#! /usr/bin/python3 -sP

from __future__ import print_function
import sys
import re
import argparse
from lxml import html

parser = argparse.ArgumentParser(
    description='Calculate word frequency in an hOCR file')
parser.add_argument(
    '-i',
    '--case-insensitive',
    action='store_true',
    default=False,
    help="ignore case")
parser.add_argument(
    '-s',
    '--spaces',
    action='store_true',
    default=False,
    help="split on spaces only")
parser.add_argument(
    '-y',
    '--dehyphenate',
    action='store_true',
    default=False,
    help="try to dehyphenate the text before analysis")
parser.add_argument(
    '-n',
    '--max',
    type=int,
    default=10,
    help="number of hits (default: %(default)s)")
parser.add_argument(
    'hocr_in',
    help="hOCR file to count frequency for (default: standard input)",
    type=argparse.FileType('r'),
    nargs='?',
    default=sys.stdin)
args = parser.parse_args()

doc = html.parse(args.hocr_in)
text = doc.find('//body').text_content().strip()
if args.case_insensitive:
    text = text.lower()
if args.dehyphenate:
    # delete blank lines
    text = re.sub(r"^\s*$\r?\n", "", text)
    # dehyphenate
    text = re.sub(r"-\r?\n", "", text)
    # replace line breaks with a space
    text = re.sub(r"\r?\n", " ", text)
wc = {}
separators = re.compile(r'\W+', re.UNICODE)
if args.spaces:
    separators = re.compile(r'\s+', re.UNICODE)
for word in separators.split(text):
    if word == '':
        continue
    wc[word] = wc[word] + 1 if word in wc else 1

for idx, word in enumerate(sorted(wc, reverse=True, key=wc.get)):
    if idx > args.max:
        break
    print("%-5d\t%s" % (wc[word], word))
