Add sitemap index and serve sitemap for subprojects

The sitemap index lists all the sitemap locations for a project
including the subprojects.
The sitemap of subprojects are also served from the parent domain.

Closes readthedocs#6841

Author: cedk

docs/user/reference/sitemaps.rst

@@ -16,12 +16,18 @@ It contains information such as:
 * How important this URL is in relation to other URLs in the site.
 * What translations are available for a page.
 
-Read the Docs automatically generates a ``sitemap.xml`` for your project,
+Read the Docs automatically generates a ``sitemap.xml`` and a
+``sitemap_index.xml`` for your project,
 
 By default the sitemap includes:
 
 * Each version of your documentation and when it was last updated, sorted by version number.
 
+By default the sitemap index includes:
+
+* The location of ``sitemap.xml``
+* The locations of the ``sitemap.xml`` of subprojects if they are set.
+
 This allows search engines to prioritize results based on the version number,
 sorted by `semantic versioning`_.
 

readthedocs/proxito/tests/test_full.py

@@ -1504,6 +1504,51 @@ def test_sitemap_all_private_versions(self):
         )
         self.assertEqual(response.status_code, 404)
 
+    def test_sitemap_subproject(self):
+        self.project.versions.update(active=True)
+        self.subproject.versions.update(active=True)
+
+        subresponse = self.client.get(
+            reverse("sitemap_xml", args=["subproject"]),
+            headers={"host": "project.readthedocs.io"},
+        )
+        response = self.client.get(
+            reverse("sitemap_xml"), headers={"host": "subproject.readthedocs.io"}
+        )
+
+        self.assertEqual(subresponse.status_code, 200)
+        self.assertEqual(response.status_code, 200)
+        self.assertEqual(subresponse.content, response.content)
+
+    def test_sitemap_index(self):
+        self.project.versions.update(active=True)
+        response = self.client.get(
+            reverse("sitemap_index_xml"), headers={"host": "project.readthedocs.io"}
+        )
+        self.assertEqual(response.status_code, 200)
+        self.assertEqual(response["Content-Type"], "application/xml")
+        expected = dedent(
+            """
+            <?xml version="1.0" encoding="UTF-8"?>
+            <sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
+
+                <sitemap>
+                    <loc>https://project.readthedocs.io/sitemap.xml</loc>
+                </sitemap>
+
+                <sitemap>
+                    <loc>https://project.readthedocs.io/projects/subproject/sitemap.xml</loc>
+                </sitemap>
+
+                <sitemap>
+                    <loc>https://project.readthedocs.io/projects/subproject-alias/sitemap.xml</loc>
+                </sitemap>
+
+            </sitemapindex>
+            """
+        ).lstrip()
+        self.assertEqual(response.content.decode(), expected)
+
     @mock.patch(
         "readthedocs.proxito.views.mixins.staticfiles_storage",
         new=StaticFileSystemStorageTest(),

readthedocs/proxito/urls.py

@@ -47,6 +47,7 @@
     ServePageRedirect,
     ServeRobotsTXT,
     ServeSitemapXML,
+    ServeSitemapIndexXML,
     ServeStaticFiles,
 )
 from readthedocs.proxito.views.utils import fast_404, proxito_404_page_handler
@@ -136,7 +137,15 @@
         name="proxito_404_handler",
     ),
     re_path(r"robots\.txt$", ServeRobotsTXT.as_view(), name="robots_txt"),
-    re_path(r"sitemap\.xml$", ServeSitemapXML.as_view(), name="sitemap_xml"),
+    re_path(
+        r"^(?:projects/(?P<subproject_slug>{project_slug})/)?"
+        r"sitemap\.xml$".format(**pattern_opts),
+        ServeSitemapXML.as_view(),
+        name="sitemap_xml",
+    ),
+    re_path(
+        r"sitemap_index\.xml$", ServeSitemapIndexXML.as_view(), name="sitemap_index_xml"
+    ),
 ]
 
 docs_urls = [

readthedocs/proxito/views/serve.py

@@ -1019,7 +1019,7 @@ class ServeSitemapXMLBase(CDNCacheControlMixin, CDNCacheTagsMixin, View):
     # Extra cache tag to invalidate only this view if needed.
     project_cache_tag = "sitemap.xml"
 
-    def get(self, request):
+    def get(self, request, subproject_slug=None):
         """
         Generate and serve a ``sitemap.xml`` for a particular ``project``.
 
@@ -1078,6 +1078,12 @@ def changefreqs_generator():
             yield from itertools.chain(changefreqs, itertools.repeat('monthly'))
 
         project = request.unresolved_domain.project
+
+        if subproject_slug:
+            project = get_object_or_404(
+                project.subprojects, alias=subproject_slug
+            ).child
+
         public_versions = Version.internal.public(
             project=project,
             only_active=True,
@@ -1164,6 +1170,60 @@ class ServeSitemapXML(SettingsOverrideObject):
     _default_class = ServeSitemapXMLBase
 
 
+class ServeSitemapIndexXMLBase(CDNCacheControlMixin, CDNCacheTagsMixin, View):
+
+    """Serve sitemap_index.xml from the domain's root."""
+
+    cache_response = True
+    project_cache_tag = "sitemap.xml"
+
+    def get(self, request):
+        """
+        Generate and serve a ``sitemap_index.xml`` for a particular
+        ``project``.
+
+        The sitemap index is generated from the project and all sub-projects.
+        """
+
+        project = request.unresolved_domain.project
+
+        locations = [
+            "{scheme}://{domain}/sitemap.xml".format(
+                scheme="https",
+                domain=project.subdomain(),
+            )
+        ]
+        for subproject in project.related_projects.all():
+            locations.append(
+                "{scheme}://{domain}/projects/{subproject}/sitemap.xml".format(
+                    scheme="https",
+                    domain=project.subdomain(),
+                    subproject=subproject.slug,
+                )
+            )
+        context = {
+            "locations": locations,
+        }
+        return render(
+            request,
+            "sitemap_index.xml",
+            context,
+            content_type="application/xml",
+        )
+
+    def _get_project(self):
+        # Method used by the CDNCacheTagsMixin class.
+        return self.request.unresolved_domain.project
+
+    def _get_version(self):
+        # This view isn't attached to a version.
+        return None
+
+
+class ServeSitemapIndexXML(SettingsOverrideObject):
+    _default_class = ServeSitemapIndexXMLBase
+
+
 class ServeStaticFiles(CDNCacheControlMixin, CDNCacheTagsMixin, ServeDocsMixin, View):
 
     """

readthedocs/templates/sitemap_index.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
+{% for loc in locations %}
+    <sitemap>
+        <loc>{{ loc }}</loc>
+    </sitemap>
+{% endfor %}
+</sitemapindex>