v0.10.0

CHANGES.rst

@@ -1,3 +1,7 @@
+0.10.0 (2020-05-25)
+-------------------
+- Fix spaces missing in some cases (#116, #227)
+
 0.9.2 (2020-04-26)
 -------------------
 - Fix #271

docs/index.rst

@@ -54,8 +54,25 @@ If you want to manage paragraphs, table rows and a whole run with its style, you
 
 **Note:**
 
-a 'run' for Microsoft Word is a sequence of characters with the same style. For example, if you create a paragraph with all characters of the same style, MS Word will create internally only one 'run' in the paragraph. Now, if you put in bold a text in the middle of this paragraph, word will transform the previous 'run' into 3 different 'runs' (normal - bold - normal).
+a 'run' for Microsoft Word is a sequence of characters with the same style.
+For example, if you create a paragraph with all characters of the same style,
+MS Word will create internally only one 'run' in the paragraph. Now,
+if you put in bold a text in the middle of this paragraph,
+word will transform the previous 'run' into 3 different 'runs' (normal - bold - normal).
 
+**Important:**
+
+Always put space after a jinja2 starting var/tag delimiter and a space before the ending one :
+
+Avoid::
+
+   {{myvariable}}
+   {%if something%}
+
+Use instead::
+
+   {{ myvariable }}
+   {% if something %}
 
 Extensions
 ++++++++++

docxtpl/__init__.py

@@ -7,7 +7,7 @@ Created : 2015-03-12
 import functools
 import io
 
-__version__ = '0.9.2'
+__version__ = '0.10.0'
 
 from lxml import etree
 from docx import Document
@@ -67,11 +67,17 @@ class DocxTemplate(object):
             fh.write(self.get_xml())
 
     def patch_xml(self, src_xml):
-        # strip all xml tags inside {% %} and {{ }} that MS word can insert
+        """ Make a lots of cleanning to have a raw xml understandable by jinja2 :
-        # into xml source also unescape html entities
+        strip all unnecessary xml tags, manage table cell background color and colspan,
+        unescape html entities, etc... """
+
+        # replace {<something>{ by {{   ( works with {{ }} {% and %} )
         src_xml = re.sub(r'(?<={)(<[^>]*>)+(?=[\{%])|(?<=[%\}])(<[^>]*>)+(?=\})', '',
                          src_xml, flags=re.DOTALL)
 
+        # replace {{<some tags>jinja2 stuff<some other tags>}} by {{jinja2 stuff}}
+        # same thing with {% ... %}
+        # "jinja2 stuff" could a variable, a 'if' etc... anything jinja2 will understand
         def striptags(m):
             return re.sub('</w:t>.*?(<w:t>|<w:t [^>]*>)', '',
                           m.group(0), flags=re.DOTALL)
@@ -101,9 +107,8 @@ class DocxTemplate(object):
         src_xml = re.sub(r'(<w:tc[ >](?:(?!<w:tc[ >]).)*){%\s*cellbg\s+([^%]*)\s*%}(.*?</w:tc>)',
                          cellbg, src_xml, flags=re.DOTALL)
 
-        # avoid {{r and {%r tags to strip MS xml tags too far
+        # ensure space preservation
-        # ensure space preservation when splitting
+        src_xml = re.sub(r'<w:t>((?:(?!<w:t>).)*)({{.*?}}|{%.*?%})',
-        src_xml = re.sub(r'<w:t>((?:(?!<w:t>).)*)({{r\s.*?}}|{%r\s.*?%})',
                          r'<w:t xml:space="preserve">\1\2',
                          src_xml, flags=re.DOTALL)
         src_xml = re.sub(r'({{r\s.*?}}|{%r\s.*?%})',

tests/preserve_spaces.py

@@ -0,0 +1,14 @@
+from docxtpl import DocxTemplate, InlineImage
+
+# With old docxtpl version, "... for spicy ..." was replaced by "... forspicy..."
+# This test is for checking that is some cases the spaces are not lost anymore
+
+tpl = DocxTemplate('templates/preserve_spaces_tpl.docx')
+
+tags = ['tag_1', 'tag_2']
+replacement = ['looking', 'too']
+
+context = dict(zip(tags, replacement))
+
+tpl.render(context)
+tpl.save('output/preserve_spaces.docx')