better xml code cleaning around Jinja2 tags

CHANGES.rst

@@ -1,3 +1,7 @@
+0.1.8 (2015-11-05)
+------------------
+- better xml code cleaning around Jinja2 tags
+
 0.1.7 (2015-09-09)
 ------------------
 - python 3 support

docs/index.rst

@@ -38,6 +38,8 @@ You save the document as a .docx file (xml format) : it will be your .docx templ
 
 Now you can use python-docx-template to generate as many word documents you want from this .docx template and context variables you will associate.
 
+Note : python-docx-template as been tested with MS Word 97, it may not work with other version.
+
 Jinja2-like syntax
 ------------------
 

docxtpl/__init__.py

@@ -26,18 +26,16 @@ class DocxTemplate(object):
         return self.docx
 
     def get_xml(self):
-        return etree.tostring(self.docx._element.body, encoding='unicode', pretty_print=True)
+        # Be careful : pretty_print MUST be set to False, otherwise patch_xml() won't work properly
+        return etree.tostring(self.docx._element.body, encoding='unicode', pretty_print=False)
 
     def write_xml(self,filename):
         with open(filename,'w') as fh:
             fh.write(self.get_xml())
 
     def patch_xml(self,src_xml):
-        # strip all xml tags inside {% %} and {{ }}
+        # strip all xml tags inside {% %} and {{ }} that MS word can insert into xml source
-        # that Microsoft word can insert into xml code for this part of the document
+        src_xml = re.sub(r'(?<={)(<[^>]*>)+(?=[\{%])|(?<=[%\}])(<[^>]*>)+(?=\})','',src_xml,flags=re.DOTALL)
-        
-        # A essayer : src_xml = re.sub(r'(?<={)(<[^>]*>)+(?=[\{%])|(?<=[%\}])(<[^>]*>)+(?=\})','',src_xml,flags=re.DOTALL)
-        
         def striptags(m):
             return re.sub('</w:t>.*?(<w:t>|<w:t [^>]*>)','',m.group(0),flags=re.DOTALL)
         src_xml = re.sub(r'{%(?:(?!%}).)*|{{(?:(?!}}).)*',striptags,src_xml,flags=re.DOTALL)

setup.py

@@ -23,8 +23,16 @@ News
 %(CHANGES)s
 """ % read('README', 'CHANGES')
 
+def get_version(pkg):
+    path = os.path.join(os.path.dirname(__file__),pkg,'__init__.py')
+    with open(path) as fh:
+        m = re.search(r'^__version__\s*=\s*[\'"]([^\'"]+)[\'"]',fh.read(),re.M)
+    if m:
+        return m.group(1)
+    raise RuntimeError("Unable to find __version__ string in %s." % path)
+
 setup(name='docxtpl',
-      version='0.1.7',
+      version=get_version('docxtpl'),
       description='Python docx template engine',
       long_description=long_description,
       classifiers=[