def datetime(self):
if self._datetime:
return self._datetime
if not self._path:
return None
try:
self._datetime = ts_parse_date_path(self._path)
except ValueError:
self._datetime = get_exif_date(self._path)
return self._datetime
def ts_to_tsnc(ts_path, tsnc_path):
log = logging.getLogger("timestreamlib")
# Get timestream images
imgs = list(ts_iter_images(ts_path))
# Make netcdf4 file
root = ncdf.Dataset(tsnc_path, 'w', format="NETCDF4")
root.createGroup('timestream')
# Make dimensions
# Kludge time!!!
root.createDimension('y', mat0.shape[0])
root.createDimension('x', mat0.shape[1])
root.createDimension('z', mat0.shape[2])
root.createDimension('t', None)
mat0 = None
mat_n = 0
while mat0 is None:
try:
mat0 = read_image(imgs[mat_n])
mat_n += 1
except IndexError:
raise ValueError("Didn't find a valid image in {}".format(ts_path))
# setup variables
times = root.createVariable("time", 'f8', ('t',))
times.units = "seconds since 1970-01-01 00:00:00.0"
times.calendar = "standard"
root.createVariable("z", 'u1', ('z',))
root.createVariable("y", 'u4', ('y',))
root.createVariable("x", 'u4', ('x',))
# create actual pixel array
px_type = 'u{:d}'.format(mat0.dtype.itemsize)
pixels = root.createVariable("pixel", px_type, ('t', 'y', 'x', 'z'),
zlib=True)
log.info("Created netcdf4 file {} with pixel array dimensions {!r}".format(
tsnc_path, pixels.shape))
# iteratively add images
count = 0
for img_n, mat in enumerate(imap(read_image, imgs)):
img = imgs[img_n]
if mat is None:
continue
n_dates = len(root.dimensions['t'])
time = ts_parse_date_path(img)
times[n_dates] = date2num([time, ], units=times.units,
calendar=times.calendar)
pixels[n_dates, :, :, :] = mat
count += 1
log.debug("Processed {}. Matrix shape is {!r}".format(img,
pixels.shape))
log.log(NOEOL, "Processed {: 5d} images.\r".format(count))
log.info("Processed {: 5d} images. ts_to_tsnc finished!".format(count))
root.close()
def ts_to_tsnc(ts_path, tsnc_path):
log = logging.getLogger("CONSOLE")
# Get timestream images
imgs = list(ts_iter_images(ts_path))
mats = ts_iter_numpy(imgs)
# Make netcdf4 file
root = ncdf.Dataset(tsnc_path, 'w', format="NETCDF4")
root.createGroup('timestream')
# Make dimensions
# Kludge time!!!
mat0 = next(ts_iter_numpy([imgs[0], ]))[1]
root.createDimension('y', mat0.shape[0])
root.createDimension('x', mat0.shape[1])
root.createDimension('z', mat0.shape[2])
root.createDimension('t', None)
# setup variables
times = root.createVariable("time", 'f8', ('t',))
times.units = "seconds since 1970-01-01 00:00:00.0"
times.calendar = "standard"
root.createVariable("z", 'u1', ('z',))
root.createVariable("y", 'u4', ('y',))
root.createVariable("x", 'u4', ('x',))
# create actual pixel array
px_type = 'u{d}'.format(mat0.dtype.itemsize)
pixels = root.createVariable("pixel", px_type, ('t', 'y', 'x', 'z'),
zlib=True)
log.info("Created netcdf4 file {} with pixel array dimensions {!r}".format(
tsnc_path, pixels.shape))
# iteratively add images
count = 0
for img, mat in mats:
n_dates = len(root.dimensions['t'])
time = ts_parse_date_path(img)
times[n_dates] = date2num([time, ], units=times.units,
calendar=times.calendar)
pixels[n_dates, :, :, :] = mat
count += 1
log.debug("Processed {}. Matrix shape is {!r}".format(img,
pixels.shape))
if count % 2 == 0:
log.log(NOEOL, "Processed {: 5d} images.\r".format(count))
log.info("Processed {: 5d} images. ts_to_tsnc finished!".format(count))
root.close()
