def denoise_file(source, destination, kappa, iterations, manipulations={}, verbose=False):
"""
Ferforms denoising of the source file and outputs the response to the
destination file.
Parameters
----------
source: str
Path to the file to denoise.
destination : str
Path to the destination (output) file.
kappa : float
The denoising factor. Allowed range: [0.0-1.0]
iterations : int
Number of interations to do with the denoising.
manipulations : dict
Manipulation dict is received to maintain an API thats similar
to the other backends, but it's not used for anything as the
functionality is missing in this backend.
Returns
-------
output : str
Empty string if OK. Error output if something didn't go right.
Example
-------
Error message returned if source does not exist
>>> denoise_file('non-existant.jpg', 'foo.jpg', 0.1, 10)
'Source file [non-existant.jpg] could not be loaded.'
Empty string is returned upon successful denoising
>>> denoise_file('../../assets/disasterbefore.jpg', '../../tmp/out.jpg', 0.1, 1)
"""
# Open image
try:
image = Image.open(source);
except IOError:
return 'Source file [%s] could not be loaded.' % source;
# Get pixel information from image
data = list(image.getdata());
if verbose:
print 'Image data read from %s' % source;
# Get image dimensions
width, height = image.size;
if verbose:
print 'Image shape:';
print image.size;
# Check if the image is a color image
if type(data[0]) is tuple:
return 'Color images are not supported in the pure python backend.';
# Give warning if manipulation is requested
if should_do_manipulation(manipulations):
return 'Manipulations are not supported i pure python backend';
if verbose:
print 'Start data processing';
# Perform denoising of image
denoised_data = denoise_image_data(data, width, height, kappa, iterations);
if verbose:
print 'Finished processing data';
# Ouput denoised image data to new file
im = Image.new("L", (width, height));
im.putdata(denoised_data);
# Save file
try:
im.save(destination);
if verbose:
print 'Wrote output data to %s' % destination;
except IOError:
return 'Destination file [%s] was not writeable.' % args.destination;
exit();
def denoise_file(source, destination, kappa, iterations, manipulations={}, verbose=False):
"""
Ferforms denoising of the source file and outputs the response to the
destination file.
Parameters
----------
source: str
Path to the file to denoise.
destination : str
Path to the destination (output) file.
kappa : float
The denoising factor. Allowed range: [0.0-1.0]
iterations : int
Number of interations to do with the denoising.
manipulations : dict
Specification of the manipulations to do. Expected keys: lr (R component
in RGB), lg (G component in RGB), lb (B component in RGB), lh (H component
in HSI), ls (S component in HSI) and li (I component in HSI).
Returns
-------
output : str
Empty string if OK. Error output if something didn't go right.
Example
-------
Error message returned if source does not exist
>>> denoise_file('non-existant.jpg', 'foo.jpg', 0.1, 10)
'Source file [non-existant.jpg] could not be loaded.'
Error message returned if color image is provided
>>> denoise_file('../../assets/disasterbeforecolor.jpg', 'foobar.jpg', 0.1, 10);
'Color images are not supported in C backend.'
Error message returned if manipulations are requested
>>> denoise_file('../../assets/disasterbefore.jpg', 'foobar.jpg', 0.1, 10, {'lr': 100});
'Manipulations are not supported i C backend'
Empty string is returned upon successful denoising
>>> denoise_file('../../assets/disasterbefore.jpg', '../../tmp/out.jpg', 0.1, 10)
''
"""
try:
data = np.array(Image.open(source))
except IOError:
return 'Source file [%s] could not be loaded.' % source;
if verbose:
print 'Image data read from %s' % source;
# Get width and height based on the data shape
if len(data.shape) > 2:
return 'Color images are not supported in C backend.';
# Give warning if manipulation is requested
if should_do_manipulation(manipulations):
return 'Manipulations are not supported i C backend';
script_path = path.join(
path.dirname(__file__),
'../../lib/denoise'
);
# Parse/split the command into arguments
arguments = arguments = [
script_path,
"%f" % kappa,
"%d" % iterations,
source,
destination
];
if verbose:
print 'Forking new process from command:\n%s' % script_path;
print 'Arguments:';
print arguments[1:];
try:
# Open a sub process with the arguments
process = subprocess.Popen(arguments, stdout=subprocess.PIPE);
# Get the piped output and return
out, err = process.communicate();
except OSError as e:
raise Exception('Execution of command failed');
return out;
