From 8ebb376bb8aab36e72ab9e41f6c553a1bf4823c1 Mon Sep 17 00:00:00 2001
From: Christian Berendt <berendt@b1-systems.de>
Date: Tue, 12 Jan 2016 09:12:26 +0100
Subject: [PATCH] [sitemap] set higher priority for files of the current
 release

Change-Id: I9dbaa787354582f2f766fcce58aff95766d242c7
---
 sitemap/generator/spiders/sitemap.py | 36 +++++++++++++++-------------
 1 file changed, 20 insertions(+), 16 deletions(-)

diff --git a/sitemap/generator/spiders/sitemap.py b/sitemap/generator/spiders/sitemap.py
index 44009d5..1c36405 100644
--- a/sitemap/generator/spiders/sitemap.py
+++ b/sitemap/generator/spiders/sitemap.py
@@ -10,7 +10,6 @@
 # License for the specific language governing permissions and limitations
 # under the License.
 
-import posixpath
 import time
 import urlparse
 
@@ -21,6 +20,19 @@ from scrapy import spiders
 
 class SitemapSpider(spiders.CrawlSpider):
     name = 'sitemap'
+    old_releases = tuple(["/%s" % old_release for old_release in [
+        'austin',
+        'bexar',
+        'cactus',
+        'diablo',
+        'essex',
+        'folsom',
+        'grizzly',
+        'havana',
+        'icehouse',
+        'juno',
+        'kilo'
+    ]])
 
     rules = [
         spiders.Rule(
@@ -52,25 +64,17 @@ class SitemapSpider(spiders.CrawlSpider):
 
     def parse_item(self, response):
         item = items.SitemapItem()
-        item['priority'] = '0.5'
-        item['changefreq'] = 'daily'
         item['loc'] = response.url
 
         path = urlparse.urlsplit(response.url).path
-        filename = posixpath.basename(path)
-
-        if filename == 'index.html' or filename == '':
+        if path.startswith(self.old_releases):
+            # weekly changefrequency and lower priority for old files
+            item['priority'] = '0.5'
+            item['changefreq'] = 'weekly'
+        else:
+            # daily changefrequency and highest priority for current files
             item['priority'] = '1.0'
-
-        weekly = [
-            'juno',
-            'icehouse',
-            'havana'
-        ]
-
-        for entry in weekly:
-            if path.startswith("/%s" % entry):
-                item['changefreq'] = 'weekly'
+            item['changefreq'] = 'daily'
 
         if 'Last-Modified' in response.headers:
             timestamp = response.headers['Last-Modified']