[yahoo] search implemented

weboob/backends/yahoo/backend.py

@@ -22,8 +22,9 @@ from xml.dom import minidom
 # TODO store datetime objects instead of strings
 # from datetime import datetime
 
-from weboob.capabilities.weather import ICapWeather, CityNotFound, Current, Forecast
+from weboob.capabilities.weather import ICapWeather, CityNotFound, Current, Forecast, City
 from weboob.tools.backend import BaseBackend
+from weboob.tools.browser import BaseBrowser
 
 
 __all__ = ['YahooBackend']
@@ -36,10 +37,51 @@ class YahooBackend(BaseBackend, ICapWeather):
     VERSION = '0.5'
     DESCRIPTION = 'Yahoo'
     LICENSE = 'GPLv3'
+    BROWSER = BaseBrowser
     WEATHER_URL = 'http://weather.yahooapis.com/forecastrss?w=%s&u=%s'
+    SEARCH_URL = 'http://fr.meteo.yahoo.com/search/weather?p=%s'
+
+    def create_default_browser(self):
+        return self.create_browser()
 
     def iter_city_search(self, pattern):
-        raise NotImplementedError()
+        # minidom doesn't seem to work with that page
+
+        #handler = urllib2.urlopen((self.SEARCH_URL % pattern).replace(' ','+'))
+        #dom = minidom.parse(handler)
+        #handler.close()
+        #results = dom.getElementById('search-results')
+        #for no in results.childNodes:
+        #    print no.nodeValue
+
+        # so i use a basic but efficient parsing
+        with self.browser:
+            content = self.browser.readurl((self.SEARCH_URL % pattern).replace(' ','+'))
+
+            page=''
+            for line in content.split('\n'):
+                if "<title>" in line and "Prévisions et Temps" in line:
+                    page="direct"
+                elif "<title>" in line and "Résultats de la recherche" in line:
+                    page="resultats"
+                    
+                if page == "resultats":
+                    if '/redirwoei/' in line:
+                        cities = line.split('/redirwoei/')
+                        for c in cities:
+                            if "strong" in c:
+                                cid = c.split("'")[0]
+                                cname = c.split("'")[1].replace("><strong>","").replace("</strong>","").split("</a>")[0]
+                                yield City(cid, cname.decode('utf-8'))
+                elif page == "direct":
+                    if 'div id="yw-breadcrumb"' in line:
+                        l = line.split('</a>')
+                        region = l[2].split('>')[-1]
+                        country = l[1].split('>')[-1]
+                        city = l[3].split('</li>')[1].replace('<li>','')
+                        cid = line.split("/?unit")[0].split('-')[-1]
+                        yield City(cid, (city+", "+region+", "+country).decode('utf-8'))
+
 
     def _get_weather_dom(self, city_id):
         handler = urllib2.urlopen(self.WEATHER_URL % (city_id, 'c'))