def _convert_page(layout, path):
text_content = []
for text_obj in _find_objects(layout._objs, (LTTextBox, LTTextLine)):
text_content.append(text_obj.get_text())
text = text_fragments(text_content)
# if len(text) < 2:
# if len(list(_find_objects(layout._objs, LTImage))):
# log.debug("Defaulting to OCR: %r, pg. %s", path, page_no)
# text = _extract_image_page(path, page_no, languages)
return text
def _convert_page(layout, path):
text_content = []
for text_obj in _find_objects(layout._objs, (LTTextBox, LTTextLine)):
text_content.append(text_obj.get_text())
text = text_fragments(text_content)
# if len(text) < 2:
# if len(list(_find_objects(layout._objs, LTImage))):
# log.debug("Defaulting to OCR: %r, pg. %s", path, page_no)
# text = _extract_image_page(path, page_no, languages)
return text
def _convert_page(layout, languages):
text_content = []
for text_obj in _find_objects(layout._objs, (LTTextBox, LTTextLine)):
text_content.append(text_obj.get_text())
text = text_fragments(text_content)
if len(text) > 3:
# TODO: invent a smarter way to decide whether to do OCR.
return text
for img_obj in _find_objects(layout._objs, LTImage):
try:
if img_obj.width < OCR_MIN_WIDTH or \
img_obj.height < OCR_MIN_HEIGHT:
continue
data = img_obj.stream.get_rawdata()
img_text = extract_image_data(data, languages=languages)
text_content.append(img_text)
except Exception as ex:
log.debug(ex)
return text_fragments(text_content)