def read_cfgrib(filename, **kwargs):
""" Read grib files using cfgrib
Args:
filename:
**kwargs:
Returns:
"""
import xarray as xr
try:
import cfgrib
import cf2cdm
except ImportError:
print("requires eccodes library, and xarray")
raise ImportError()
kwargs['engine'] = 'cfgrib'
kwargs['backend_kwargs'] = kwargs.get('backend_kwargs', {})
if 'indexpath' not in kwargs['backend_kwargs']:
kwargs['backend_kwargs'].update({'indexpath': ''}) # no indexfiles
data = xr.open_dataset(filename, **kwargs)
return cf2cdm.translate_coords(data, cf2cdm.CDS)
def dump(inpaths, variable, cdm, engine):
# type: (T.List[str], str, str, str) -> None
import xarray as xr
import cf2cdm
# NOTE: noop if no input argument
if len(inpaths) == 0:
return
if len(inpaths) == 1:
# avoid to depend on dask when passing only one file
ds = xr.open_dataset(inpaths[0], engine=engine) # type: ignore
else:
ds = xr.open_mfdataset(inpaths, engine=engine,
combine="by_coords") # type: ignore
if cdm:
coord_model = getattr(cf2cdm, cdm)
ds = cf2cdm.translate_coords(ds, coord_model=coord_model)
if variable:
ds_or_da = ds[variable]
else:
ds_or_da = ds
print(ds_or_da)
def to_netcdf(inpaths, outpath, cdm, engine):
import cf2cdm
import xarray as xr
# NOTE: noop if no input argument
if len(inpaths) == 0:
return
if not outpath:
outpath = os.path.splitext(inpaths[0])[0] + '.nc'
ds = xr.open_mfdataset(inpaths, engine=engine)
if cdm:
coord_model = getattr(cf2cdm, cdm)
ds = cf2cdm.translate_coords(ds, coord_model=coord_model)
ds.to_netcdf(outpath)
def to_netcdf(inpaths, outpath, cdm, engine, backend_kwargs_json):
# type: (T.List[str], str, str, str, str) -> None
import json
import xarray as xr
import cf2cdm
# NOTE: noop if no input argument
if len(inpaths) == 0:
return
if not outpath:
outpath = os.path.splitext(inpaths[0])[0] + ".nc"
if backend_kwargs_json is not None:
try:
# Assume a json format string
backend_kwargs = json.loads(backend_kwargs_json)
except json.JSONDecodeError:
# Then a json file
with open(backend_kwargs_json, "r") as f:
backend_kwargs = json.load(f)
else:
backend_kwargs = {}
if len(inpaths) == 1:
# avoid to depend on dask when passing only one file
ds = xr.open_dataset(
inpaths[0],
engine=engine,
backend_kwargs=backend_kwargs,
) # type: ignore
else:
ds = xr.open_mfdataset(
inpaths,
engine=engine,
combine="by_coords",
backend_kwargs=backend_kwargs,
) # type: ignore
if cdm:
coord_model = getattr(cf2cdm, cdm)
ds = cf2cdm.translate_coords(ds, coord_model=coord_model)
ds.to_netcdf(outpath)
def to_netcdf(inpaths, outpath, cdm, engine):
# type: (T.List[str], str, str, str) -> None
import xarray as xr
import cf2cdm
# NOTE: noop if no input argument
if len(inpaths) == 0:
return
if not outpath:
outpath = os.path.splitext(inpaths[0])[0] + ".nc"
if len(inpaths) == 1:
# avoid to depend on dask when passing only one file
ds = xr.open_dataset(inpaths[0], engine=engine) # type: ignore
else:
ds = xr.open_mfdataset(inpaths, engine=engine,
combine="by_coords") # type: ignore
if cdm:
coord_model = getattr(cf2cdm, cdm)
ds = cf2cdm.translate_coords(ds, coord_model=coord_model)
ds.to_netcdf(outpath)
def add(self, name, filename=None, directory=None, rename=None, cfunits=False, close=True, xwargs=None, **kwargs):
""" Add dataset to radiosonde class object [container]
Args:
name (str): used as filename and/or as dataset name
filename (str): filename to read from (netcdf)
directory (str): directory of radiosonde store, default config rasodir
rename (dict): rename variables
cfunits (bool): apply cfunits
xwargs (dict): xarray open_dataset keywords
"""
if rename is None:
rename = {}
if xwargs is None:
xwargs = {}
from .. import config
if self.directory is not None:
directory = self.directory
if 'engine' not in xwargs.keys():
xwargs['engine'] = 'h5netcdf'
if directory is None:
if 'rasodir' in config and os.path.isdir(config.rasodir):
directory = config.rasodir + '/' + str(self.ident) + '/'
else:
directory = './' + str(self.ident) + '/'
message(directory, **kwargs)
if filename is not None:
if '.nc' not in filename:
print("Warning can read only NetCDF / Xarray open_dataset formats")
if '*' not in filename:
if not os.path.isfile(filename):
if directory is not None:
if not os.path.isfile(directory + '/' + filename):
filename = directory + '/' + str(self.ident) + '/' + filename
else:
filename = directory + '/' + filename
if filename is None:
if os.path.isfile(directory + '/' + name + '.nc'):
filename = directory + '/' + name + '.nc'
else:
message("Not found:", name, directory, **kwargs)
return
if '*' in filename:
if 'combine' not in xwargs.keys():
xwargs['combine'] = 'by_coords'
try:
message("Reading ...", filename, **kwargs)
self.data[name] = xr.open_mfdataset(filename, **xwargs)
except OSError:
message("Reading ...", directory + '/' + filename, **kwargs)
self.data[name] = xr.open_mfdataset(directory + '/' + filename, **xwargs)
else:
message("Reading ...", filename, **kwargs)
self.data[name] = xr.open_dataset(filename, **xwargs)
if close:
# add this to make sure that a file is read and closed (some error with h5netcdf, h5py)
if hasattr(self.data[name], 'close'):
self.data[name].load()
self.data[name].close()
self.data[name] = self.data[name].rename(rename)
if cfunits:
import cf2cdm
for i, j in self.data[name].coords.items():
if str(j.dtype) == 'datetime64[ns]':
if 'standard_name' in j.attrs:
del j.attrs['standard_name']
self.data[name] = cf2cdm.translate_coords(self.data[name], cf2cdm.CDS)
# merge dictionaries and append common
self.attrs.__dict__ = dict_add(vars(self.attrs), dict(self.data[name].attrs))
if 'ident' in self.attrs:
if self.ident != self.attrs['ident']:
message("Warning different idents: ", self.ident, ">", self.attrs['ident'],
**update_kw('level', -1, **kwargs))
if self.ident == "":
self.ident = self.attrs['ident']
