diff --git a/scripts/common.py b/scripts/common.py index 661ce1ddb8b0a56f8ffbcf507a64f173225d9936..e9d318624ee0f4281aad628029e5a45ec4de8ab8 100644 --- a/scripts/common.py +++ b/scripts/common.py @@ -619,6 +619,7 @@ def is_multicolumn(filename, algorithm='computational_geometry'): def _get_number_of_columns_computational_geometry(xml_document, num_beams=5): page, = xml_document.xpath('//div[@class="ocr_page"]') page_match = re.match(BBOX_REGEX, page.attrib['title']) + assert page_match is not None page_height = float(page_match.group('y1')) - float(page_match.group('y0')) from shapely.geometry import Polygon, LineString @@ -626,6 +627,7 @@ def _get_number_of_columns_computational_geometry(xml_document, num_beams=5): paragraphs = [] for paragraph_number, paragraph in enumerate(page.xpath('//p[@class="ocr_par" and @title]')): paragraph_match = re.match(BBOX_REGEX, paragraph.attrib['title']) + assert paragraph_match is not None coords = [ (float(paragraph_match.group('x0')), float(paragraph_match.group('y0'))), (float(paragraph_match.group('x0')), float(paragraph_match.group('y1'))), diff --git a/scripts/evaluate_speed_piechart.py b/scripts/evaluate_speed_piechart.py index 2e23498eb5a411843cfa96c4b36a9f89776b7a72..20846425bb9f122e54c80c5d66487d392dd1d6e1 100644 --- a/scripts/evaluate_speed_piechart.py +++ b/scripts/evaluate_speed_piechart.py @@ -39,6 +39,7 @@ def evaluate(): if exit_code != 0: continue match = re.fullmatch(COMMAND_REGEX, command) + assert match is not None command = list(dropwhile(lambda x: x is None, match.groups()))[0] if command not in LABEL_MAP: if command not in missing_commands: