dt', action='store', default=60.0, type=float,
help='frame duration')
parser.add_argument('--width', metavar='distance in km',
default=400.0,
dest='width', action='store', default=400.0,
type=float, help='total width of the grid')
parser.add_argument('--height', metavar='distance in km',
default=400.0,
dest='height', action='store', default=400.0,
type=float, help='total height of the grid')
parser.add_argument('--nevents', metavar='minimum events per flash',
type=int, dest='min_events', default=1,
type=int, dest='min_events', action='store', default=1,
help='minimum number of events per flash')
parser.add_argument('--ngroups', metavar='minimum groups per flash',
type=int, dest='min_groups', default=1,
type=int, dest='min_groups', action='store', default=1,
help='minimum number of groups per flash')
parser.add_argument('--subdivide-grid', metavar='sqrt(number of subgrids)',
parser.add_argument('--fixed_grid',
action='store_true', dest='fixed_grid',
help='grid to the geostationary fixed grid')
parser.add_argument('--subdivide_grid', metavar='sqrt(number of subgrids)',
action='store', dest='subdivide_grid',
type=int, default=1,
help="subdivide the grid this many times along "
"each dimension")
parser.add_argument('--goes-position',
help="One of [east|west|test|auto]. "
"Requires '--goes-sector'.")
parser.add_argument('--goes-sector',
help="One of [full|conus|meso]. "
"Requires goes_position. If sector is "
"meso, ctr_lon and ctr_lat are interpreted as "
"the ctr_x and ctr_y of the fixed grid")
# parser.add_argument('--split-events',
# action='store_true',
# help='Split GLM event polygons when gridding')
help=("subdivide the grid this many times along "
"each dimension"))
parser.add_argument('--goes_position', default='none',
action='store', dest='goes_position',
help=("One of [east|west|test]. "
"Also requires goes_sector."))
parser.add_argument('--goes_sector', default='none',
action='store', dest='goes_sector',
help=("One of [full|conus|meso]. "
"Also requires goes_position. If sector is "
"meso, ctr_lon and ctr_lat are interpreted as "
"the ctr_x and ctr_y of the fixed grid. "
"Omit if you are creating a fully custom grid "
"with --width and --height arguments."))
parser.add_argument('--corner_points', metavar='filename.pickle',
action='store', dest='corner_points',
help=("name of file containing a pickled "
"corner point lookup table"))
parser.add_argument('--split_events', dest='split_events',
action='store_true',
help='Split GLM event polygons when gridding')
parser.add_argument('--ellipse', dest='ellipse_rev', default=-1,
type=int,
action='store', type=int,
help='Lightning ellipse revision. -1 (default)=infer'
' from date in each GLM file, 0=value at launch,'
' 1=late 2018 revision')
parser.add_argument('--float-output', dest='output_scale_and_offset',
parser.add_argument('--float_output', dest='output_scale_and_offset',
default=True,
action='store_false',
help='write all output variables as floating point')
parser.add_argument(dest='filenames', metavar='filename', nargs='+')
parser.add_argument('--lma', dest='is_lma',
action='store_true',
help='grid LMA h5 files instead of GLM data')
# parser.add_argument('-v', dest='verbose', action='store_true',
# help='verbose mode')
return parser
##### END PARSING #####
import numpy as np
import subprocess, glob
from datetime import datetime, timedelta
import os
from functools import partial
import logging
class MyFormatter(logging.Formatter):
""" Custom class to allow logging of microseconds"""
converter=datetime.fromtimestamp
def formatTime(self, record, datefmt=None):
ct = self.converter(record.created)
if datefmt:
s = ct.strftime(datefmt)
else:
t = ct.strftime("%Y-%m-%d %H:%M:%S")
s = "%s,%03d" % (t, record.msecs)
return s
logoutfile = logging.FileHandler("make_GLM_grid.log")
formatter = MyFormatter(fmt='%(levelname)s %(asctime)s %(message)s',
datefmt='%Y-%m-%dT%H:%M:%S.%f')
logoutfile.setFormatter(formatter)
logging.basicConfig(handlers = [logoutfile],
level=logging.DEBUG)
# Separate from log setup - actually log soemthign specific to this module.
log = logging.getLogger(__name__)
log.info("Starting GLM Gridding")
def nearest_resolution(args):
""" Uses args.dx to find the closest resolution specified by the
......@@ -104,47 +148,14 @@ def nearest_resolution(args):
resln = '{0:4.1f}km'.format(closest_resln).replace(' ', '')
return resln
# if provided "auto" position, we determine the sensor from the filename
def get_goes_position(filenames):
if all("_G16_" in f for f in filenames):
return "east"
if all("_G17_" in f for f in filenames):
return "west"
# we require that all files are from the same sensor and raise an exception if not
raise ValueError("position 'auto' but could not determine position - did you provide a mix of satellites?")
def get_start_end(filenames, start_time=None, end_time=None):
"""Compute start and end time of data based on filenames."""
base_filenames = [os.path.basename(p) for p in filenames]
try:
filename_infos = [parse_glm_filename(f) for f in base_filenames]
# opsenv, algorithm, platform, start, end, created = parse_glm_filename(f)
filename_starts = [info[3] for info in filename_infos]
filename_ends = [info[4] for info in filename_infos]
except ValueError:
filename_starts = None
filename_ends = None
if args.start is not None:
start_time = datetime.strptime(args.start, '%Y-%m-%dT%H:%M:%S')
elif filename_starts is not None:
start_time = min(filename_starts)
if args.end is not None:
end_time = datetime.strptime(args.end, '%Y-%m-%dT%H:%M:%S')
elif filename_ends is not None:
end_time = max(filename_ends)
if start_time is None or end_time is None:
raise ValueError("Could not determine start/end time")
return start_time, end_time
def grid_setup(args):
from lmatools.grid.make_grids import write_cf_netcdf_latlon, write_cf_netcdf_noproj, write_cf_netcdf_fixedgrid
from lmatools.grid.make_grids import dlonlat_at_grid_center, grid_h5flashfiles
from glmtools.grid.make_grids import grid_GLM_flashes
from glmtools.io.glm import parse_glm_filename
from lmatools.io.LMA_h5_file import parse_lma_h5_filename
from lmatools.grid.fixed import get_GOESR_grid, get_GOESR_coordsys
# When passed None for the minimum event or group counts, the gridder will skip
# the check, saving a bit of time.
min_events = int(args.min_events)
......@@ -154,120 +165,168 @@ def grid_setup(args):
if min_groups <= 1:
min_groups = None
if args.is_lma:
filename_parser = parse_lma_h5_filename
start_idx = 0
end_idx = 1
else:
filename_parser = parse_glm_filename
start_idx = 3
end_idx = 4
glm_filenames = args.filenames
base_filenames = [os.path.basename(p) for p in glm_filenames]
try:
start_time, end_time = get_start_end(args.filenames, args.start, args.end)
filename_infos = [filename_parser(f) for f in base_filenames]
# opsenv, algorithm, platform, start, end, created = parse_glm_filename(f)
filename_starts = [info[start_idx] for info in filename_infos]
filename_ends = [info[end_idx] for info in filename_infos]
except ValueError:
log.error("Non-standard filenames provided, use --start and --end to specify data times.")
raise
date = datetime(start_time.year, start_time.month, start_time.day)
os.makedirs(args.output_dir, exist_ok=True)
output = os.path.join(args.output_dir, "{dataset_name}") # GLMTools expects a template in addition to the path
proj_name = 'geos'
log.error("One or more GLM files has a non-standard filename.")
log.error("Assuming that --start and --end have been passed directly.")
if args.goes_position == "auto":
# auto-create the goes-position from the input filename
args.goes_position = get_goes_position(args.filenames)
from glmtools.io.glm import parse_glm_filename
if args.start is not None:
start_time = datetime.strptime(args.start[:19], '%Y-%m-%dT%H:%M:%S')
else:
start_time = min(filename_starts)
if args.end is not None:
end_time = datetime.strptime(args.end[:19], '%Y-%m-%dT%H:%M:%S')
else:
# Used to use max(filename_ends), but on 27 Oct 2020, the filename
# ends started to report the time of the last event in the file,
# causing a slight leakage (usually less than a second) into the
# next minute. This caused two minutes of grids to be produced for every
# three twenty second files passed to this script.
# Instead, we now assume every LCFA file is 20 s long, beginning with
# the start time. No doubt in the future we will see filenames that no
# longer start on an even minute boundary.
end_time = max(filename_starts) + timedelta(0, 20)
if args.goes_position is not None and args.goes_sector is not None:
resln = nearest_resolution(args)
view = get_GOESR_grid(position=args.goes_position,
view=args.goes_sector,
resolution=resln)
nadir_lon = view['nadir_lon']
dx = dy = view['resolution']
nx, ny = view['pixelsEW'], view['pixelsNS']
geofixcs, grs80lla = get_GOESR_coordsys(sat_lon_nadir=nadir_lon)
date = datetime(start_time.year, start_time.month, start_time.day)
if 'centerEW' in view:
x_ctr, y_ctr = view['centerEW'], view['centerNS']
elif args.goes_sector == 'meso':
# use ctr_lon, ctr_lat to get the center of the mesoscale FOV
outpath = args.outdir
if args.fixed_grid:
proj_name = 'geos'
if (args.goes_position != 'none') & (args.goes_sector != 'none'):
resln = nearest_resolution(args)
view = get_GOESR_grid(position=args.goes_position,
view=args.goes_sector,
resolution=resln)
nadir_lon = view['nadir_lon']
dx = dy = view['resolution']
nx, ny = view['pixelsEW'], view['pixelsNS']
geofixcs, grs80lla = get_GOESR_coordsys(sat_lon_nadir=nadir_lon)
if 'centerEW' in view:
x_ctr, y_ctr = view['centerEW'], view['centerNS']
elif args.goes_sector == 'meso':
# use ctr_lon, ctr_lat to get the center of the mesoscale FOV
x_ctr, y_ctr, z_ctr = geofixcs.fromECEF(
*grs80lla.toECEF(args.ctr_lon, args.ctr_lat, 0.0))
elif (args.goes_position != 'none') & (args.goes_sector == 'none'):
# Requires goes_position, a center, and a width. Fully flexible
# in resolution, i.e., doesn't slave it to one of the GOES-R specs
view = get_GOESR_grid(position=args.goes_position,
view='full',
resolution='1.0km')
nadir_lon = view['nadir_lon']
dx1km = dy1km = view['resolution']
geofixcs, grs80lla = get_GOESR_coordsys(sat_lon_nadir=nadir_lon)
x_ctr, y_ctr, z_ctr = geofixcs.fromECEF(
*grs80lla.toECEF(args.ctr_lon, args.ctr_lat, 0.0))
elif args.goes_position is not None and args.goes_sector is None:
# Requires goes_position, a center, and a width. Fully flexible
# in resolution, i.e., doesn't slave it to one of the GOES-R specs
view = get_GOESR_grid(position=args.goes_position,
view='full',
resolution='1.0km')
nadir_lon = view['nadir_lon']
dx1km = dy1km = view['resolution']
*grs80lla.toECEF(args.ctr_lon, args.ctr_lat, 0.0))
# Convert the specified resolution in km given by args.dx to
# a delta in fixed grid coordinates using the 1 km delta from the
# GOES-R PUG.
dx, dy = args.dx * dx1km, args.dy * dy1km
nx, ny = int(args.width/args.dx), int(args.height/args.dy)
else:
raise ValueError("Gridding on the fixed grid requires "
"goes_position and dx. For goes_sector='meso', also specify "
"ctr_lon and ctr_lat. Without goes_sector, also include width "
"and height.")
# Need to use +1 here to convert to xedge, yedge expected by gridder
# instead of the pixel centroids that will result in the final image
nx += 1
ny += 1
x_bnd = (np.arange(nx, dtype='float') - (nx)/2.0)*dx + x_ctr + 0.5*dx
y_bnd = (np.arange(ny, dtype='float') - (ny)/2.0)*dy + y_ctr + 0.5*dy
log.debug(("initial x,y_ctr", x_ctr, y_ctr))
log.debug(("initial x,y_bnd", x_bnd.shape, y_bnd.shape))
x_bnd = np.asarray([x_bnd.min(), x_bnd.max()])
y_bnd = np.asarray([y_bnd.min(), y_bnd.max()])
geofixcs, grs80lla = get_GOESR_coordsys(sat_lon_nadir=nadir_lon)
x_ctr, y_ctr, z_ctr = geofixcs.fromECEF(
*grs80lla.toECEF(args.ctr_lon, args.ctr_lat, 0.0))
ctr_lon, ctr_lat, ctr_alt = grs80lla.fromECEF(
*geofixcs.toECEF(x_ctr, y_ctr, 0.0))
fixed_grid = geofixcs
log.debug((x_bnd, y_bnd, dx, dy, nx, ny))
# Convert the specified resolution in km given by args.dx to
# a delta in fixed grid coordinates using the 1 km delta from the
# GOES-R PUG.
dx, dy = args.dx * dx1km, args.dy * dy1km
nx, ny = int(args.width / args.dx), int(args.height / args.dy)
output_writer = partial(write_cf_netcdf_fixedgrid, nadir_lon=nadir_lon)
else:
raise ValueError("Gridding on the fixed grid requires "
"goes_position and dx. For goes_sector='meso', also specify "
"ctr_lon and ctr_lat. Without goes_sector, also include width "
"and height.")
# Need to use +1 here to convert to xedge, yedge expected by gridder
# instead of the pixel centroids that will result in the final image
nx += 1
ny += 1
x_bnd = (np.arange(nx, dtype='float') - (nx) / 2.0) * dx + x_ctr + 0.5 * dx
y_bnd = (np.arange(ny, dtype='float') - (ny) / 2.0) * dy + y_ctr + 0.5 * dy
log.debug(("initial x,y_ctr", x_ctr, y_ctr))
log.debug(("initial x,y_bnd", x_bnd.shape, y_bnd.shape))
x_bnd = np.asarray([x_bnd.min(), x_bnd.max()])
y_bnd = np.asarray([y_bnd.min(), y_bnd.max()])
geofixcs, grs80lla = get_GOESR_coordsys(sat_lon_nadir=nadir_lon)
ctr_lon, ctr_lat, ctr_alt = grs80lla.fromECEF(
*geofixcs.toECEF(x_ctr, y_ctr, 0.0))
fixed_grid = geofixcs
log.debug((x_bnd, y_bnd, dx, dy, nx, ny))
output_writer = partial(write_cf_netcdf_fixedgrid, nadir_lon=nadir_lon)
gridder = grid_GLM_flashes
output_filename_prefix = 'GLM'
grid_kwargs = dict(proj_name=proj_name,
base_date=date, do_3d=False,
dx=dx, dy=dy, frame_interval=float(args.dt),
x_bnd=x_bnd, y_bnd=y_bnd,
ctr_lat=ctr_lat, ctr_lon=ctr_lon, outpath=output,
min_points_per_flash=min_events,
output_writer=output_writer, subdivide=args.subdivide_grid,
output_filename_prefix=output_filename_prefix,
output_kwargs={'scale_and_offset': args.output_scale_and_offset},
spatial_scale_factor=1.0)
#if args.fixed_grid:
# grid_kwargs['fixed_grid'] = True
# grid_kwargs['nadir_lon'] = nadir_lon
# if args.split_events:
grid_kwargs['clip_events'] = True
# Default
proj_name='latlong'
output_writer = write_cf_netcdf_latlon
ctr_lat = float(args.ctr_lat)
ctr_lon = float(args.ctr_lon)
dx_km=float(args.dx)*1.0e3
dy_km=float(args.dy)*1.0e3
width, height = 1000.0*float(args.width), 1000.0*float(args.height)
x_bnd_km = (-width/2.0, width/2.0)
y_bnd_km = (-height/2.0, height/2.0)
dx, dy, x_bnd, y_bnd = dlonlat_at_grid_center(ctr_lat, ctr_lon,
dx=dx_km, dy=dy_km,
x_bnd = x_bnd_km, y_bnd = y_bnd_km )
# tuples of the corners
corners = np.vstack([(x_bnd[0], y_bnd[0]), (x_bnd[0], y_bnd[1]),
(x_bnd[1], y_bnd[1]), (x_bnd[1], y_bnd[0])])
# print(x_bnd, y_bnd)
if args.is_lma:
gridder = grid_h5flashfiles
output_filename_prefix='LMA'
else:
gridder = grid_GLM_flashes
output_filename_prefix='GLM'
grid_kwargs=dict(proj_name=proj_name,
base_date = date, do_3d=False,
dx=dx, dy=dy, frame_interval=float(args.dt),
x_bnd=x_bnd, y_bnd=y_bnd,
ctr_lat=ctr_lat, ctr_lon=ctr_lon, outpath = outpath,
min_points_per_flash = min_events,
output_writer = output_writer, subdivide=args.subdivide_grid,
output_filename_prefix=output_filename_prefix,
output_kwargs={'scale_and_offset':args.output_scale_and_offset},
spatial_scale_factor=1.0)
if args.fixed_grid:
grid_kwargs['fixed_grid'] = True
grid_kwargs['nadir_lon'] = nadir_lon
if args.split_events:
grid_kwargs['clip_events'] = True
if min_groups is not None:
grid_kwargs['min_groups_per_flash'] = min_groups
grid_kwargs['energy_grids'] = ('total_energy',)
if (proj_name == 'pixel_grid') or (proj_name == 'geos'):
if args.is_lma:
grid_kwargs['energy_grids'] = True
else:
grid_kwargs['energy_grids'] = ('total_energy',)
if (proj_name=='pixel_grid') or (proj_name=='geos'):
grid_kwargs['pixel_coords'] = fixed_grid
grid_kwargs['ellipse_rev'] = args.ellipse_rev
return gridder, args.filenames, start_time, end_time, grid_kwargs
# if args.corner_points:
# grid_kwargs['corner_pickle'] = args.corner_points
return gridder, glm_filenames, start_time, end_time, grid_kwargs
if __name__ == '__main__':
import sys
parser = create_parser()
args = parser.parse_args()
# Configure logging
levels = [logging.ERROR, logging.WARN, logging.INFO, logging.DEBUG]
logging.basicConfig(level=levels[min(3, args.verbosity)], filename=args.log_fn)
if levels[min(3, args.verbosity)] > logging.DEBUG:
import warnings
warnings.filterwarnings("ignore")
log.info("Starting GLM Gridding")
log.debug("Starting script with: %s", sys.argv)
from multiprocessing import freeze_support
freeze_support()
gridder, glm_filenames, start_time, end_time, grid_kwargs = grid_setup(args)
gridder(glm_filenames, start_time, end_time, **grid_kwargs)
gridder(glm_filenames, start_time, end_time, **grid_kwargs)
\ No newline at end of file
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment