def explore_tomo_dataset(folder, file_prefix):
if not(os.path.isdir(folder)): raise IOError('explore_tomo_dataset(): Not a folder : %s' % folder)
pref = os.path.join(folder, file_prefix)
# TODO : more flexible file suffix
fl = glob.glob(pref + "????.edf")
fl.sort()
if fl == []:
print('Warning: explore_tomo_dataset(): could not find any EDF file in %s matching the file prefix %s' % (folder, file_prefix))
n = len(fl)
# Read first projection file, assuming the others follow the same format
proj0 = edf_read(fl[0])
npx_R, npx_C = proj0.shape
res = {}
res['width'] = npx_C
res['height'] = npx_R
res['datatype'] = proj0.dtype
res['number'] = n
res['files'] = fl
return res
def apply_processing(myprocessing, dataset, options, extra_args=None):
"""
Apply a processing on a whole dataset. The processing function is defined by the user.
myprocessing : function taking a numpy.ndarray (and possibly extra_args) as an input, and returning a numpy.ndarray
type: function
dataset : dataset information returned by explore_tomo_folder()
type : dict
folder_out : folder where the results will be saved
type: str
file_prefix_out : file prefix for the reconstruction (eg. "rec_")
type: str
extra_args : (optionnal) extra arguments that should be passed to the user's processing function
type: tuple
TODO :
- Output format other than EDF ; and suffix possibly different from _0123.edf
- It assumes that the data is stored "alphabetically" (data of file_2718.edf comes after data of file_2717.edf)
"""
# Arg check
# ----------
if not isinstance(dataset, dict):
raise ValueError('apply_processing(): second argument should be a Python dictionary')
if not callable(myprocessing):
raise ValueError('apply_processing(): first argument should be a callable custom function. It has to take an array as an input and return an array.')
if options and not isinstance(options, dict):
raise ValueError('apply_processing() : options should be a dictionary')
# Parse the execution options
# ----------
_verbose = True # Verbose is true by default.
_start = 0
_number = dataset['number']
_end = _number -1
# Turns into a case insensitive search
options = dict((k.lower(), options[k]) for k in options)
if options.has_key('verbose'):
_verbose = True if bool(options['verbose']) else False
if options.has_key('file_start'): tmp_str1 = get_tomo_file_num(options['file_start'])
else: tmp_str1 = '0'
if options.has_key('file_end'): tmp_str2 = get_tomo_file_num(options['file_end'])
else: tmp_str2 = str(_end)
try:
_start = int(tmp_str1)
_end = int(tmp_str2)
except ValueError:
raise ValueError('apply_processing(): options file_start and file_end should be in format "path_0123.edf"') # TODO
if options.has_key('start'): _start = int(options['start'])
if options.has_key('end'): _end = int(options['end'])
if (_start > _end): raise ValueError('apply_processing(): option start or file_start (%d) is greater than end or file_end (%d)' % (_start, _end))
if (_end > _number): raise ValueError('apply_processing(): option end or file_end (%d) is greater than available files (%d)' % (_end, _number))
try:
folder_out = options['output_folder']
file_prefix_out = options['output_file_prefix']
except KeyError:
raise Exception('apply_processing(): please provide "output_folder" and "output_file_prefix" in the options')
# Processing
# ----------
for i in range(_start, _end +1): # TODO : what if the ordering gives a result different from "for f in dataset['files']" ?
f = dataset['files'][i]
if _verbose: print("Processing file %s" % os.path.split(f)[-1])
arr = edf_read(f)
if extra_args:
arr_out = myprocessing(arr, extra_args)
else:
arr_out = myprocessing(arr)
fname_out = os.path.join(folder_out, file_prefix_out) + str("%04d" % i) + ".edf"
edf_write(arr_out, fname_out)
if _verbose: print("Wrote %s" % fname_out)