hocr-split

#!/usr/bin/env python

# split an hOCR file into individual pages

from __future__ import print_function
import argparse
import re

from lxml import etree, html

################################################################
# main program
################################################################

parser = argparse.ArgumentParser(
    description="split a multipage hOCR file into single pages")
parser.add_argument("file", help="hOCR file", type=argparse.FileType('r'))
parser.add_argument("pattern", help="naming pattern, e.g. 'base-%%03d.html'")
parser.add_argument("--start-from-zero", action="store_true",
                    help="start number of page from zero, e.g. 'base-000.html'")
args = parser.parse_args()

assert re.search('%[0-9]*d', args.pattern)

doc = etree.parse(args.file, html.XHTMLParser())
pages = doc.xpath("//*[@class='ocr_page']")
assert pages != []

container = pages[0].getparent()
index = 1

if args.start_from_zero:
    index = 0

for new_page in pages:
    container_pages = container.xpath("//*[@class='ocr_page']")
    for page in container_pages:
        container.remove(page)
    container.append(new_page)
    doc.write((args.pattern % index), pretty_print=True)
    index += 1