wikihelp: Ignore the first level="1" heading.

It is usually (== everywhere but 2 or 3 files) the only level 1 heading in the
file, and additionaly exactly the same as the page title.  When we skip that,
we get much nicer table of content, better structure, etc.

1 files changed, 17 insertions, 1 deletions

diff --git a/helpcontent2/to-wiki/wikiconv2.py b/helpcontent2/to-wiki/wikiconv2.py
index acc2b97685..a9fd773845 100755
--- a/helpcontent2/to-wiki/wikiconv2.py
+++ b/helpcontent2/to-wiki/wikiconv2.py
@@ -281,6 +281,11 @@ class XhpFile(ElementBase):
     def __init__(self):
         ElementBase.__init__(self, None, None)
 
+        # we want to ignore the 1st level="1" heading, because in most of the
+        # cases, it is the only level="1" heading in the file, and it is the
+        # same as the page title
+        self.ignore_heading = True
+
     def start_element(self, parser, name, attrs):
         if name == 'body':
             # ignored, we flatten the structure
@@ -301,7 +306,18 @@ class XhpFile(ElementBase):
         elif name == 'meta':
             self.parse_child(Meta(attrs, self))
         elif name == 'paragraph':
-            self.parse_child(Paragraph(attrs, self))
+            ignore_this = False
+            try:
+                if attrs['role'] == 'heading' and int(attrs['level']) == 1 \
+                        and self.ignore_heading and parser.follow_embed:
+                    self.ignore_heading = False
+                    ignore_this = True
+            except:
+                pass
+            if ignore_this:
+                self.parse_child(Ignore(attrs, self, 'paragraph'))
+            else:
+                self.parse_child(Paragraph(attrs, self))
         elif name == 'section':
             self.parse_child(Section(attrs, self))
         elif name == 'sort':