sporada/ocr
1
0
Fork 0
mirror of https://gitlab.ub.uni-bielefeld.de/sfb1288inf/ocr.git synced 2025-07-01 19:00:34 +00:00
Code Issues Projects Releases Wiki Activity

Codestyle

Browse Source
This commit is contained in:
Patrick Jentsch
2019-05-20 11:10:40 +02:00
parent 93de923b4e
commit e1462152fe
2 changed files with 16 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
hocrtotei
View File

@ -5,7 +5,9 @@ from xml.sax.saxutils import escape
import argparse
import xml.etree.ElementTree as ET
parser = argparse.ArgumentParser(description='hocrtotei merges several hOCR files in order of their occurrence on command line to one TEI result file.')
parser = argparse.ArgumentParser(
description='Merges several hOCR files in order of their occurrence on command line to one TEI result file.'
)
parser.add_argument(
'i',
metavar='hOCR-sourcefile',
@ -17,7 +19,7 @@ parser.add_argument(
)
args = parser.parse_args()
output_file = open(args.o, "w")
output_file = open(args.o, 'w')
output_file.write(
'<?xml version="1.0" encoding="UTF-8"?>\n'
@ -37,11 +39,11 @@ output_file.write(
for index, input_file in enumerate(args.i):
tree = ET.parse(input_file)
output_file.write(' <pb n="%i"/>\n' % (index + 1))
for para in tree.findall(".//*[@class='ocr_par']"):
for para in tree.findall('.//*[@class="ocr_par"]'):
output_file.write(' <p>\n')
for line in para.findall(".//*[@class='ocr_line']"):
for line in para.findall('.//*[@class="ocr_line"]'):
first_word_in_line = True
for word in line.findall(".//*[@class='ocrx_word']"):
for word in line.findall('.//*[@class="ocrx_word"]'):
if word.text is not None:
output_file.write((' ' if first_word_in_line else ' ') + escape(word.text.strip()))
first_word_in_line = False