From 03f7d77da6ce807effa5494f161aab07fdd2498c Mon Sep 17 00:00:00 2001 From: Michael Williamson Date: Sat, 26 Dec 2015 11:01:22 +0000 Subject: Always use mc:Fallback when reading mc:AlternateContent elements --- mammoth/docx/office_xml.py | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) (limited to 'mammoth/docx') diff --git a/mammoth/docx/office_xml.py b/mammoth/docx/office_xml.py index 4dffb44..0a6a8a8 100644 --- a/mammoth/docx/office_xml.py +++ b/mammoth/docx/office_xml.py @@ -1,4 +1,5 @@ -from .xmlparser import parse_xml +from ..lists import flat_map +from .xmlparser import parse_xml, XmlElement _namespaces = [ @@ -16,5 +17,15 @@ _namespaces = [ def read(fileobj): - return parse_xml(fileobj, _namespaces) + return _collapse_alternate_content(parse_xml(fileobj, _namespaces))[0] + +def _collapse_alternate_content(node): + if isinstance(node, XmlElement): + if node.name == "mc:AlternateContent": + return node.find_child("mc:Fallback").children + else: + node.children = flat_map(_collapse_alternate_content, node.children) + return [node] + else: + return [node] -- cgit v1.2.3