[damned-lies] Support doap files where root is <rdf:RDF>

[Claude Paroz <claudep src gnome org>]

commit a094cacd0933a75f8246ba8b4ba0b1d279d3a6ab
Author: Claude Paroz <claude 2xlibre net>
Date:   Tue Oct 26 22:31:43 2010 +0200

    Support doap files where root is <rdf:RDF>

 stats/doap.py |   61 +++++++++++++++++++++++++++++++++-----------------------
 1 files changed, 36 insertions(+), 25 deletions(-)
---
diff --git a/stats/doap.py b/stats/doap.py
index e3f7e04..427c84a 100644
--- a/stats/doap.py
+++ b/stats/doap.py
@@ -5,29 +5,40 @@ from urllib import unquote
 
 from people.models import Person
 
-def parse_maintainers(doap_tree):
-    maint_tags = doap_tree.getroot().findall("{http://usefulinc.com/ns/doap#}maintainer";)
-    pers_attrs = [
-        "{http://xmlns.com/foaf/0.1/}name";,
-        "{http://xmlns.com/foaf/0.1/}mbox";,
-        "{http://api.gnome.org/doap-extensions#}userid";
-    ]
-    maintainers = []
-    for maint in maint_tags:
-        name, mbox, uid = [maint[0].find(attr) for attr in pers_attrs]
-        maint = {'name': name.text, 'email': None, 'account': None}
-        if mbox is not None:
-            maint['email'] = unquote(mbox.items()[0][1].replace("mailto:";, ""))
-        if uid is not None:
-            maint['account'] = uid.text
-        maintainers.append(maint)
-    return maintainers
+class DoapParser(object):
+    def __init__(self, doap_path):
+        self.tree = parse(doap_path)
+        self.project_root = self.tree.getroot()
+        if not self.project_root.tag.endswith("Project"):
+            # Try to get Project from root children (root might be <rdf:RDF>)
+            for child in self.project_root.getchildren():
+                if child.tag.endswith("Project"):
+                    self.project_root = child
+                    break
 
-def parse_homepage(doap_tree):
-    homepage_tag = doap_tree.getroot().find("{http://usefulinc.com/ns/doap#}homepage";)
-    if homepage_tag is not None:
-        return homepage_tag.attrib.get('{http://www.w3.org/1999/02/22-rdf-syntax-ns#}resource')
-    return None
+    def parse_maintainers(self):
+        maint_tags = self.project_root.findall("{http://usefulinc.com/ns/doap#}maintainer";)
+        pers_attrs = [
+            "{http://xmlns.com/foaf/0.1/}name";,
+            "{http://xmlns.com/foaf/0.1/}mbox";,
+            "{http://api.gnome.org/doap-extensions#}userid";
+        ]
+        maintainers = []
+        for maint in maint_tags:
+            name, mbox, uid = [maint[0].find(attr) for attr in pers_attrs]
+            maint = {'name': name.text, 'email': None, 'account': None}
+            if mbox is not None:
+                maint['email'] = unquote(mbox.items()[0][1].replace("mailto:";, ""))
+            if uid is not None:
+                maint['account'] = uid.text
+            maintainers.append(maint)
+        return maintainers
+
+    def parse_homepage(self):
+        homepage_tag = self.project_root.find("{http://usefulinc.com/ns/doap#}homepage";)
+        if homepage_tag is not None:
+            return homepage_tag.attrib.get('{http://www.w3.org/1999/02/22-rdf-syntax-ns#}resource')
+        return None
 
 def update_doap_infos(module):
     """ Should only be called inside an "update-stats" context of a master branch,
@@ -35,13 +46,13 @@ def update_doap_infos(module):
     doap_path = os.path.join(module.get_head_branch().co_path(), "%s.doap" % module.name)
     if not os.access(doap_path, os.F_OK):
         return
-    tree = parse(doap_path)
+    tree = DoapParser(doap_path)
 
     # *********** Update maintainers
     def slugify(val):
         value = unicode(re.sub('[^\w\s-]', '', val).strip().lower())
         return re.sub('[-\s]+', '-', value)
-    doap_maintainers = parse_maintainers(tree)
+    doap_maintainers = tree.parse_maintainers()
     current_maintainers = dict([(m.email or m.username, m) for m in module.maintainers.all()])
 
     # Using email or username as unique identifier
@@ -65,7 +76,7 @@ def update_doap_infos(module):
         module.maintainers.remove(maint)
 
     # *********** Update homepage
-    home = parse_homepage(tree)
+    home = tree.parse_homepage()
     if home and home != module.homepage:
         module.homepage = home
         module.save()