def get_stack(self, image_data, outfile=None, mask_path=None):
stack = None
if mask_path:
self.mask = self._get_mask_from_raster(mask_path)
if isinstance(image_data, str):
self.saved_array = image_data
stack = load(image_data)
elif isinstance(image_data, ImageStack):
self.data = image_data
stack = self._get_stack_channels()
if outfile:
print('saving image stack {}'.format(outfile))
save(outfile, stack)
self.final_shape = 1, stack.shape[1], stack.shape[2]
stack = stack.reshape((stack.shape[0], stack.shape[1] * stack.shape[2]))
stack[stack == 0.] = np.nan
if mask_path:
ms = self.mask.shape
msk = np.repeat(self.mask.reshape((ms[0], ms[1] * ms[2])), stack.shape[0], axis=0)
stack = marray(stack, mask=msk)
self.masked_data_stack = marray(stack, mask=np.isnan(stack))
self.n = self.masked_data_stack.shape[0]
del stack
self.new_array = zeros((1, self.masked_data_stack.shape[1]), dtype=float16)
def run(self):
if not self.vars:
print 'depthManipNC:\tERROR:\tvariables to save are no use:', self.vars
return
if not exists(self.fni):
print 'depthManipNC:\tERROR:\tinputfile name does not exists:', self.fni
return
if self.depthFlags == '':
print 'depthManipNC:\tWARNING:\tNo depth flags given, assuming surface values only.'
self.depthFlags = zeros(len(self.vars), dtype=int)
if len(self.vars) != len(self.depthFlags):
print 'depthManipNC:\tERROR:\tVariables do not match depth flags:', len(
self.vars), '!=', len(self.depthFlags)
return
self.varflag = {}
for var, flag in zip(self.vars, self.depthFlags):
self.varflag[var] = flag
if self.debug:
print 'depthManipNC:\tINFO:\topening dataset:\t', self.fni
nci = Dataset(self.fni, 'r') #Quiet =True)
#if self.depthFlags and 'zbnd' not in nci.variables.keys():
# print 'depthManipNC:\tERROR:\tdepthFlags is ',self.depthFlags,'but inputfile name does contain \'zbnd\''
# return
#check that there are some overlap between input vars and nci:
for v in self.vars:
if v in nci.variables.keys(): continue
print 'depthManipNC:\tERROR:\tvariable,', v, ', not found in ', self.fni
return
#create dataset and header.
if self.debug:
print 'depthManipNC:\tINFO:\tCreating a new dataset:\t', self.fno
nco = Dataset(self.fno, 'w')
for a in nci.ncattrs():
if self.debug:
print 'depthManipNC:\tINFO:\tcopying attribute: \t\"' + a + '\":\t', nci.getncattr(
a)
nco.setncattr(a, nci.getncattr(a))
appendToDesc = 'Reprocessed on ' + todaystr() + ' by ' + getuser(
) + ' using depthManipNC.py'
try:
nco.Notes = nci.Notes + '\n\t\t' + appendToDesc
except:
nco.Notes = appendToDesc
# list of variables to save, assuming some conventions
save = list(
set(nci.variables.keys()).intersection(set(self.alwaysInclude)))
save = list(set(sorted(save + self.vars)))
# create dimensions:
for d in nci.dimensions.keys():
if d in [
'time',
]: nco.createDimension(d, None)
elif d in [
'depth',
'z',
]: nco.createDimension(d, 1)
else: nco.createDimension(d, len(nci.dimensions[d]))
# create Variables:
for var in save:
nco.createVariable(var,
nci.variables[var].dtype,
nci.variables[var].dimensions,
zlib=True,
complevel=5)
# Long Names:
for var in save:
varln = ''
long_name = ''
try:
long_name = nci.variables[var].long_name
except:
if self.debug:
print 'depthManipNC:\tWarning:\tNo long_name for ', var
if var in self.vars:
long_name += ' ' + self.depthStrings[str(self.varflag[var])]
if self.timemean: long_name.replace('Daily', '')
if self.timemean: long_name.replace('Monthly', '')
nco.variables[var].long_name = long_name
if self.debug:
print 'depthManipNC:\tInfo:\tAdding long_name:', var, long_name
# Units:
for var in save:
units = ''
try:
units = nci.variables[var].units
except:
if self.debug:
print 'depthManipNC:\tWarning:\tNo units for ', var
if var in self.vars:
if self.varflag[var] == 1: units = units.replace('m^3', 'm^2')
nco.variables[var].units = units
if self.debug:
print 'depthManipNC:\tInfo:\tAdding units:', var, units
if 'zbnd' in nci.variables.keys(): self.zbnd = nci.variables['zbnd'][:]
if 'bathymetry' in nci.variables.keys():
self.bathy = nci.variables['bathymetry'][:]
# Fill Values:
for var in save:
if var not in self.vars: #no change
arr = nci.variables[var][:]
else:
flag = self.varflag[var]
if self.debug:
print 'depthManipNC:\tInfo:\tFilling var:', var, 'flag:', flag
if flag == 1:
arr = (nci.variables[var][:] * abs(
(self.zbnd[:, :, :, :, 1] - self.zbnd[:, :, :, :, 0]))
).sum(1)
arr = arr[:, None, :, :]
elif flag in [-2, -1, 0]:
arr = nci.variables[var][:, flag, :, :]
elif flag in [
-15,
]:
arr = self.bottomLayer(nci, var)
else:
arr = nci.variables[var][:, flag, :, :].mean(1)
arr = arr[:, None, :, :]
#while len(arr.shape) < len(nci.variables[var].dimensions): arr = marray(arr[None,:])
timeav = False
if self.timemean and len(
intersection(['time', 't'],
nci.variables[var].dimensions)):
timeav = True
if timeav:
if self.debug:
print 'depthManipNC:\tInfo:\tSaving time averaged var:', var
arr = marray([
arr.mean(0),
])
while len(arr.shape) < len(nci.variables[var].dimensions):
arr = marray(arr[None, :])
if self.debug:
print 'depthManipNC:\tInfo:\tSaving var:', var, arr.shape, '\tdims:', nci.variables[
var].dimensions
nco.variables[var][:] = arr
# Close netcdfs:
nco.close()
nci.close()
if self.debug:
print 'depthManipNC:\tINFO:\tsuccessfully created:\t', self.fno
return
def run(self):
if not self.vars:
print 'pruneNC:\tERROR:\tvariables to save are no use:', self.vars
return
if not exists(self.fni):
print 'pruneNC:\tERROR:\tinputfile name does not exists:', self.fni
return
nci = Dataset(self.fni, 'r') #Quiet =True)
if self.depthInt:
print 'FAIL: maybe you should look at the depthManip.py class instead. This one only removes variables from a netcdf.'
return
#check that there are some overlap between input vars and nci:
for v in self.vars:
if v in nci.variables.keys(): continue
print 'pruneNC:\tERROR:\tvariable,', v, ', not found in ', self.fni
return
#create dataset and header.
if self.debug:
print 'pruneNC:\tINFO:\tCreating a new dataset:\t', self.fno
nco = Dataset(self.fno, 'w')
for a in nci.ncattrs():
if self.debug:
print 'pruneNC:\tINFO:\tcopying attribute: \t\"' + a + '\":\t', nci.getncattr(
a)
nco.setncattr(a, nci.getncattr(a))
appendToDesc = 'Reprocessed on ' + todaystr() + ' by ' + getuser(
) + ' using pruneNC.py'
try:
nco.Notes = nci.Notes + '\n\t\t' + appendToDesc
except:
nco.Notes = appendToDesc
# list of variables to save, assuming some conventions
save = list(set(nci.variables.keys()).intersection(set(alwaysInclude)))
save = list(set(sorted(save + self.vars)))
# create dimensions:
for d in nci.dimensions.keys():
if d in [
'time',
]: nco.createDimension(d, None)
else: nco.createDimension(d, len(nci.dimensions[d]))
# create Variables:
for var in save:
nco.createVariable(var,
nci.variables[var].dtype,
nci.variables[var].dimensions,
zlib=True,
complevel=5)
# Long Names:
for var in save:
try:
long_name = nci.variables[var].long_name
except:
print 'pruneNC:\tWarning:\tNo long_name for ', var
long_name = var
#if self.timemean: long_name.replace('Daily', 'Monthly')
nco.variables[var].long_name = long_name
if self.debug:
print 'pruneNC:\t Adding long_name for ', var, long_name
# Units:
for var in save:
try:
nco.variables[var].units = nci.variables[var].units
except:
print 'pruneNC:\tWarning:\tNo units for ', var
# Fill Values:
for var in save:
if self.debug: print 'pruneNC:\tINFO:\tCopying ', var, ' ...'
arr = nci.variables[var][:]
if len(intersection(['time', 't'], nci.variables[var].dimensions)):
#####
# Take the time mean of the whole file
if self.timemean:
if self.debug:
print 'pruneNC:\tInfo:\tSaving time averaged var:', var
arr = marray([
arr.mean(0),
])
#while len(arr.shape) < len(nci.variables[var].dimensions): arr = marray(arr[None,:])
if self.timeindex == None: pass
else:
#####
# time index is an integer
if type(self.timeindex) in [type(1), type(1.)]:
arr = marray(arr[self.timeindex])
if self.debug:
print "pruneNC:\t", self.timeindex, arr.shape
#####
# time index is a slice
if type(self.timeindex) == type(slice(
0,
1,
)):
arr = marray(arr[self.timeindex]).mean(0)
# make sure it's the right shaped array
while len(arr.shape) < len(nci.variables[var].dimensions):
arr = marray(arr[None, ...])
if self.debug:
print 'pruneNC:\tInfo:\tSaving var:', var, arr.shape, '\tdims:', nci.variables[
var].dimensions
nco.variables[var][:] = arr
# Close netcdfs:
nco.close()
nci.close()
if self.debug:
print 'pruneNC:\tINFO:\tsuccessfully created:\t', self.fno
return
def run(self):
if not self.vars:
print 'depthManipNC:\tERROR:\tvariables to save are no use:', self.vars
return
if not exists(self.fni):
print 'depthManipNC:\tERROR:\tinputfile name does not exists:', self.fni
return
if self.depthFlags =='':
print 'depthManipNC:\tWARNING:\tNo depth flags given, assuming surface values only.'
self.depthFlags = zeros(len(self.vars), dtype=int)
if len(self.vars)!= len(self.depthFlags):
print 'depthManipNC:\tERROR:\tVariables do not match depth flags:', len(self.vars), '!=', len(self.depthFlags)
return
self.varflag={}
for var,flag in zip(self.vars, self.depthFlags):self.varflag[var] = flag
if self.debug: print 'depthManipNC:\tINFO:\topening dataset:\t', self.fni
nci = Dataset(self.fni,'r')#Quiet =True)
#if self.depthFlags and 'zbnd' not in nci.variables.keys():
# print 'depthManipNC:\tERROR:\tdepthFlags is ',self.depthFlags,'but inputfile name does contain \'zbnd\''
# return
#check that there are some overlap between input vars and nci:
for v in self.vars:
if v in nci.variables.keys():continue
print 'depthManipNC:\tERROR:\tvariable,' ,v,', not found in ',self.fni
return
#create dataset and header.
if self.debug: print 'depthManipNC:\tINFO:\tCreating a new dataset:\t', self.fno
nco = Dataset(self.fno,'w')
for a in nci.ncattrs():
if self.debug: print 'depthManipNC:\tINFO:\tcopying attribute: \t\"'+a+'\":\t', nci.getncattr(a)
nco.setncattr(a,nci.getncattr(a))
appendToDesc= 'Reprocessed on '+todaystr()+' by '+getuser()+' using depthManipNC.py'
try: nco.Notes = nci.Notes + '\n\t\t'+appendToDesc
except: nco.Notes = appendToDesc
# list of variables to save, assuming some conventions
alwaysInclude = ['time', 'lat','lon', 'latbnd', 'lonbnd']
save = list(set(nci.variables.keys()).intersection(set(alwaysInclude) ) )
save = list(set(sorted(save + self.vars)))
# create dimensions:
for d in nci.dimensions.keys():
if d in ['time',]: nco.createDimension(d, None)
elif d in ['depth', 'z',]: nco.createDimension(d, 1)
else: nco.createDimension(d, len(nci.dimensions[d]))
# create Variables:
for var in save: nco.createVariable(var, nci.variables[var].dtype, nci.variables[var].dimensions,zlib=True,complevel=5)
# Long Names:
for var in save:
varln = ''
long_name = ''
try: long_name=nci.variables[var].long_name
except:
if self.debug: print 'depthManipNC:\tWarning:\tNo long_name for ', var
if var in self.vars:
long_name += ' '+self.depthStrings[str(self.varflag[var])]
if self.timemean: long_name.replace('Daily', '')
if self.timemean: long_name.replace('Monthly', '')
nco.variables[var].long_name = long_name
if self.debug: print 'depthManipNC:\tInfo:\tAdding long_name:',var,long_name
# Units:
for var in save:
units = ''
try: units=nci.variables[var].units
except:
if self.debug: print 'depthManipNC:\tWarning:\tNo units for ', var
if var in self.vars:
if self.varflag[var] == 1: units = units.replace('m^3', 'm^2')
nco.variables[var].units=units
if self.debug: print 'depthManipNC:\tInfo:\tAdding units:',var,units
if 'zbnd' in nci.variables.keys(): self.zbnd =nci.variables['zbnd'][:]
if 'bathymetry' in nci.variables.keys(): self.bathy =nci.variables['bathymetry'][:]
# Fill Values:
for var in save:
if var not in self.vars: #no change
arr=nci.variables[var][:]
else:
flag = self.varflag[var]
if self.debug: print 'depthManipNC:\tInfo:\tFilling var:',var, 'flag:', flag
if flag == 1:
arr = (nci.variables[var][:] * abs((self.zbnd[:,:,:,:,1]-self.zbnd[:,:,:,:,0]))).sum(1)
arr= arr[:,None,:,:]
elif flag in [-2,-1,0]: arr =nci.variables[var][:,flag,:,:]
elif flag in [-15,]: arr =self.bottomLayer(nci, var)
else:
arr =nci.variables[var][:,flag,:,:].mean(1)
arr= arr[:,None,:,:]
#while len(arr.shape) < len(nci.variables[var].dimensions): arr = marray(arr[None,:])
timeav=False
if self.timemean and len(intersection(['time','t'], nci.variables[var].dimensions)): timeav=True
if timeav:
if self.debug: print 'depthManipNC:\tInfo:\tSaving time averaged var:',var
arr = marray([arr.mean(0),])
while len(arr.shape) < len(nci.variables[var].dimensions): arr = marray(arr[None,:])
if self.debug: print 'depthManipNC:\tInfo:\tSaving var:',var, arr.shape, '\tdims:', nci.variables[var].dimensions
nco.variables[var][:] = arr
# Close netcdfs:
nco.close()
nci.close()
if self.debug: print 'depthManipNC:\tINFO:\tsuccessfully created:\t', self.fno
return
def _saveNC_(self):
print 'GreenSeasXLtoNC:\tINFO:\tCreating a new dataset:\t', self.fno
nco = Dataset(self.fno,'w')
nco.setncattr('CreatedDate','This netcdf was created on the '+ucToStr(date.today()) +' by '+getuser()+' using GreenSeasXLtoNC.py')
nco.setncattr('Original File',self.fni)
for att in self.attributes.keys():
print 'GreenSeasXLtoNC:\tInfo:\tAdding Attribute:', att, self.attributes[att]
nco.setncattr(ucToStr(att), ucToStr(self.attributes[att]))
nco.createDimension('i', None)
nco.createVariable('index', 'i4', 'i',zlib=True,complevel=5)
for v in self.saveCols:
if v in self.dataIsAString:continue
print 'GreenSeasXLtoNC:\tInfo:\tCreating var:',v,self.ncVarName[v], self.dataTypes[v]
nco.createVariable(self.ncVarName[v], self.dataTypes[v], 'i',zlib=True,complevel=5)
nco.variables['index'].long_name = 'Excel Row index'
for v in self.saveCols:
if v in self.dataIsAString:continue
print 'GreenSeasXLtoNC:\tInfo:\tAdding var long_name:',v,self.ncVarName[v], self.lineTitles[v]
nco.variables[self.ncVarName[v]].long_name = self.lineTitles[v]
nco.variables['index'].units = ''
for v in self.saveCols:
if v in self.dataIsAString:continue
print 'GreenSeasXLtoNC:\tInfo:\tAdding var units:',v,self.ncVarName[v], self.unitTitles[v]
nco.variables[self.ncVarName[v]].units = self.unitTitles[v].replace('[','').replace(']','')
nco.variables['index'].metadata = ''
for v in self.saveCols:
if v in self.dataIsAString:continue
print 'GreenSeasXLtoNC:\tInfo:\tAdding meta data:',v#, self.metadata[v]
nco.variables[self.ncVarName[v]].metadata = self.metadata[v]
nco.variables['index'].xl_column = '-1'
for v in self.saveCols:
if v in self.dataIsAString:continue
print 'GreenSeasXLtoNC:\tInfo:\tAdding excell column name:',v,self.ncVarName[v],':', self.colnames[v]
nco.variables[self.ncVarName[v]].xl_column = self.colnames[v]
arr=[]
for a,val in enumerate(self.index.values()):
if not self.saveRows[a]: continue
if self.rowcounts[a] ==0 :continue
arr.append(val+1) # accounting for python 0th position is excels = 1st row
nco.variables['index'][:] = marray(arr)
for v in self.saveCols:
if v in self.dataIsAString:continue
print 'GreenSeasXLtoNC:\tInfo:\tSaving data:',v,self.ncVarName[v],
arr = []
for a,val in enumerate(self.data[v]):
if not self.saveRows[a]: continue
if self.rowcounts[a]==0:continue
arr.append(val)
nco.variables[self.ncVarName[v]][:] = marray(arr)
print len(arr)
print 'GreenSeasXLtoNC:\tInfo:\tCreated ',self.fno
nco.close()
def _getData_(self):
#loading file metadata
header = [h.value for h in self.datasheet.row(self.headR)]
units = [h.value for h in self.datasheet.row(self.unitR)]
lastMetaColumn = 20
locator = [h.value for h in self.datasheet.row(self.locR)[:lastMetaColumn]]
ckey={}
for n,l in enumerate(locator):
if l in [ 'time','t', 'Date& Time (local)']: ckey['time']=n
if l.lower() in [ 'lat', 'latitude']: ckey['lat']=n
if l.lower() in [ 'lon','long', 'longitude']: ckey['lon']=n
if l in [ 'Depth of sample [m]']: ckey['z']=n
if l in [ 'Depth of Sea [m]',]: ckey['bathy']=n
if l in [ 'UTC offset',]: ckey['tOffset']=n
if l in ['Institute',]: ckey['Institute']=n
bad_cells = [xl_cellerror,xl_cellempty,xl_cellblank]
metadataTitles = {r:h.value for r,h in enumerate(self.datasheet.col(self.metaC)[:self.maxMDR]) if h.ctype not in bad_cells}
endofHeadRow=max(metadataTitles.keys())
#create excel coordinates for netcdf.
colnames = {h: colname(h) for h,head in enumerate(self.datasheet.row(0))} # row number doesn't matter here
# which columns are we saving?
saveCols={}
lineTitles={}
unitTitles={}
attributes={}
for l,loc in enumerate(locator):
if loc in ['', None]:continue
print 'GreenSeasXLtoNC:\tInfo:\tFOUND:\t',l,'\t',loc, 'in locator'
saveCols[l] = True
lineTitles[l]=loc
unitTitles[l]=''
if loc.find('[') > 0:
unitTitles[l]=loc[loc.find('['):].replace(']','')
if header[5].find('Note')>-1:
attributes['Note'] = header[5]
header[5]=''
# flag for saving all columns:
if 'all' in self.datanames:
for head in header[lastMetaColumn:]:
if head == '': continue
self.datanames.append(head)
# add data columns titles to output to netcdf.
for h,head in enumerate(header):
if head == '':continue
if h in saveCols.keys(): continue
for d in self.datanames:
if h in saveCols.keys(): continue
if head.lower().find(d.lower()) > -1:
print 'GreenSeasXLtoNC:\tInfo:\tFOUND:\t',h,'\t',d, 'in ',head
saveCols[h] = True
lineTitles[h] = header[h]
unitTitles[h] = units[h]
saveCols = sorted(saveCols.keys())
print 'GreenSeasXLtoNC:\tInfo:\tInterograting columns:',saveCols
# Meta data for those columns with only one value:
ncVarName={}
allNames=[]
for h in saveCols:
name = self._getNCvarName_(lineTitles[h])
#ensure netcdf variable keys are unique:
if name in allNames:name+='_'+ucToStr(colnames[h])
allNames.append(name)
ncVarName[h] = name
# make an index to link netcdf back to spreadsheet
index = {}
for r in xrange(len(self.datasheet.col(saveCols[0])[self.maxMDR:])):
index[r] = r+self.maxMDR
#create data dictionary
data={}
tunit='seconds since 1900-00-00'
unitTitles[ckey['time']] = tunit
for d in saveCols:
tmpdata= self.datasheet.col(d)[self.maxMDR:]
arr = []
if d == ckey['time']: # time
for a in tmpdata[:]:
if a.ctype in bad_cells:
arr.append(default_fillvals['i8'])
else:
try: arr.append(int64(date2num(parse(a.value),units=tunit)))
except:
try:
arr.append(int(a.value))
print 'GreenSeasXLtoNC:\tWarning: Can not read time effecitvely:',int(a.value)
except: arr.append(default_fillvals['i8'])
data[d] = marray(arr)
continue
isaString = self._isaString_(lineTitles[d])
if isaString: #strings
for a in tmpdata[:]:
if a.ctype in bad_cells:
arr.append(default_fillvals['S1'])
else:
try: arr.append(ucToStr(a.value))
except: arr.append(default_fillvals['S1'])
else: # data
for a in tmpdata[:]:
if a.ctype in bad_cells:
arr.append(default_fillvals['f4'])
else:
try: arr.append(float(a.value))
except: arr.append(default_fillvals['f4'])
data[d] = marray(arr)
fillvals = default_fillvals.values()
# count number of data in each column:
print 'GreenSeasXLtoNC:\tInfo:\tCount number of data in each column...' # can be slow
datacounts = {d:0 for d in saveCols}
for d in saveCols:
for i in data[d][:]:
if i in ['', None, ]: continue
if i in fillvals: continue
datacounts[d]+=1
print 'GreenSeasXLtoNC:\tInfo:\tMax number of entries to in a column:', max(datacounts.values())
# list data columns with no data or only one value
removeCol=[]
for h in saveCols:
if datacounts[h] == 0:
print 'GreenSeasXLtoNC:\tInfo:\tNo data for column ',h,lineTitles[h],'[',unitTitles[h],']'
removeCol.append(h)
continue
col = sorted(data[h])
if col[0] == col[-1]:
if col[0] in fillvals:
print 'GreenSeasXLtoNC:\tInfo:\tIgnoring masked column', h, lineTitles[h],'[',unitTitles[h],']'
removeCol.append(h)
continue
print 'GreenSeasXLtoNC:\tInfo:\tonly one "data": ',lineTitles[h],'[',unitTitles[h],']','value:', col[0]
removeCol.append(h)
attributes[makeStringSafe(ucToStr(lineTitles[h]))] = ucToStr(col[0])
for r in removeCol:saveCols.remove(r)
print 'GreenSeasXLtoNC:\tInfo:\tnew file attributes:', attributes
print 'GreenSeasXLtoNC:\tInfo:\tFigure out which rows should be saved...'
saveRows = {a: False for a in index.keys()} #index.keys() are rows in data. #index.values are rows in excel.
rowcounts = {a: 0 for a in index.keys()}
for r in sorted(saveRows.keys()):
if data[ckey['time']][r] in ['', None,]:
print 'No time value:',r, data[ckey['time']][r]
continue
if data[ckey['time']][r] in fillvals:
print 'No time value:',r, data[ckey['time']][r]
continue
for d in saveCols:
if d<lastMetaColumn:continue
if data[d][r] in ['', None, ]: continue
if data[d][r] in fillvals: continue
rowcounts[r] += 1
saveRows[r] = True
print 'GreenSeasXLtoNC:\tInfo:\tMaximum number of rows to save: ',max(rowcounts.values()) #
#rowcounts = {d:0 for d in saveRows.keys()}
#for r in sorted(rowcounts.keys()):
# #if saveRows[r] == False: continue
# for d in saveCols:
# if d<20:continue
# if data[d][r] in ['', None, ]:continue
# if data[d][r] in fillvals: continue
# rowcounts[r] += 1
# get data type (ie float, int, etc...):
# netcdf4 requries some strange names for datatypes:
# ie f8 instead of numpy.float64
dataTypes={}
dataIsAString=[]
for h in saveCols:
dataTypes[h] = marray(data[h]).dtype
print 'GreenSeasXLtoNC:\tInfo:\ttype: ',ncVarName[h], h,'\t',dataTypes[h]
if dataTypes[h] == float64: dataTypes[h] = 'f8'
elif dataTypes[h] == int32: dataTypes[h] = 'i4'
elif dataTypes[h] == int64: dataTypes[h] = 'i8'
else:
dataTypes[h] = 'S1'
dataIsAString.append(h)
print 'GreenSeasXLtoNC:\tInfo:\tCreate MetaData...'
#create metadata.
metadata = {}
for h in saveCols:
if h in dataIsAString:continue
datacol = self.datasheet.col(h)[:]
colmeta = {metadataTitles[mdk]: datacol[mdk] for mdk in metadataTitles.keys() if metadataTitles[mdk] not in ['', None]}
md=' '
if len(colmeta.keys())> 20:
print 'Too many metadata'
print 'GreenSeasXLtoNC:\tWarning:\tMetadata reading failed. Please Consult original excel file for more info.'
metadata[h] = 'Metadata reading failed. Please Consult original excel file for more info.'
continue
for mdt,mdc in zip(colmeta.keys(),colmeta.values() ):
if mdc in ['', None]:continue
md +=ucToStr(mdt)+':\t'+ucToStr(mdc)+'\n '
#print md
metadata[h] = md
# save all info as public variables, so that it can be accessed if netCDF creation fails:
self.saveCols = saveCols
self.saveRows = saveRows
self.rowcounts=rowcounts
self.ncVarName = ncVarName
self.dataTypes = dataTypes
self.dataIsAString=dataIsAString
self.metadata=metadata
self.colnames = colnames
self.data = data
self.lineTitles = lineTitles
self.unitTitles = unitTitles
self.attributes = attributes
self.index = index
def run(self):
if not exists(self.fni):
print 'convertToOneDNC:\tERROR:\tinputfile name does not exists:', self.fni
return
nci = Dataset(self.fni,'r')
if not self.vars:
self.vars = nci.variables.keys() # save all
#check that there are some overlap between input vars and nci:
for v in self.vars:
if v in nci.variables.keys():continue
print 'convertToOneDNC:\tERROR:\tvariable,' ,v,', not found in ',self.fni
return
#create dataset and header.
if self.debug: print 'convertToOneDNC:\tINFO:\tCreating a new dataset:\t', self.fno
nco = Dataset(self.fno,'w')
for a in nci.ncattrs():
if self.debug: print 'convertToOneDNC:\tINFO:\tcopying attribute: \t\"'+a+'\":\t', nci.getncattr(a)
nco.setncattr(a,nci.getncattr(a))
appendToDesc= 'Reprocessed on '+todaystr()+' by '+getuser()+' using convertToOneDNC.py'
try: nco.Notes = nci.Notes + '\n\t\t'+appendToDesc
except: nco.Notes = appendToDesc
save = list(set(nci.variables.keys()).intersection(set(alwaysInclude) ) )
save = sorted(list(set(sorted(save + self.vars))))
# test to find out which coordinates should be saved.
if not self.dictToKeep:
CoordsToKeep,save=getCoordsToKeep(nci,save,newMask=self.newMask,debug = self.debug)
else:
CoordsToKeep = self.dictToKeep
# create dimensions:
#for d in nci.dimensions.keys():
# if d in ['time',]: nco.createDimension(d, None)
# else: nco.createDimension(d, len(nci.dimensions[d]))
nco.createDimension('index', None)
# create Variables:
nco.createVariable('index', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_t', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_z', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_y', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_x', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
for var in save:
nco.createVariable(var, nci.variables[var].dtype, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
# Long Names:
nco.variables['index'].long_name='index'
nco.variables['index_t'].long_name='index - time'
nco.variables['index_z'].long_name='index - depth'
nco.variables['index_y'].long_name='index - latitude'
nco.variables['index_x'].long_name='index - longitude'
for var in save:
try: long_name=nci.variables[var].long_name
except:
if self.debug: print 'convertToOneDNC:\tWarning:\tNo long_name for ', var
long_name = var
nco.variables[var].long_name=long_name
if self.debug: print 'convertToOneDNC:\t Adding long_name for ', var, long_name
# Units:
nco.variables['index'].units=''
nco.variables['index_t'].units=''
nco.variables['index_z'].units=''
nco.variables['index_y'].units=''
nco.variables['index_x'].units=''
for var in save:
try: nco.variables[var].units=nci.variables[var].units
except: print 'convertToOneDNC:\tWarning:\tNo units for ', var
# Fill Values:
sorted_Coords = sorted(CoordsToKeep.iteritems(), key=itemgetter(1))
data={}
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index ...' , len(sorted_Coords)
# nco.variables['index'][:] = [ int(a[1]) for a in sorted_Coords]
nco.variables['index'][:] = array([ a[1] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index t ...'
nco.variables['index_t'][:] = array([a[0][0] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index z ...'
nco.variables['index_z'][:] = array([a[0][1] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index y ...'
nco.variables['index_y'][:] = array([a[0][2] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index x ...'
nco.variables['index_x'][:] = array([a[0][3] for a in sorted_Coords])
nco.sync()
for var in save:
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying ', var, ' ...'
arr = nci.variables[var][:]
outarr = []
if arr.ndim ==1:
if var.lower() in ['time','time_counter','t']: d = 0
if var.lower() in ['depth','deptht',]: d = 1
if var.lower() in ['latbnd','lat','latitude']: d = 2
if var.lower() in ['lonbnd','lon','longitude']:d = 3
#for c in (CoordsToKeep.keys()):
for c in sorted_Coords:
outarr.append(arr[c[0][d]])
try: print var, d
except: var, "not found"
elif arr.ndim ==2:
if var.lower() in ['nav_lat','nav_lon']: d = (2,3)
print var, 'lendth : 2', d
for c in sorted_Coords:
#for c in sorted(CoordsToKeep.keys()):
outarr.append(arr[(c[0][2:])])
else:
#for c in sorted(CoordsToKeep.keys()):
for c in sorted_Coords:
outarr.append(arr[c[0]])
outarr= marray(outarr)
if self.debug: print 'convertToOneDNC:\tINFO:\tSaving var:',var, arr.shape, '->', outarr.shape , 'coords:',len(sorted_Coords)
nco.variables[var][:] =outarr
nco.sync()
# Close netcdfs:
nco.close()
nci.close()
if self.debug: print 'convertToOneDNC:\tINFO:\tsuccessfully created:\t', self.fno
return
def run(self):
if not self.vars:
print 'pruneNC:\tERROR:\tvariables to save are no use:', self.vars
return
if not exists(self.fni):
print 'pruneNC:\tERROR:\tinputfile name does not exists:', self.fni
return
nci = Dataset(self.fni,'r')#Quiet =True)
if self.depthInt:
print 'FAIL: maybe you should look at the depthManip.py class instead. This one only removes variables from a netcdf.'
return
#check that there are some overlap between input vars and nci:
for v in self.vars:
if v in nci.variables.keys():continue
print 'pruneNC:\tERROR:\tvariable,' ,v,', not found in ',self.fni
return
#create dataset and header.
if self.debug: print 'pruneNC:\tINFO:\tCreating a new dataset:\t', self.fno
nco = Dataset(self.fno,'w')
for a in nci.ncattrs():
if self.debug: print 'pruneNC:\tINFO:\tcopying attribute: \t\"'+a+'\":\t', nci.getncattr(a)
nco.setncattr(a,nci.getncattr(a))
appendToDesc= 'Reprocessed on '+todaystr()+' by '+getuser()+' using pruneNC.py'
try: nco.Notes = nci.Notes + '\n\t\t'+appendToDesc
except: nco.Notes = appendToDesc
# list of variables to save, assuming some conventions
alwaysInclude = ['time', 'lat','lon', 'latbnd', 'lonbnd']
save = list(set(nci.variables.keys()).intersection(set(alwaysInclude) ) )
save = list(set(sorted(save + self.vars)))
# create dimensions:
for d in nci.dimensions.keys():
if d in ['time',]: nco.createDimension(d, None)
else: nco.createDimension(d, len(nci.dimensions[d]))
# create Variables:
for var in save: nco.createVariable(var, nci.variables[var].dtype, nci.variables[var].dimensions,zlib=True,complevel=5)
# Long Names:
for var in save:
try: long_name=nci.variables[var].long_name
except:
print 'pruneNC:\tWarning:\tNo long_name for ', var
long_name = var
if self.timemean: long_name.replace('Daily', 'Monthly')
nco.variables[var].long_name=long_name
if self.debug: print 'pruneNC:\t Adding long_name for ', var, long_name
# Units:
for var in save:
try: nco.variables[var].units=nci.variables[var].units
except: print 'pruneNC:\tWarning:\tNo units for ', var
# Fill Values:
for var in save:
if self.debug: print 'pruneNC:\tINFO:\tCopying ', var, ' ...'
arr = nci.variables[var][:]
if self.timemean and len(intersection(['time','t'], nci.variables[var].dimensions)):
if self.debug: print 'pruneNC:\tInfo:\tSaving time averaged var:',var
arr = marray([arr.mean(0),])
while len(arr.shape) < len(nci.variables[var].dimensions): arr = marray(arr[None,:])
if self.debug: print 'pruneNC:\tInfo:\tSaving var:',var, arr.shape, '\tdims:', nci.variables[var].dimensions
nco.variables[var][:] =arr
# Close netcdfs:
nco.close()
nci.close()
if self.debug: print 'pruneNC:\tINFO:\tsuccessfully created:\t', self.fno
return
def run(self):
if not exists(self.fni):
print 'convertToOneDNC:\tERROR:\tinputfile name does not exists:', self.fni
assert False
return
nci = Dataset(self.fni,'r')
if not self.vars:
self.vars = nci.variables.keys() # save all
#check that there are some overlap between input vars and nci:
for v in self.vars:
if v in nci.variables.keys():continue
print 'convertToOneDNC:\tERROR:\tvariable,' ,v,', not found in ',self.fni
return
#create dataset and header.
if self.debug: print 'convertToOneDNC:\tINFO:\tCreating a new dataset:\t', self.fno
nco = Dataset(self.fno,'w')
for a in nci.ncattrs():
if self.debug: print 'convertToOneDNC:\tINFO:\tcopying attribute: \t\"'+a+'\":\t', nci.getncattr(a)
nco.setncattr(a,nci.getncattr(a))
appendToDesc= 'Reprocessed on '+todaystr()+' by '+getuser()+' using convertToOneDNC.py'
try: nco.Notes = nci.Notes + '\n\t\t'+appendToDesc
except: nco.Notes = appendToDesc
save = list(set(nci.variables.keys()).intersection(set(alwaysInclude) ) )
save = sorted(list(set(sorted(save + self.vars))))
# test to find out which coordinates should be saved.
if not self.dictToKeep:
CoordsToKeep,save=getCoordsToKeep(nci,save,newMask=self.newMask,debug = self.debug)
else:
CoordsToKeep = self.dictToKeep
print "convertToOneDNC:\tinfo:\tCoordsToKeep:", save#, self.dictToKeep
# create dimensions:
nco.createDimension('index', None)
# create Variables:
nco.createVariable('index', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_t', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_z', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_y', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
nco.createVariable('index_x', int64, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
for var in save:
if var in ['index','index_t','index_z','index_y','index_x']:continue
nco.createVariable(var, nci.variables[var].dtype, ['index',],zlib=True,complevel=5)#,chunksizes=10000)
# Long Names:
nco.variables['index'].long_name='index'
nco.variables['index_t'].long_name='index - time'
nco.variables['index_z'].long_name='index - depth'
nco.variables['index_y'].long_name='index - latitude'
nco.variables['index_x'].long_name='index - longitude'
for var in save:
try: long_name=nci.variables[var].long_name
except:
if self.debug: print 'convertToOneDNC:\tWARNING:\tNo long_name for ', var
long_name = var
nco.variables[var].long_name=long_name
if self.debug: print 'convertToOneDNC:\tINFO:\tAdding long_name for ', var, '\t',long_name
# Units:
nco.variables['index'].units=''
nco.variables['index_t'].units=''
nco.variables['index_z'].units=''
nco.variables['index_y'].units=''
nco.variables['index_x'].units=''
for var in save:
try:
nco.variables[var].units=nci.variables[var].units
if self.debug: print 'convertToOneDNC:\tINFO:\tAdding units for ', var, '\t',nci.variables[var].units
except: print 'convertToOneDNC:\tWARNING:\tNo units for ', var
# Fill Values:
def itemsgetter(a):
return a[1][0]
sorted_Coords = sorted(CoordsToKeep.iteritems(), key=itemsgetter)
print "convertToOneDNC:\tINFO:\tsorted_Coords:",sorted_Coords[0],sorted_Coords[-1]
data={}
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index ...' , len(sorted_Coords)
# nco.variables['index'][:] = [ int(a[1]) for a in sorted_Coords]
nco.variables['index'][:] = array([ a[1][0] for a in sorted_Coords])
nco.sync()
# 4D coordinates:
if len(sorted_Coords[0][0]) ==4:
print "4D:", sorted_Coords[0][0]
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index t ...'
nco.variables['index_t'][:] = array([a[0][0] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index z ...'
nco.variables['index_z'][:] = array([a[0][1] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index y ...'
nco.variables['index_y'][:] = array([a[0][2] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index x ...'
nco.variables['index_x'][:] = array([a[0][3] for a in sorted_Coords])
nco.sync()
if len(sorted_Coords[0][0]) ==3:
print "3D:", sorted_Coords[0][0]
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index t ...'
nco.variables['index_t'][:] = array([a[0][0] for a in sorted_Coords])
nco.sync()
#if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index z ...'
nco.variables['index_z'][:] = zeros(len(sorted_Coords))
#nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index y ...'
nco.variables['index_y'][:] = array([a[0][1] for a in sorted_Coords])
nco.sync()
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying index x ...'
#tempArr = []
#for a in sorted_Coords:
# print a[0]
# tempArr.append(a[0][2])
nco.variables['index_x'][:] = array([a[0][2] for a in sorted_Coords])
nco.sync()
errorScript = "If it is failing here, then you need to check that your dimensions are names correctly in netcdf_manip/alwaysInclude.py:timeNames, depthNames, lonnames and latnames"
for var in save:
if self.debug: print 'convertToOneDNC:\tINFO:\tCopying ', var, ' ...'
arr = nci.variables[var][:]
if len(arr)==0:
print 'convertToOneDNC:\tWarning:\tIt looks like the netcdf ',self.fni,'does not contain the variable', var
outarr = []
if arr.ndim ==1 and len(sorted_Coords[0][0]) == 4:
if var.lower() in ['time','time_counter','t','month', 'time_centered']: d = 0
if var.lower() in timeNames: d = 0
if var.lower() in depthNames: d = 1
if var.lower() in latnames: d = 2
if var.lower() in lonnames: d = 3
print var, 'convertToOneDNC:\tINFO:\tndim: (1-4)',arr.ndim, var, sorted_Coords[0][0], d, #arr[0]
for c in sorted_Coords:
try: outarr.append(arr[c[0][d]])
except: raise AssertionError(errorScript)
try: print var, d
except: var, "not found"
elif arr.ndim ==1 and len(sorted_Coords[0][0]) ==1:
d = 0
print var, 'convertToOneDNC:\tINFO:\tndim: (1-1)',arr.ndim,var, sorted_Coords[0][0], d
for c in sorted_Coords:
try: outarr.append(arr[c[0][d]])
except: raise AssertionError(errorScript)
try: print var, d
except: var, "not found"
elif arr.ndim ==1 and len(sorted_Coords[0][0]) ==3:
if var.lower() in ['time','time_counter','t','month','time_centered',]: d = 0
if var.lower() in timeNames: d = 0
#if var.lower() in depthNames: d = 1
if var.lower() in latnames: d = 1
if var.lower() in lonnames: d = 2
#for c in (CoordsToKeep.keys()):
print var, 'convertToOneDNC:\tINFO:\tndim: (1-3)',arr.ndim,var, sorted_Coords[0][0], d
for c in sorted_Coords:
try: outarr.append(arr[c[0][d]])
except: raise AssertionError(errorScript)
try: print var, d
except: var, "not found"
elif arr.ndim ==2:
if var.lower() in ['nav_lat','nav_lon']:
d = (2,3)
elif var.lower() in ['deptht','depthu','depthv','latitude','longitude','depth',]:
d = (0,1)
elif var.lower() in ['mask'] and len(sorted_Coords[0][0]) ==3:
d = (1,2) # because of MLD datasets.
else: d = (0,1)
print var, 'convertToOneDNC:\tINFO:\tndim: (2)',arr.ndim,var, sorted_Coords[0][0], d
for c in sorted_Coords: #for c in sorted(CoordsToKeep.keys()):
try: outarr.append(arr[(c[0][d[0]],c[0][d[1]])])
except: raise AssertionError(errorScript)
elif arr.ndim ==3:
print var, 'convertToOneDNC:\tINFO:\tndim: (3)',arr.ndim,var, sorted_Coords[0][0]
for c in sorted_Coords:
if len(c[0]) == 4: outarr.append(arr[(c[0][0],c[0][2],c[0][3])])
if len(c[0]) == 3: outarr.append(arr[(c[0][0],c[0][1],c[0][2])])
elif arr.ndim ==4:
print var, 'convertToOneDNC:\tINFO:\tndim: (4)',arr.ndim,var, sorted_Coords[0][0]
for c in sorted_Coords:
#print 'convertToOneD:',var, c, c[0], arr.shape, arr[c[0]]
outarr.append(arr[ (c[0][0],c[0][1],c[0][2],c[0][3]) ])
else:
print "How many dimensions?", arr.ndim, len(sorted_Coords[0][0])
assert False
outarrnm = np.array(outarr)
print var, [outarrnm.min(),outarrnm.max()], '(no mask)'
outarr = marray(outarr)
print var, [outarr.min(),outarr.max()], '(masked)'
if self.debug: print 'convertToOneDNC:\tINFO:\tSaving var:',var, arr.shape, '->', outarr.shape , 'coords:',len(sorted_Coords), [outarr.min(),outarr.max()]
nco.variables[var][:] = outarr
nco.sync()
# Close netcdfs:
nco.close()
nci.close()
if self.debug: print 'convertToOneDNC:\tINFO:\tsuccessfully created:\t', self.fno
#nc = Dataset(self.fno, 'r')
#lon = nc.variables['lon'][:]
#print lon.min(), lon.max()
#if self.fno.find('Model')>-1: assert 0
return