qurator-spk/page2tsv
1
0
Fork 0
mirror of https://github.com/qurator-spk/page2tsv.git synced 2025-06-09 19:39:54 +02:00
Code Issues Releases Wiki Activity

cli: produce TSV if no words are transcribed

Browse source
This commit is contained in:
Konstantin Baierer 2022-02-21 17:00:03 +01:00
parent 93ee53c8e2
commit fe0c355e5a
1 changed files with 10 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

16
tsvtools/cli.py
View file

@ -101,12 +101,16 @@ def page2tsv(page_xml_file, tsv_out_file, purpose, image_url, ner_rest_endpoint,
text_equiv.get_Unicode(), len(urls), left, right, top, bottom, text_line.id)) text_equiv.get_Unicode(), len(urls), left, right, top, bottom, text_line.id))
else: else:
for word in words: for word in words:
for text_equiv in word.get_TextEquiv(): # XXX TODO make this configurable
# transform OCR coordinates using `scale_factor` to derive textequiv = ''
# correct coordinates for the web presentation image list_textequivs = word.get_TextEquiv()
left, top, right, bottom = [int(scale_factor * x) for x in bbox_from_points(word.get_Coords().points)] if list_textequivs:
tsv.append((region_idx, len(line_info) - 1, left + (right - left) / 2.0, textequiv = list_textequivs[0].get_Unicode()
text_equiv.get_Unicode(), len(urls), left, right, top, bottom, text_line.id)) # transform OCR coordinates using `scale_factor` to derive
# correct coordinates for the web presentation image
left, top, right, bottom = [int(scale_factor * x) for x in bbox_from_points(word.get_Coords().points)]
tsv.append((region_idx, len(line_info) - 1, left + (right - left) / 2.0,
textequiv, len(urls), left, right, top, bottom, text_line.id))
line_info = pd.DataFrame(line_info, columns=['url_id', 'left', 'right', 'top', 'bottom', 'conf', 'line_id']) line_info = pd.DataFrame(line_info, columns=['url_id', 'left', 'right', 'top', 'bottom', 'conf', 'line_id'])