def getTimeIndxs(ncFile):
#Validate case of list of dates vs start and end date...
st = ed = None
if START_DATE and END_DATE:
st = ncu.parseDate(START_DATE)
ed = ncu.parseDate(END_DATE)
return ncu.getPeriodIndexes(ncFile, st, ed)
def getFileDate(fileName):
#Would be best if provided some kind of format or regex to exactly match the format of the date in the file name
#Regex not generic (just works for TRMM data)
dat = re.match(r'^.+\.(\d{10})', fileName).group(1)
return ncu.parseDate('-'.join([dat[:4], dat[4:6], dat[6:8]]) + 'T' + dat[8:])
#Means want to extract from a bunch of files, get list of files contained in it
print('multi file mode...')
ncFiles = ncu.getNCFiles(DATA)
#See if all the files (only the first, lets be optimist) can be opened with MFDataset (NC4 and time dim), else
#Open one by one.
if isPartOfMFDataset(ncFiles[0]):
print('Opening all files as MFDataset...')
else:
print('Files not eligible for MFDataset, opening one by one...')
#This could be that the files dont have a time dimension (then the date must be on the file name)
#or that the files are not in a supported version of netCDF (and could have time dimension anyways)
#Lets asume that the dates are in the file name then
ncFiles = [(getFileDate(f), f) for f in ncFiles]
#Filter files to get only the wanted period
if START_DATE and END_DATE:
ncFiles = [f for f in ncFiles if f[0] >= ncu.parseDate(START_DATE)
and f[0] <= ncu.parseDate(END_DATE)]
#else select files from wanted date list
ncFiles = sorted(ncFiles)
#Could be more specific (if for some case there are also a time dimension, but not now)
#Now get the indexes of the area wanted to extract
latIndexes = lonIndexes = None
lats = lons = None
with Dataset(ncFiles[0][1], 'r') as ncFile:
latIndexes, lonIndexes = getLocIndxs(ncFile)
lats = ncFile.variables[ncu.LAT_KEY][latIndexes]
lons = ncFile.variables[ncu.LON_KEY][lonIndexes]
#Once the dates (files) and location (indexes) are set, start writing the data to the file...
if FORMAT == FMT_CSV:
num2date(timeStep[0], timeUnits))
else:
#This may never happen, except for bad parsed date...
print('Can\'t extract time data, bad date format!')
#Once the dates and location (indexes) are set, start writing the data to the file...
else:
print('Files not eligible for MFDataset, opening one by one...')
#This could be that the files dont have a time dimension (then the date must be on the file name)
#or that the files are not in a supported version of netCDF (and could have time dimension anyways)
#Lets asume that the dates are in the file name then
ncFiles = [(getFileDate(f), f) for f in ncFiles]
#Filter files to get only the wanted period
if START_DATE and END_DATE:
ncFiles = [
f for f in ncFiles if f[0] >= ncu.parseDate(START_DATE)
and f[0] <= ncu.parseDate(END_DATE)
]
#else select files from wanted date list
ncFiles = sorted(ncFiles)
#Could be more specific (if for some case there are also a time dimension, but not now)
#Now get the indexes of the area wanted to extract
latIndexes = lonIndexes = None
lats = lons = None
with Dataset(ncFiles[0][1], 'r') as ncFile:
latIndexes, lonIndexes = getLocIndxs(ncFile)
lats = ncFile.variables[ncu.LAT_KEY][latIndexes]
lons = ncFile.variables[ncu.LON_KEY][lonIndexes]
#Once the dates (files) and location (indexes) are set, start writing the data to the file...