from box import load_box_file
def box_widths(boxes, path):
"""Returns a dict mapping box width -> count."""
counts = defaultdict(int)
for box in boxes:
width = box.right - box.left
if 16 <= width and width <= 19:
sys.stderr.write('%s: %s %s\n' % (width, path, box))
counts[width] += 1
return counts
def counts_to_list(counts):
high_val = max(counts.keys())
return [counts[x] for x in range(0, high_val + 1)]
if __name__ == '__main__':
count_lists = []
for path in sys.argv[1:]:
boxes = load_box_file(path)
count_lists.append(counts_to_list(box_widths(boxes, path)))
max_width = max(len(x) for x in count_lists)
print('\t'.join([''] + [str(x) for x in range(0, max_width)]))
for path, counts in zip(sys.argv[1:], count_lists):
print('%s\t%s' % (path, '\t'.join(str(x) for x in counts)))
h = box.top - box.bottom
assert h > 0
if w < 21: return [box] # probably just a single letter.
if h > w: return [box] # maybe it's just large, not wide
num_ways = int(round(w / 12.0))
assert num_ways > 1, w
boxes = []
for i in range(0, num_ways):
b = copy.deepcopy(box)
b.left = box.left + int(round((1.0 * i / num_ways * w)))
b.right = box.left + int(round((1.0 * (i + 1) / num_ways * w)))
boxes.append(b)
return boxes
def split_boxes(boxes):
out_boxes = []
for box in boxes:
out_boxes += split_box(box)
return out_boxes
if __name__ == '__main__':
for path in sys.argv[1:]:
boxes = load_box_file(path)
out_boxes = split_boxes(boxes)
out_path = path.replace('.box', '.split.box')
open(out_path, 'w').write('\n'.join(str(x) for x in out_boxes))
def padded_box(box, pad_width, pad_height):
"""Adds some additional margin around the box."""
return BoxLine(box.letter, box.left - pad_width, box.top + pad_height,
box.right + pad_width, box.bottom - pad_height, box.page)
def crop_image_to_box(im, box):
"""
Returns a new image containing the pixels inside box.
This accounts for BoxLine measuring pixels from the bottom up, whereas
Image objects measure from the top down.
"""
w, h = im.size
box = [
int(round(v))
for v in (box.left, h - box.top, box.right, h - box.bottom)
]
return im.crop(box)
if __name__ == '__main__':
_, box_path, image_path, out_image_path = sys.argv
boxes = load_box_file(box_path)
big_box = find_box_extrema(boxes)
pad_box = padded_box(big_box, 20, 20)
im = Image.open(image_path)
cropped_im = crop_image_to_box(im, pad_box)
cropped_im.save(out_image_path)
# From http://stackoverflow.com/questions/600268/mkdir-p-functionality-in-python
def mkdir_p(path):
try:
os.makedirs(path)
except OSError as exc: # Python >2.5
if exc.errno == errno.EEXIST:
pass
else:
raise
if __name__ == '__main__':
_, box_path, image_path, output_dir = sys.argv
boxes = box.load_box_file(box_path)
im = Image.open(image_path)
w, h = im.size
for idx, box in enumerate(boxes):
x1, x2 = box.left, box.right
y1, y2 = h - box.top, h - box.bottom
assert x2 > x1
assert y2 > y1
char_im = im.crop((x1, y1, x2, y2))
out_dir, out_file = path_for_letter(output_dir, image_path, idx,
box.letter)
out_path = os.path.join(out_dir, out_file)
mkdir_p(out_dir)
def padded_box(box, pad_width, pad_height):
"""Adds some additional margin around the box."""
return BoxLine(box.letter,
box.left - pad_width,
box.top + pad_height,
box.right + pad_width,
box.bottom - pad_height,
box.page)
def crop_image_to_box(im, box):
"""Returns a new image containing the pixels inside box.
This accounts for BoxLine measuring pixels from the bottom up, whereas
Image objects measure from the top down.
"""
w, h = im.size
box = [int(round(v)) for v in (box.left, h - box.top, box.right, h - box.bottom)]
return im.crop(box)
if __name__ == '__main__':
_, box_path, image_path, out_image_path = sys.argv
boxes = load_box_file(box_path)
big_box = find_box_extrema(boxes)
pad_box = padded_box(big_box, 20, 20)
im = Image.open(image_path)
cropped_im = crop_image_to_box(im, pad_box)
cropped_im.save(out_image_path)
return os.path.join(output_dir, letter), '%s.%s.png' % (image_base, idx)
# From http://stackoverflow.com/questions/600268/mkdir-p-functionality-in-python
def mkdir_p(path):
try:
os.makedirs(path)
except OSError as exc: # Python >2.5
if exc.errno == errno.EEXIST:
pass
else: raise
if __name__ == '__main__':
_, box_path, image_path, output_dir = sys.argv
boxes = box.load_box_file(box_path)
im = Image.open(image_path)
w, h = im.size
for idx, box in enumerate(boxes):
x1, x2 = box.left, box.right
y1, y2 = h - box.top, h - box.bottom
assert x2 > x1
assert y2 > y1
char_im = im.crop((x1, y1, x2, y2))
out_dir, out_file = path_for_letter(output_dir, image_path, idx, box.letter)
out_path = os.path.join(out_dir, out_file)
mkdir_p(out_dir)
char_im.save(out_path)