Ticket #8995: sitemaps.diff

django/contrib/sitemaps/__init__.py

@@ -1 @@
-from django.core import urlresolvers, paginator
+import logging
 import urllib
 
-PING_URL = "http://www.google.com/webmasters/tools/ping"
+from django.conf import settings
+from django.contrib.sites.models import Site
+from django.core import urlresolvers, paginator
 
 class SitemapNotFound(Exception):
     pass
 
-def ping_google(sitemap_url=None, ping_url=PING_URL):
+class Pinger(object):
     """
-    Alerts Google that the sitemap for the current site has been updated.
-    If sitemap_url is provided, it should be an absolute path to the sitemap
-    for this site -- e.g., '/sitemap.xml'. If sitemap_url is not provided, this
-    function will attempt to deduce it by using urlresolvers.reverse().
+
+    Superclass for alerting search engines that the sitemap for the current
+    site has been updated.  If sitemap_url is provided to the constructor, it
+    should be an absolute path to the sitemap for this site -- e.g.,
+    '/sitemap.xml'. If sitemap_url is not provided, this function will attempt
+    to deduce it by using urlresolvers.reverse().
+
+    The ping_url is specific to each search engine, thus each subclass, but
+    can be overridden in the constructor arguments.
+
     """
-    if sitemap_url is None:
-        try:
-            # First, try to get the "index" sitemap URL.
-            sitemap_url = urlresolvers.reverse('django.contrib.sitemaps.views.index')
-        except urlresolvers.NoReverseMatch:
+    logger = None
+    name = None
+    ping_url = None
+    sitemap_url = None
+    
+    def __init__(self, sitemap_url=None, ping_url=None):
+        if sitemap_url is None:
             try:
-                # Next, try for the "global" sitemap URL.
-                sitemap_url = urlresolvers.reverse('django.contrib.sitemaps.views.sitemap')
+                # First, try to get the "index" sitemap URL.
+                sitemap_url = urlresolvers.reverse('django.contrib.sitemaps.views.index')
             except urlresolvers.NoReverseMatch:
-                pass
+                try:
+                    # Next, try for the "global" sitemap URL.
+                    sitemap_url = urlresolvers.reverse('django.contrib.sitemaps.views.sitemap')
+                except urlresolvers.NoReverseMatch:
+                    raise SitemapNotFound("You didn't provide a sitemap_url, and the sitemap URL couldn't be auto-detected.")
 
-    if sitemap_url is None:
-        raise SitemapNotFound("You didn't provide a sitemap_url, and the sitemap URL couldn't be auto-detected.")
+        self.sitemap_url = "http://%s%s" % (Site.objects.get_current().domain, sitemap_url or get_sitemap_url())
+        if ping_url:
+            self.ping_url = ping_url
+        self.logger = logging.getLogger("django.contrib.sitemaps.%s" % str(self.__class__.__name__))
+    
+    def ping(self):
+        try:
+            if settings.DEBUG:
+                self.logger.debug("Pinging %s with sitemap %s..." % (self.name, self.sitemap_url))
+            params = urllib.urlencode({'sitemap' : self.sitemap_url})
+            u = urllib.urlopen("%s?%s" % (self.ping_url, params))
+            if settings.DEBUG:
+                self.logger.debug(u.read())
+            self.logger.info("Pinged %s with sitemap %s." % (self.name, self.sitemap_url))
+        except Exception, e:
+            self.logger.error("%s ping failed: %s" % (self.name, e))
 
-    from django.contrib.sites.models import Site
-    current_site = Site.objects.get_current()
-    url = "http://%s%s" % (current_site.domain, sitemap_url)
-    params = urllib.urlencode({'sitemap':url})
-    urllib.urlopen("%s?%s" % (ping_url, params))
+class AskPinger(Pinger):
+    name = 'Ask'
+    ping_url = 'http://submissions.ask.com/ping'
 
+class GooglePinger(Pinger):
+    name = 'Google'
+    ping_url = 'http://www.google.com/webmasters/tools/ping'
+
+class LiveSearchPinger(Pinger):
+    name = 'Live Search'
+    ping_url = 'http://webmaster.live.com/ping.aspx'
+
+def ping_google(sitemap_url=None, ping_url=None):
+    GooglePinger(sitemap_url=sitemap_url).ping()
+
+def ping_search_engines(sitemap_url=None):
+    pingers = getattr(settings, 'SITEMAP_PINGERS', [AskPinger, GooglePinger, LiveSearchPinger])
+    for pinger in pingers:
+        pinger = pinger(sitemap_url=sitemap_url)
+        pinger.ping()
+
 class Sitemap(object):
     # This limit is defined by Google. See the index documentation at
     # http://sitemaps.org/protocol.php#index.
     limit = 50000
+    protocol = 'http'
 
     def __get(self, name, obj, default=None):
         try:
@@ -64 +108 @@
         current_site = Site.objects.get_current()
         urls = []
         for item in self.paginator.page(page).object_list:
-            loc = "http://%s%s" % (current_site.domain, self.__get('location', item))
+            loc = "%s://%s%s" % (self.protocol, current_site.domain, self.__get('location', item))
             url_info = {
                 'location':   loc,
                 'lastmod':    self.__get('lastmod', item, None),

django/contrib/sitemaps/management/commands/ping_search_engines.py

@@ +1 @@
+import logging
+import sys
+
+from django.conf import settings
+from django.core.management.base import BaseCommand
+from django.contrib.sitemaps import ping_search_engines
+
+
+class Command(BaseCommand):
+    help = "Ping all supported search engines with an updated sitemap and optionally, the url of the sitemap."
+
+    def execute(self, *args, **options):
+        logging.root.addHandler(logging.StreamHandler(sys.stderr))
+        logging.root.setLevel(settings.DEBUG and logging.DEBUG or logging.INFO)
+        if len(args) == 1:
+            sitemap_url = args[0]
+        else:
+            sitemap_url = None
+        ping_search_engines(sitemap_url=sitemap_url)
+

docs/ref/contrib/sitemaps.txt

@@ -209 +209 @@
 
         .. _sitemaps.org documentation: http://www.sitemaps.org/protocol.html#prioritydef
 
+    .. attribute:: Sitemap.protocol
+
+        **Optional.** A string containing the protocol to be used in the URL of each
+        item in the sitemap. The default is 'http'; if your site is all HTTPS, set it to 'https'.
+
 Shortcuts
 =========
 
@@ -348 +353 @@
 ping the Google server's through the command line manage.py interface::
 
     python manage.py ping_google [/sitemap.xml]
+
+New in Django development version
+---------------------------------
+
+Support for pinging other search engines is now available. There is a
+new function, :func:`ping_search_engines`, that can be used just like 
+:func:`ping_google`. It notifies Ask and Live Search in addition to 
+Google. There is a corresponding manage.py command, which again 
+works just like ping_google:
+
+    python manage.py ping_search_engines [/sitemap.xml]