From 80c94ccf4f001f91bdef4ee4b9efed954bd2be85 Mon Sep 17 00:00:00 2001
From: Jonathan Golder <jonathan@golderweb.de>
Date: Sat, 11 Mar 2017 11:30:19 +0100
Subject: [PATCH 1/2] Replace underscores in article titles

Remove underscores in article titles and replace with spaces to have
canonical state for all articles
Therefore we need to split title and posible anchors in heading parser

Related Task: [https://fs.golderweb.de/index.php?do=details&task_id=114 FS#114]
---
 lib/redfam.py | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)

diff --git a/lib/redfam.py b/lib/redfam.py
index dc1535d..6abb7ae 100644
--- a/lib/redfam.py
+++ b/lib/redfam.py
@@ -280,8 +280,21 @@ class RedFamParser( RedFam ):
         # (Task FS#77)
         heading = mwparser.parse( str( heading ) )
 
-        # Save destinations of wikilinks in headings
-        return [ str( link.title ) for link in heading.ifilter_wikilinks() ]
+        articlesList = []
+        for link in heading.ifilter_wikilinks():
+            article = str( link.title )
+
+            # Split in title and anchor part
+            article = article.split("#", 1)
+            # Replace underscores in title with spaces
+            article[0] = article[0].replace("_", " ")
+            # Rejoin title and anchor
+            article = "#".join(article)
+
+            # Add to list
+            articlesList.append(article)
+
+        return articlesList
 
     def add_beginning( self, beginning ):
         """

From 0f930082b4e4a77731557691d7ee4d727a781750 Mon Sep 17 00:00:00 2001
From: Jonathan Golder <jonathan@golderweb.de>
Date: Sat, 11 Mar 2017 11:40:41 +0100
Subject: [PATCH 2/2] Also canonicalise anchor parts of articles

Replace spaces in anchors with underscores as spaces are not correct
there

Related Task: [https://fs.golderweb.de/index.php?do=details&task_id=114 FS#114]
---
 lib/redfam.py | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/lib/redfam.py b/lib/redfam.py
index 6abb7ae..6c045c6 100644
--- a/lib/redfam.py
+++ b/lib/redfam.py
@@ -288,6 +288,11 @@ class RedFamParser( RedFam ):
             article = article.split("#", 1)
             # Replace underscores in title with spaces
             article[0] = article[0].replace("_", " ")
+
+            if len(article) > 1:
+                # other way round, replace spaces with underscores in anchors
+                article[1] = article[1].replace(" ", "_")
+
             # Rejoin title and anchor
             article = "#".join(article)