Geolocalistaion des bibliothèques de Montréal

locater.py

# -*- coding: utf-8 -*-
import json
import requests
import urllib
from urllib import urlopen
import codecs
import time

url = u"http://maps.googleapis.com/maps/api/geocode/json?address={0}&sensor=false"

def format_spaces(input):
    return input.replace(" ", "+")

def get_coords(place):
    place = format_spaces(place)
    request = url.format(place)
    response = urlopen(request.encode("utf8"))
    j = json.loads(response.read())
    try:
        coords = j["results"][0]["geometry"]["location"]
    except:
        print j
    return coords["lat"], coords["lng"]

if __name__ == '__main__':
    with open("adresses_bibliotheques.txt") as f:
        file = codecs.getreader("utf-8")(f).read()
        lines = file.split('\n')
        lines.pop()

    locations = []
    for i,l in enumerate(lines):
        name, adresse = l.split(': ')
        adresse = adresse.replace('<br>', ' ')
        lat, lng = get_coords(adresse)
        out = {
                "name": name,
                "latitude": lat,
                "longitude": lng,
                "adresse": adresse
              }
        locations.append(out)
        if i%5:
            time.sleep(0.5)
        #print json.dumps(out, indent=4, sort_keys=True, ensure_ascii=False), ","
        #for key, val in out.items():
            #print key, val
    print json.dumps(locations, indent=4, sort_keys=True, ensure_ascii=False)

scrape.py

#!/usr/bin/env python

from urllib import urlopen
import json

url = "http://ville.montreal.qc.ca/pls/portal/ubst.lieu_wqp.get_coord_horaire?p_langue={0}&p_id={1}"

#individual page:
def scrape_page(id, lang="fr"):
    response = urlopen(url.format(lang, id)).read()
    if response == "{\n":
        return "","",""
    response = response.decode("iso-8859-1")
    j = json.loads(response)
    horaire = []
    #for i in range(1,8):
        #horaire.append(j.get("HREGAJ" + i))

    return j.get("adresse"), j.get("legende"), horaire

if __name__ == '__main__':
    for i in range(60):
        adresse, city, hr = scrape_page(i)
        if adresse and city:
            result = u"{}: {}".format(city, adresse)
            print result.encode("utf-8")