fix street names in OSM Overpass XML file

osm_street_parse.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# Copyright (c) 2020 Nasridinov Renat
# MIT License

import argparse
import logging
import re 
import sys
from lxml import etree
from pathlib import Path

logging.basicConfig(
    format='%(levelname)s:%(message)s', filename='street.log',
    filemode='w',level=logging.INFO)

STREETS = {
    "^вул\.?(?:иця)?(.+?)$": "вулиця",
    "^ул\.?(?:ица)?(.+?)$": "улица",
    "^пров(?!еряем)\.?(?:улок)?(.+?)$": "провулок",
    "^провулок\s(.+?)$": "провулок",
    # "^пров\.?(?:улок)?(.+?)$": "провулок",
    "^переулок\s(.+?)$": "переулок",
    "^пер\.?(?:еул\.(?:ок)?)?(.+?)$": "переулок",
    "^проезд\s(.+?)$": "проезд",
    "^проулок\s(.+?)$": "проулок",
    "^проїзд\s(.+?)$": "проїзд",
    "^пр(?:\.\s)?(?:(?:осп(?:\.\s)?(?:ект)?)?)?\s(.+?)$": "проспект",
}

def reverse_street(value):
    _ = re.search("\(|\)", value)
    if _ is not None:
        logging.info(value)
    _ = re.search("\d+-го", value)
    if _ is not None:
        logging.info(value)
    match = [y for y in [(x, re.match(x, value, re.I)) for x in
            STREETS.keys()] if y[1] is not None]
    if not match:
        logging.error(value)
        return
    k = match[0][1].group(1).strip(), STREETS[match[0][0]]
    return(' '.join(k))


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument('xmlfile', help='file to be processed',  type=str)
    parser.add_argument('outfile', help='file to be write', type=str,
        default='osm_new.xml', nargs='?')
    args = parser.parse_args()
    osm_filepath = Path(args.xmlfile)
    if not osm_filepath.is_file():
        print("No file found.\n")
        sys.exit(1)
    try:
        osm_tree = etree.parse(args.xmlfile)
    except:
        print("Input file processing error, maybe it is not "
            "valid XML file?\n")
        sys.exit(1)
    tags = osm_tree.xpath('//tag')
    counter = 0
    for tag in tags:
        k, v = tag.attrib.values()
        if not k or k == "name:en":
            continue
        if re.match("^(?:old_)?name(?:\:(?:ru|uk))?$", k):
            if re.match("^.+? (?:вулиця|пров?улок|площа|проспект|улица|"\
                "переулок|площадь|про[еї]зд)$", v):
                continue
            new_name = reverse_street(v)
            if new_name:
                tag.attrib['v'] = new_name
                parent = tag.xpath("./parent::*")
                parent[0].attrib['action']='modify'
                counter += 1

    if counter == 0:
        print("No elements processed")
        sys.exit(0)
    else:
        print(f"{counter} elements processed")

    try:
        with open(args.outfile, 'wb') as f:
            f.write(etree.tostring(osm_tree,
                pretty_print=False,xml_declaration=True,encoding="UTF-8"))
    except Exception as e:
        print("There is an error occurred during output file "
             f"write process:\n{e.strerror}\n")
        sys.exit(1)