[meteofrance] fix #1431, call the url that retrieve all the search results

modules/meteofrance/browser.py

@@ -21,11 +21,7 @@
 import urllib
 
 from weboob.tools.browser import BaseBrowser
-from weboob.tools.json import json as simplejson
-from weboob.capabilities.weather import City
-
-from .pages.meteo import WeatherPage
-import re
+from .pages.meteo import WeatherPage, SearchCitiesPage
 
 __all__ = ['MeteofranceBrowser']
 
@@ -34,29 +30,23 @@ class MeteofranceBrowser(BaseBrowser):
     DOMAIN = 'www.meteofrance.com'
     PROTOCOL = 'http'
     ENCODING = 'utf-8'
-    USER_AGENT = BaseBrowser.USER_AGENTS['wget']
     WEATHER_URL = '{0}://{1}/previsions-meteo-france/{{city_name}}/{{city_id}}'.format(PROTOCOL, DOMAIN)
-    CITY_SEARCH_URL = '{0}://{1}/mf3-rpc-portlet/rest/lieu/facet/previsions/search/{{city_pattern}}'\
-                      .format(PROTOCOL, DOMAIN)
+    CITY_SEARCH_URL = 'http://www.meteofrance.com/recherche/resultats'
     PAGES = {
         WEATHER_URL.format(city_id=".*", city_name=".*"): WeatherPage,
+        CITY_SEARCH_URL: SearchCitiesPage,
         }
 
     def __init__(self, *args, **kwargs):
         BaseBrowser.__init__(self, *args, **kwargs)
 
     def iter_city_search(self, pattern):
-        searchurl = self.CITY_SEARCH_URL.format(city_pattern=urllib.quote_plus(pattern.encode('utf-8')))
-        response = self.openurl(searchurl)
-        return self.parse_cities_result(response)
-
-    def parse_cities_result(self, datas):
-        cities = simplejson.loads(datas.read(), self.ENCODING)
-        re_id = re.compile('\d{5}', re.DOTALL)
-        for city in cities:
-            if re_id.match(city['codePostal']):
-                mcity = City(int(city['codePostal']), u'%s' % city['slug'])
-                yield mcity
+        datas = {'facet': 'previsions',
+                 'search-type': 'previsions',
+                 'query': pattern}
+        self.location(self.CITY_SEARCH_URL, data=urllib.urlencode(datas))
+        assert self.is_on_page(SearchCitiesPage)
+        return self.page.iter_cities()
 
     def iter_forecast(self, city_id):
         mcity = self.iter_city_search(city_id).next()

modules/meteofrance/pages/meteo.py

@@ -20,14 +20,24 @@
 
 
 from weboob.tools.browser import BasePage
-from weboob.capabilities.weather import Forecast, Current
+from weboob.capabilities.weather import Forecast, Current, City
 
 import datetime
-
+import re
 
 __all__ = ['WeatherPage']
 
 
+class SearchCitiesPage(BasePage):
+    def iter_cities(self):
+        list = self.document.getroot().xpath('//ul[@class="list-style-1"]/li/a')
+        for a in list:
+            m = re.search('/.*/(.*)/(\d{5})', a.attrib.get('href'))
+            if m:
+                mcity = City(int(m.group(2)), u'%s' % m.group(1))
+                yield mcity
+
+
 class WeatherPage(BasePage):
     def get_temp_without_unit(self, temp_str):
         # It seems that the mechanize module give us some old style