qurator-spk/page2tsv
1
0
Fork 0
mirror of https://github.com/qurator-spk/page2tsv.git synced 2025-06-30 21:59:54 +02:00
Code Issues Releases Wiki Activity

support segmentation only Page-XML

Browse source
This commit is contained in:
Kai 2021-11-19 11:27:46 +01:00
parent ee5f03ce07
commit ed90193c45
1 changed files with 16 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

19
tsvtools/cli.py
View file

@ -116,14 +116,27 @@ def page2tsv(page_xml_file, tsv_out_file, purpose, image_url, ner_rest_endpoint,
line_info.append((len(urls), left, right, top, bottom, conf, text_line.id))
for word in text_line.get_Word():
for text_equiv in word.get_TextEquiv():
words = [word for word in text_line.get_Word()]
if len(words) <= 0:
for text_equiv in text_line.get_TextEquiv():
# 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)]
left, top, right, bottom = [int(scale_factor * x) for x in
bbox_from_points(text_line.get_Coords().points)]
tsv.append((region_idx, len(line_info) - 1, left + (right - left) / 2.0,
text_equiv.get_Unicode(), len(urls), left, right, top, bottom, text_line.id))
else:
for word in words:
for text_equiv in word.get_TextEquiv():
# 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,
text_equiv.get_Unicode(), 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'])