ocr/hocr2tei
2022-02-03 10:40:50 +01:00

69 lines
2.0 KiB
Python
Executable File

#!/usr/bin/env python3.7
# coding=utf-8
''' Convert hOCR to TEI XML. '''
from argparse import ArgumentParser
from lxml import html
from xml.sax.saxutils import escape
import re
parser = ArgumentParser(description='Convert hOCR to TEI XML.')
parser.add_argument(
'-i', '--input-file',
help='Input file',
required=True
)
parser.add_argument(
'-o', '--output-file',
help='Output file',
required=True
)
args = parser.parse_args()
tei = ''
tei += '<TEI xmlns="http://www.tei-c.org/ns/1.0">\n'
tei += ' <teiHeader>\n'
tei += ' <fileDesc>\n'
tei += ' <titleStmt>\n'
tei += ' <title></title>\n'
tei += ' </titleStmt>\n'
tei += ' <publicationStmt>\n'
tei += ' <p></p>\n'
tei += ' </publicationStmt>\n'
tei += ' <sourceDesc>\n'
tei += ' <p></p>\n'
tei += ' </sourceDesc>\n'
tei += ' </fileDesc>\n'
tei += ' </teiHeader>\n'
tei += ' <text>\n'
tei += ' <body>\n'
hocr = html.parse(args.input_file)
for ocr_page in hocr.findall('.//div[@class="ocr_page"]'):
ocr_page_title_attrib = ocr_page.attrib.get('title')
facsimile = re.search(r'image \"(.*?)\"', ocr_page_title_attrib).group(1)
page_number = re.search(r'ppageno (\d+)', ocr_page_title_attrib).group(1)
tei += f' <pb facs="{facsimile}" n="{page_number}"/>\n'
for ocr_par in ocr_page.findall('.//p[@class="ocr_par"]'):
tei += ' <p>\n'
for ocr_line in ocr_par.findall('.//span[@class="ocr_line"]'):
tei += ' <lb/>'
is_first_word_in_line = True
for ocrx_word in ocr_line.findall('.//span[@class="ocrx_word"]'):
if ocrx_word.text is not None:
if not is_first_word_in_line:
tei += ' '
tei += escape(ocrx_word.text)
is_first_word_in_line = False
tei += '\n'
tei += ' </p>\n'
tei += ' </body>\n'
tei += ' </text>\n'
tei += '</TEI>\n'
with open(args.output_file, 'w') as f:
f.write(tei)