def run(self):
"""Execute of Code logic"""
source, dest = self._inputs
dest = trim_file_ext(self._inputs[1])
print('New Dest {}'.format(dest))
self._inputs = (source, dest)
super().run()
def main(image_filename, image_dir=None):
"""
convert image -> text file
:param image_filename: filename with path
:return:
"""
source_image = image_filename
image = cv2.imread(source_image, 0)
line_segments = get_line_segments(image)
line_image_offset = 0
i = 1
if not image_dir:
# TODO: Need to remove config usage from here
image_dir = os.path.join(config.TEXT_IMAGES,
trim_file_ext(os.path.basename(source_image)))
if os.path.isdir(image_dir):
print('Skipping the run as {} already exists'.format(image_dir))
return
else:
os.mkdir(image_dir)
for start, end in line_segments:
if abs(start - end) < 10:
continue
text_image = image[start - line_image_offset:end +
line_image_offset, :]
plt.imsave(os.path.join(image_dir, str(i) + '.png'), text_image)
i += 1
def update_data(image=None):
image_key = trim_file_ext(image)
if image_key in GetNewImage.pending_images:
GetNewImage.pending_images.remove(image_key)
GetNewImage.completed_images.append(image_key)
del GetNewImage.receipt_images[image_key]
print('Image Annotations Done for {}'.format(image))
def main(source_file, destination_file=None):
img = cv2.imread(source_file)
out_path = trim_file_ext(source_file) + '.tesseract.txt'
text = pytesseract.image_to_string(img,
lang='eng',
config=config.TESSERACT_CONFIG)
with open(out_path, "w") as fd:
fd.write(string_parser(text))
return out_path
def get_specific_image(image_file):
image_name = trim_file_ext(image_file)
send_info = {
'id': GetNewImage.receipt_images[image_name]['file'],
# 'url': GetNewImage.receipt_images[image_name]['url'],
'url': get_image_url(GetNewImage.receipt_images[image_name]),
'folder': '',
"annotations": []
}
return send_info
def rotate_image_with_east(img_filename, save_file=None):
image = plt.imread(img_filename)
text_filename = os.path.join(
os.path.dirname(img_filename),
trim_file_ext(os.path.basename(img_filename)) + '.txt')
raw_data, bag = read_data(text_filename)
new_plot_points = trim_outliers1(raw_data)
cnt = np.array(new_plot_points)
rect = cv2.minAreaRect(cnt)
box = cv2.boxPoints(rect)
box = np.int0(box)
warped = four_point_transform(image, box)
if save_file:
plt.imsave(save_file, warped)
else:
plt.figure(figsize=(10, 10))
plt.imshow(warped)
def gen_cropper_binarisation(image_name):
image_id = trim_file_ext(image_name)
_regenerate_binarisation_images(image_id)
return 'ok - _generate_binarisation_images'
def update_cropper_data(image_file):
image_id = trim_file_ext(image_file)
GetNewImage.receipt_images[image_id]['cropper_url'] = os.path.join(config.CROPPER_ROOT_DIR, image_file)
def update_binarisation_data(image_file):
image_id = trim_file_ext(image_file)
GetNewImage.receipt_images[image_id]['binarisation_url'] = os.path.join(config.BINARIZE_ROOT_DIR, image_file)