add support for lfooter (updated) and top level pages...

(other than index.html).

1 files changed, 39 insertions, 17 deletions

diff --git a/pyblog b/pyblog
index 411b261d..fe3b1f16 100755
--- a/pyblog
+++ b/pyblog
@@ -259,33 +259,39 @@ def generate_index(feed):
     os.remove(tmppath)
 
 
-def make_sitemap_url_element(atomlink, atomupdated, changefreq, priority):
+def make_sitemap_url_element(link, updated=None, changefreq=None, priority=None):
     """Make a sitemap <url> element.
 
     Parameters
     ----------
-    atomlink : xml.etree.ElementTree.Element
-        atom:link element, e.g., <link href="http://zmwangx.github.io/"/>
-    atomupdated : xml.etree.ElementTree.Element
-        atom:updated element, e.g., <updated>2015-05-05T22:38:42-07:00</updated>
-    changefreq : {"always", "hourly", "daily", "weekly", "monthly", "yearly", "never"}
-    priority : {1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1}
+    link : str or xml.etree.ElementTree.Element
+        If using an xml.etree.ElementTree.Element element, then it shall
+        be an atom:link element, e.g., <link href="http://zmwangx.github.io/"/>.
+    updated : str or xml.etree.ElementTree.Element, optional
+        If using an xml.etree.ElementTree.Element element, then it shall
+        be an atom:updated element, e.g.,
+        <updated>2015-05-05T22:38:42-07:00</updated>.
+    changefreq : {"always", "hourly", "daily", "weekly", "monthly", "yearly", "never"}, optional
+    priority : {1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1}, optional
 
     """
 
     urlelem = ET.Element("url")
     loc = ET.Element("loc")
-    loc.text = atomlink.attrib["href"]
-    lastmod = ET.Element("lastmod")
-    lastmod.text = atomupdated.text
-    changefreq_elem = ET.Element("changefreq")
-    changefreq_elem.text = changefreq
-    priority_elem = ET.Element("priority")
-    priority_elem.text = "%.1f" % priority
+    loc.text = link.attrib["href"] if isinstance(link, ET.Element) else link
     urlelem.append(loc)
-    urlelem.append(lastmod)
-    urlelem.append(changefreq_elem)
-    urlelem.append(priority_elem)
+    if updated is not None:
+        lastmod = ET.Element("lastmod")
+        lastmod.text = updated.text if isinstance(updated, ET.Element) else updated
+        urlelem.append(lastmod)
+    if changefreq is not None:
+        changefreq_elem = ET.Element("changefreq")
+        changefreq_elem.text = changefreq
+        urlelem.append(changefreq_elem)
+    if priority is not None:
+        priority_elem = ET.Element("priority")
+        priority_elem.text = "%.1f" % priority
+        urlelem.append(priority_elem)
     return urlelem
 
 
@@ -294,6 +300,22 @@ def generate_sitemap(feed):
     sitemap = ET.Element("urlset", xmlns="http://www.sitemaps.org/schemas/sitemap/0.9")
     # index
     sitemap.append(make_sitemap_url_element(BLOG_HOME, feed.updated, "daily", 1.0))
+    # other top level pages
+    for name in os.listdir(BUILDDIR):
+        if not name.endswith(".html") or name == "index.html":
+            continue
+        link = "{home}/{path}".format(home=BLOG_HOME, path=name)
+        fullpath = os.path.join(BUILDDIR, name)
+        # try to extract updated time
+        updated = None
+        with open(fullpath, encoding="utf-8") as htmlobj:
+            soup = bs4.BeautifulSoup(htmlobj.read())
+            if soup.article.footer is not None:
+                updated_tag = soup.article.footer.find(attrs={"class": "updated"})
+                if updated_tag is not None:
+                    updated = updated_tag.text
+        sitemap.append(make_sitemap_url_element(link, updated, "monthly", 0.9))
+
     # blog entries
     for entry in feed.entries:
         sitemap.append(make_sitemap_url_element(entry.link, entry.updated, "monthly", 0.9))