def open_file(self, file_name, mode=None, format='nc'):
"""
open_file() [public]
Purpose: Opens a file with the given mode.
Parameters: file_name [type=string]
Name of the file to open.
mode [type=string]
How to open the file (e.g. 'r' for reading, 'rw' for reading and writing, etc.)
Passing None defaults to opening for reading only.
Returns: [nothing]
"""
# Set the mode default
if mode is None: mode = 'r'
# If the file is already open, close it.
if self._df is not None:
self.close()
self._file_name = file_name
self._format = format
# Figure out whether we're reading, writing, or both.
self._read = (mode.find('r') > -1)
self._write = (mode.find('w') > -1 or mode == 'r+')
# if self._write:
# warning("DataIO: Writing is currently not supported. Opening as read-only.")
# self._write = False
if reader.__name__ == "Nio":
# If the reader is PyNIO open it with PyNIO's open command
self._df = reader.open_file(file_name, mode=mode, format=format)
elif reader.__name__ == "scipy.io.netcdf":
# If the reader is scipy, open it with scipy's open command
if self._read:
self._df = reader.netcdf_file(file_name, 'r')
elif self._write:
self._df = reader.netcdf_file(file_name, 'w')
elif reader.__name__ == "Dataset":
# If the reader is netCDF4, open it with that open command
if self._read:
self._df = reader(file_name, mode='r')
elif self._write:
self._df = reader(file_name, mode='a')
return
def open_file(self, file_name, mode=None):
"""
open_file() [public]
Purpose: Opens a file with the given mode.
Parameters: file_name [type=string]
Name of the file to open.
mode [type=string]
How to open the file (e.g. 'r' for reading, 'rw' for reading and writing, etc.)
Passing None defaults to opening for reading only.
Returns: [nothing]
"""
# Set the mode default
if mode is None: mode = 'r'
# If the file is already open, close it.
if self._df is not None:
self.close()
self._file_name = file_name
# Figure out whether we're reading, writing, or both.
self._read = (mode.find('r') > -1)
self._write = (mode.find('w') > -1 or mode == 'r+')
# if self._write:
# warning("DataIO: Writing is currently not supported. Opening as read-only.")
# self._write = False
if reader.__name__ == "Nio":
# If the reader is PyNIO open it with PyNIO's open command
self._df = reader.open_file(file_name, mode=mode)
elif reader.__name__ == "scipy.io.netcdf":
# If the reader is scipy, open it with scipy's open command
if self._read:
self._df = reader.netcdf_file(file_name, 'r')
elif self._write:
self._df = reader.netcdf_file(file_name, 'w')
elif reader.__name__ == "Dataset":
# If the reader is netCDF4, open it with that open command
if self._read:
self._df = reader(file_name, mode='r')
elif self._write:
self._df = reader(file_name, mode='a')
return
def close(self,out_file_name=''):
"""
close() [public]
Purpose: Closes the file and resets the class to its initial state. The file is set for writing, and any changes have occurred, create a
new file and copy all the variables over, without anything that's been "removed."
Parameters: [none]
Returns: [nothing]
"""
self._attempted_close = True
if self._write:
# Copy the data to another file, excluding the removed variables ...
tmp_file_name = "%s.tmp" % self._file_name
tmp_file = None
if reader.__name__ == "Nio":
tmp_file = reader.open_file(tmp_file_name, format='nc', mode='w')
elif reader.__name__ == "scipy.io.netcdf":
tmp_file = reader.netcdf_file(tmp_file_name, mode='w')
elif reader.__name__ == "Dataset":
tmp_file = reader(tmp_file_name,'w')
# Copy dimensions over
for dimension, length in self._df.dimensions.iteritems():
if reader.__name__ == "Nio":
tmp_file.create_dimension(dimension, length)
elif reader.__name__ in ["scipy.io.netcdf", "Dataset"]:
tmp_file.createDimension(dimension, len(length))
# Create new dimensions added by splicing
for variable, splice_info in self._sp_variables.iteritems():
dim_names, index_lists = splice_info
for dim_name, indexes in zip(dim_names, index_lists):
if dim_name not in tmp_file.dimensions.keys():
if reader.__name__ == "Nio":
tmp_file.create_dimension(dim_name, len(np.unique(indexes)))
elif reader.__name__ in ["scipy.io.netcdf", "Dataset"]:
tmp_file.createDimension(dim_name, len(np.unique(indexes)))
# Copy variable definitions over
for variable, data in self._df.variables.iteritems():
if variable not in self._rm_variables:
# If this variable is going to be spliced, find its new dimensions
if variable not in self._sp_variables:
dim_names = data.dimensions
else:
dim_names, index_lists = self._sp_variables[variable]
if reader.__name__ == "Nio":
tmp_file.create_variable(variable, data.typecode(), dim_names)
elif reader.__name__ == "scipy.io.netcdf":
tmp_file.createVariable(variable, data.dtype, dim_names)
elif reader.__name__ == "Dataset":
if hasattr(data, "_FillValue"):
tmp_file.createVariable(variable, data.dtype, dim_names, fill_value=data._FillValue)
else:
tmp_file.createVariable(variable, data.dtype, dim_names)
# Copy file attributes over
for attribute in dir(self._df):
if not hasattr(tmp_file, attribute) and attribute not in self._rm_attributes['__file__']:
setattr(tmp_file, attribute, getattr(self._df, attribute))
# Copy variable attributes over
for variable, data in tmp_file.variables.iteritems():
try:
for attribute in dir(self._df.variables[variable]):
if not hasattr(data, attribute) and attribute not in self._rm_attributes[variable]:
setattr(data, attribute, getattr(self._df.variables[variable], attribute))
except KeyError:
for attribute in dir(self._df.variables[variable]):
if not hasattr(data, attribute):
setattr(data, attribute, getattr(self._df.variables[variable], attribute))
# Copy variable data over (NetCDF4 Python appears to require this be done after setting all the variable attributes)
for variable, data in self._df.variables.iteritems():
if variable not in self._rm_variables:
tmp_variable = np.copy(self._df.variables[variable][:])
# Overwrite the data if we need to ...
if variable in self._ov_variables:
tmp_variable = np.copy(self._ov_variables[variable])
# Do splicing ...
if variable in self._sp_variables:
dim_names, index_lists = self._sp_variables[variable]
tmp_variable = np.copy(self._df.variables[variable][np.ix_(*index_lists)])
# Put it in the file ...
tmp_file.variables[variable][:] = tmp_variable
# for dimension, length in tmp_file.dimensions.iteritems():
# print dimension, length
self._df.close()
if self._write:
# Move temporary file to original file's location
import shutil
if len(out_file_name) > 0:
shutil.move(tmp_file_name,out_file_name)
else:
shutil.move(tmp_file_name, self._file_name)
self._df = None
self._file_name = ""
self._read = False
self._write = False
self._rm_variables = []
self._sp_variables = {}
self._rm_attributes = { '__file__':[] }
return
def close(self, out_file_name=''):
"""
close() [public]
Purpose: Closes the file and resets the class to its initial state. The file is set for writing, and any changes have occurred, create a
new file and copy all the variables over, without anything that's been "removed."
Parameters: [none]
Returns: [nothing]
"""
self._attempted_close = True
if self._write:
# Copy the data to another file, excluding the removed variables ...
tmp_file_name = "%s.tmp" % self._file_name
tmp_file = None
if reader.__name__ == "Nio":
tmp_file = reader.open_file(tmp_file_name,
mode='w',
format=self._format)
elif reader.__name__ == "scipy.io.netcdf":
tmp_file = reader.netcdf_file(tmp_file_name, mode='w')
elif reader.__name__ == "Dataset":
tmp_file = reader(tmp_file_name, 'w')
# Copy dimensions over
for dimension, length in self._df.dimensions.iteritems():
if reader.__name__ == "Nio":
tmp_file.create_dimension(dimension, length)
elif reader.__name__ in ["scipy.io.netcdf", "Dataset"]:
tmp_file.createDimension(dimension, len(length))
# Create new dimensions added by splicing
for variable, splice_info in self._sp_variables.iteritems():
dim_names, index_lists = splice_info
for dim_name, indexes in zip(dim_names, index_lists):
if dim_name not in tmp_file.dimensions.keys():
if reader.__name__ == "Nio":
tmp_file.create_dimension(dim_name,
len(np.unique(indexes)))
elif reader.__name__ in ["scipy.io.netcdf", "Dataset"]:
tmp_file.createDimension(dim_name,
len(np.unique(indexes)))
# Copy variable definitions over
for variable, data in self._df.variables.iteritems():
if variable not in self._rm_variables:
# If this variable is going to be spliced, find its new dimensions
if variable not in self._sp_variables:
dim_names = data.dimensions
else:
dim_names, index_lists = self._sp_variables[variable]
if reader.__name__ == "Nio":
tmp_file.create_variable(variable, data.typecode(),
dim_names)
elif reader.__name__ in ["scipy.io.netcdf", "Dataset"]:
tmp_file.createVariable(variable, data.dtype,
dim_names)
# Copy file attributes over
for attribute in dir(self._df):
if attribute[:2] != "__" and not hasattr(
tmp_file, attribute
) and attribute not in self._rm_attributes['__file__']:
setattr(tmp_file, attribute, getattr(self._df, attribute))
# Copy variable attributes over
for variable, data in tmp_file.variables.iteritems():
try:
for attribute in dir(self._df.variables[variable]):
if not hasattr(
data, attribute
) and attribute not in self._rm_attributes[variable]:
setattr(
data, attribute,
getattr(self._df.variables[variable],
attribute))
except KeyError:
for attribute in dir(self._df.variables[variable]):
if not hasattr(data, attribute):
setattr(
data, attribute,
getattr(self._df.variables[variable],
attribute))
# Copy variable data over (NetCDF4 Python appears to require this be done after setting all the variable attributes)
for variable, data in self._df.variables.iteritems():
if variable not in self._rm_variables:
tmp_variable = np.copy(self._df.variables[variable][:])
# Overwrite the data if we need to ...
if variable in self._ov_variables:
tmp_variable = np.copy(self._ov_variables[variable])
# Do splicing ...
if variable in self._sp_variables:
dim_names, index_lists = self._sp_variables[variable]
tmp_variable = np.copy(
self._df.variables[variable][np.ix_(*index_lists)])
# Put it in the file ...
tmp_file.variables[variable][:] = tmp_variable
# for dimension, length in tmp_file.dimensions.iteritems():
# print dimension, length
self._df.close()
if self._write:
# Move temporary file to original file's location
import shutil
if len(out_file_name) > 0:
shutil.move(tmp_file_name, out_file_name)
else:
shutil.move(tmp_file_name, self._file_name)
self._df = None
self._file_name = ""
self._format = ""
self._read = False
self._write = False
self._rm_variables = []
self._sp_variables = {}
self._rm_attributes = {'__file__': []}
return
