SRTM parser for Cartopy

gistfile1.py

"""Construct an SRTM lookup file for Cartopy."""
from HTMLParser import HTMLParser
import json
import os
import urllib2


#: Base URL where SRTM data files (with extensions .hgt.zip) are located.
SRTM_URL = 'http://e4ftl01.cr.usgs.gov/SRTM/SRTMGL3.003/2000.02.11/'

#: Name of JSON file to write output to.
SRTM_JSON = 'srtm.json'


class SRTMParser(HTMLParser):
    """Parser for SRTM data download page."""

    def __init__(self):
        HTMLParser.__init__(self)
        self.targets = {}

    def handle_starttag(self, tag, attrs):
        if tag == 'a':
            target = None
            for attribute, value in attrs:
                if attribute == 'href':
                    target = value
            if target is not None and target.endswith('.hgt.zip'):
                self.targets[target.split('.')[0]] = os.path.join(
                    SRTM_URL, target)


if __name__ == '__main__':
    resource = urllib2.urlopen(SRTM_URL)
    html = resource.read()
    resource.close()
    parser = SRTMParser()
    parser.feed(html)
    with open(SRTM_JSON, 'w') as f:
        json.dump(parser.targets, f)