spaces + misc pages

siemens · Sep 25, 2019 · b64e6d2 · b64e6d2
1 parent 025776e
commit b64e6d2
Showing 1 changed file with 6 additions and 1 deletion.
diff --git a/confluence_dumper.py b/confluence_dumper.py
@@ -133,7 +133,12 @@ def handle_html_references(html_content, page_duplicate_file_names, page_file_ma
     xpath_expr = '//a[contains(@href, "/display/")]'
     for link_element in html_tree.xpath(xpath_expr):
         if not link_element.get('class'):
-            page_title = link_element.attrib['href'].split('/')[4]
+            print("LINK - "+link_element.attrib['href'])
+            try:
+                page_title = link_element.attrib['href'].split('/')[4]
+            except:
+                page_title = link_element.attrib['href'].split('/')[3]
+
             page_title = page_title.replace('+', ' ')
             decoded_page_title = utils.decode_url(page_title)
             offline_link = provide_unique_file_name(page_duplicate_file_names, page_file_matching, decoded_page_title,