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

try other coordinate computation

Browse source
This commit is contained in:
Kai 2021-05-07 10:50:09 +02:00
parent f320904503
commit e5b635ec2d
1 changed files with 8 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

12
tsvtools/cli.py
View file

@ -107,8 +107,10 @@ def page2tsv(page_xml_file, tsv_out_file, purpose, image_url, ner_rest_endpoint,
for region_idx, region in enumerate(pcgts.get_Page().get_AllRegions(classes=['Text'], order='reading-order')): for region_idx, region in enumerate(pcgts.get_Page().get_AllRegions(classes=['Text'], order='reading-order')):
for text_line in region.get_TextLine(): for text_line in region.get_TextLine():
points = [int(scale_factor * float(pos)) for coords in text_line.get_Coords() for p in # points = [int(scale_factor * float(pos)) for coords in text_line.get_Coords() for p in
coords.attrib['points'].split(' ') for pos in p.split(',')] # coords.attrib['points'].split(' ') for pos in p.split(',')]
points = [int(scale_factor * float(pos)) for pos in text_line.get_Coords().points]
x_points, y_points = points[0::2], points[1::2] x_points, y_points = points[0::2], points[1::2]
@ -128,8 +130,10 @@ def page2tsv(page_xml_file, tsv_out_file, purpose, image_url, ner_rest_endpoint,
# transform OCR coordinates using `scale_factor` to derive # transform OCR coordinates using `scale_factor` to derive
# correct coordinates for the web presentation image # correct coordinates for the web presentation image
points = [int(scale_factor * float(pos)) for coords in text_line.get_Coords() for p in # points = [int(scale_factor * float(pos)) for coords in text_equiv.get_Coords() for p in
coords.attrib['points'].split(' ') for pos in p.split(',')] # coords.attrib['points'].split(' ') for pos in p.split(',')]
points = [int(scale_factor * float(pos)) for pos in text_equiv.get_Coords().points]
x_points, y_points = points[0::2], points[1::2] x_points, y_points = points[0::2], points[1::2]