def clean_datadir(
exp,
run,
keep_files=['input.nml', 'diag_table', 'field_table',
'git_hash_used.txt']):
"""Remove the `run` directory from output data, retaining only small
configuration files."""
outdir = exp.get_outputdir(run)
for file in keep_files:
filepath = P(outdir, 'run', file)
if os.path.isfile(filepath):
sh.cp(filepath, P(outdir, file))
exp.log.info('Copied %s to %s' % (file, outdir))
sh.rm('-r', P(outdir, 'run'))
exp.log.info('Deleted %s directory' % P(outdir, 'run'))
def interpolate_output(infile,
outfile,
all_fields=True,
var_names=[],
p_levs="input"):
"""Interpolate data from sigma to pressure levels. Includes option to remove original file.
This is a very thin wrapper around the plevel.sh script found in
`postprocessing/plevel_interpolation/scripts/plevel.sh`. Read the documentation
in that script for more information.
The interpolator must also be compiled before use. See `postprocessing/plevel_interpolation/README`
for instructions.
infile: The path of a netcdf file to interpolate over.
outfile: The path to save the output to.
all_fields: if True, interpolate all fields.
var_names: a list of fields to interpolate. This is needed in addition to
`all_fields=True` if you wish to recalculate e.g. `slp` or `height`.
p_levs: The list of pressure values, in pascals, to interpolate onto. Can be:
* A list of integer pascal values
* "input": Interpolate onto the pfull values in the input file
* "even": Interpolate onto evenly spaced in Pa levels.
Outputs to outfile.
"""
interpolator = sh.Command(
P(GFDL_BASE, 'postprocessing', 'plevel_interpolation', 'scripts',
'plevel.sh'))
# Select from pre-chosen pressure levels, or input new ones in hPa in the format below.
if isinstance(p_levs, str):
if p_levs.upper() == "INPUT":
with xr.open_dataset(infile, decode_times=False) as dat:
levels = dat.pfull.data * 100
elif p_levs.upper() == "EVEN":
#plev = ' -p "100000 95000 90000 85000 80000 75000 70000 65000 60000 55000 50000 45000 40000 35000 30000 25000 20000 15000 10000 5000" '
levels = [
100000, 95000, 90000, 85000, 80000, 75000, 70000, 65000, 60000,
55000, 50000, 45000, 40000, 35000, 30000, 25000, 20000, 15000,
10000, 5000
]
else:
raise ValueError("Unknown p_levs type '{}'".format(p_levs))
else:
levels = p_levs
plev = " ".join("{:.0f}".format(x) for x in reversed(sorted(levels)))
if all_fields:
interpolator = interpolator.bake('-a')
var_names = ' '.join(var_names)
interpolator('-i', infile, '-o', outfile, '-p', plev, var_names)
def delete_all_restarts(exp, exceptions=None):
"""Remove the restart files for a given experiment except those given.
e.g. remove_restarts(exp, [3,6,9,12])"""
if exceptions:
exceptions = [exp.restartfmt % i for i in exceptions]
all_restarts = os.listdir(exp.restartdir)
restarts_to_remove = [
file for file in all_restarts if file not in exceptions
]
for file in restarts_to_remove:
sh.rm(P(exp.restartdir, file))
exp.log.info('Deleted restart file %s' % file)
def keep_only_certain_restart_files(exp, max_num_files, interval=12):
try:
# sh.ls(sh.glob(P(exp.workdir,'restarts','res_*.cpio'))) #TODO get max_num_files calculated in line, rather than a variable to pass.
#First defines a list of ALL the restart file numbers
files_to_remove = range(0, max_num_files)
#Then defines a list of the ones we want to KEEP
files_to_keep = range(0, max_num_files, interval)
#Then we remove the files we want to keep from the list of all files, giving a list of those we wish to remove
for x in files_to_keep:
files_to_remove.remove(x)
#Then we remove them.
for entry in files_to_remove:
sh.rm(P(exp.workdir, 'restarts', 'res_' + str(entry) + '.cpio'))
except sh.ErrorReturnCode_1:
log.warning(
'Tried to remove some restart files, but the last one doesnt exist'
)