Welcome to python-docx-template's documentation!

Quickstart

To install using pip:

pip install docxtpl

or using conda:

conda install docxtpl --channel conda-forge

Usage:

from docxtpl import DocxTemplate

doc = DocxTemplate("my_word_template.docx")
context = { 'company_name' : "World company" }
doc.render(context)
doc.save("generated_doc.docx")

Introduction

This package uses 2 major packages :

python-docx for reading, writing and creating sub documents
jinja2 for managing tags inserted into the template docx

python-docx-template has been created because python-docx is powerful for creating documents but not for modifying them.

The idea is to begin to create an example of the document you want to generate with microsoft word, it can be as complex as you want : pictures, index tables, footer, header, variables, anything you can do with word. Then, as you are still editing the document with microsoft word, you insert jinja2-like tags directly in the document. You save the document as a .docx file (xml format) : it will be your .docx template file.

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

As the Jinja2 package is used, one can use all jinja2 tags and filters inside the word document. Nevertheless there are some restrictions and extensions to make it work inside a word document:

Restrictions

The usual jinja2 tags, are only to be used inside the same run of a same paragraph, it can not be used across several paragraphs, table rows, runs. If you want to manage paragraphs, table rows and a whole run with its style, you must use special tag syntax as explained in next chapter.

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).

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

Split and merge text

You can merge a jinja2 tag with previous line by using {%-
You can merge a jinja2 tag with next line by using -%}

A text containing Jinja2 tags may be unreadable if too long:

My house is located {% if living_in_town %} in urban area {% else %} in countryside {% endif %} and I love it.

One can use ENTER or SHIFT+ENTER to split a text like below, then use {%- and -%} to tell docxtpl to merge the whole thing:

My house is located
{%- if living_in_town -%}
 in urban area
{%- else -%}
 in countryside
{%- endif -%}
 and I love it.

IMPORTANT : Use an unbreakable space (CTRL+SHIFT+SPACE) when a space is wanted at line beginning or ending.

IMPORTANT 2 : {%- xxx -%} tags must be alone in a line : do not add some text before or after on the same line.

Display variables

As part of jinja2, one can used double braces:

{{ <var> }}

But if <var> is a RichText object, you must specify that you are changing the actual 'run':

{{r <var> }}

Note the r right after the openning braces.

IMPORTANT : Do not use the r variable in your template because {{r}} could be interpreted as a {{r without variable specified. Nevertheless, you can use a bigger variable name starting with 'r'. For example {{render_color}} will be interpreted as {{ render_color }} not as {{r ender_color}}.

IMPORTANT : Do not use 2 times {{r in the same run. Use RichText.add() method to concatenate several strings and styles at python side and only one {{r at template side.

Cell color

There is a special case when you want to change the background color of a table cell, you must put the following tag at the very beginning of the cell:

{% cellbg <var> %}

<var> must contain the color's hexadecimal code without the hash sign

Column spanning

If you want to dynamically span a table cell over many column (this is useful when you have a table with a dynamic column count), you must put the following tag at the very beginning of the cell to span:

{% colspan <var> %}

<var> must contain an integer for the number of columns to span. See tests/test_files/dynamic_table.py for an example.

Escaping

In order to display {%, %}, {{ or }}, one can use:

{_%, %_}, {_{ or  }_}

RichText

When you use {{ <var> }} tag in your template, it will be replaced by the string contained within var variable. BUT it will keep the current style. If you want to add dynamically changeable style, you have to use both : the {{r <var> }} tag AND a RichText object within var variable. You can change color, bold, italic, size and so on, but the best way is to use Microsoft Word to define your own character style ( Home tab -> modify style -> manage style button -> New style, select ‘Character style’ in the form ), see example in tests/richtext.py Instead of using RichText(), one can use its shortcut : R()

Important : When you use {{r }} it removes the current character styling from your docx template, this means that if you do not specify a style in RichText(), the style will go back to a microsoft word default style. This will affect only character styles, not the paragraph styles (MSWord manages this 2 kind of styles).

Hyperlink with RichText

You can add an hyperlink to a text by using a Richtext with this syntax:

tpl=DocxTemplate('your_template.docx')
rt = RichText('You can add an hyperlink, here to ')
rt.add('google',url_id=tpl.build_url_id('http://google.com'))

Put rt in your context, then use {{r rt}} in your template

Inline image

You can dynamically add one or many images into your document (tested with JPEG and PNG files). just add {{ <var> }} tag in your template where <var> is an instance of doxtpl.InlineImage:

myimage = InlineImage(tpl,'test_files/python_logo.png',width=Mm(20))

You just have to specify the template object, the image file path and optionnally width and/or height. For height and width you have to use millimeters (Mm), inches (Inches) or points(Pt) class. Please see tests/inline_image.py for an example.

Sub-documents

A template variable can contain a complex and built from scratch with python-docx word document. To do so, get first a sub-document object from template object and use it as a python-docx document object, see example in tests/subdoc.py.

Escaping, newline, new paragraph, Listing

When you use a {{ <var> }}, you are modifying an XML word document, this means you cannot use all chars, especially <, > and &. In order to use them, you must escape them. There are 4 ways :

context = { 'var':R('my text') } and {{r <var> }} in the template (note the r),

context = { 'var':'my text'} and {{ <var>|e }} in your word template

context = { 'var':escape('my text')} and {{ <var> }} in the template.

enable autoescaping when calling render method: tpl.render(context, autoescape=True) (default is autoescape=False)

The RichText() or R() offers newline, new paragraph, and page break features : just use \n, \a, or \f in the text, they will be converted accordingly.

See tests/escape.py example for more informations.

Another solution, if you want to include a listing into your document, that is to escape the text and manage n, a, and f you can use the Listing class :

in your python code:

context = { 'mylisting':Listing('the listing\nwith\nsome\nlines \a and some paragraph \a and special chars : <>&') }

in your docx template just use {{ mylisting }} With Listing(), you will keep the current character styling (except after a \a as you start a new paragraph).

Replace docx pictures

It is not possible to dynamically add images in header/footer, but you can change them. The idea is to put a dummy picture in your template, render the template as usual, then replace the dummy picture with another one. You can do that for all medias at the same time. Note: the aspect ratio will be the same as the replaced image Note2 : Specify the filename that has been used to insert the image in the docx template (only its basename, not the full path)

Syntax to replace dummy_header_pic.jpg:

tpl.replace_pic('dummy_header_pic.jpg','header_pic_i_want.jpg')

The replacement occurs in headers, footers and the whole document's body.

Replace docx medias

It is not possible to dynamically add other medias than images in header/footer, but you can change them. The idea is to put a dummy media in your template, render the template as usual, then replace the dummy media with another one. You can do that for all medias at the same time. Note: for images, the aspect ratio will be the same as the replaced image Note2 : it is important to have the source media files as they are required to calculate their CRC to find them in the docx. (dummy file name is not important)

Syntax to replace dummy_header_pic.jpg:

tpl.replace_media('dummy_header_pic.jpg','header_pic_i_want.jpg')

WARNING : unlike replace_pic() method, dummy_header_pic.jpg MUST exist in the template directory when rendering and saving the generated docx. It must be the same file as the one inserted manually in the docx template. The replacement occurs in headers, footers and the whole document's body.

Replace embedded objects

It works like medias replacement, except it is for embedded objects like embedded docx.

Syntax to replace embedded_dummy.docx:

tpl.replace_embedded('embdded_dummy.docx','embdded_docx_i_want.docx')

WARNING : unlike replace_pic() method, embdded_dummy.docx MUST exist in the template directory when rendering and saving the generated docx. It must be the same file as the one inserted manually in the docx template. The replacement occurs in headers, footers and the whole document's body.

Note that replace_embedded() may not work on other documents than embedded docx. Instead, you should use zipname replacement:

tpl.replace_zipname(
    'word/embeddings/Feuille_Microsoft_Office_Excel1.xlsx',
    'my_excel_file.xlsx')

The zipname is the one you can find when you open docx with WinZip, 7zip (Windows) or unzip -l (Linux). The zipname starts with "word/embeddings/". Note that the file to be replaced is renamed by MSWord, so you have to guess a little bit...

This works for embedded MSWord file like Excel or PowerPoint file, but won't work for others like PDF, Python or even Text files : For these ones, MSWord generate an oleObjectNNN.bin file which is no use to be replaced as it is encoded.

Microsoft Word 2016 special cases

MS Word 2016 will ignore \t tabulations. This is special to that version. Libreoffice or Wordpad do not have this problem. The same thing occurs for line beginning with a jinja2 tag provinding spaces : They will be ignored. To solve these problem, the solution is to use Richtext:

tpl.render({
    'test_space_r' : RichText('          '),
    'test_tabs_r': RichText(5*'\t'),
})

And in your template, use the {{r notation:

{{r test_space_r}} Spaces will be preserved
{{r test_tabs_r}} Tabs will be displayed

Tables

You can span table cells horizontally in two ways, by using colspan tag (see tests/dynamic_table.py):

{% colspan <number of column to span> %}

or within a for loop (see tests/horizontal_merge.py):

{% hm %}

You can also merge cells vertically within a for loop (see tests/vertical_merge.py):

{% vm %}

Jinja custom filters

render() accepts jinja_env optionnal argument : you may pass a jinja environment object. By this way you will be able to add some custom jinja filters:

from docxtpl import DocxTemplate
import jinja2

def multiply_by(value, by):
   return value * by

doc = DocxTemplate("my_word_template.docx")
context = { 'price_dollars' : 5.00 }
jinja_env = jinja2.Environment()
jinja_env.filters['multiply_by'] = multiply_by
doc.render(context,jinja_env)
doc.save("generated_doc.docx")

Then in your template, you will be able to use:

Euros price : {{ price_dollars|multiply_by(0.88) }}

Examples

The best way to see how it works is to read examples, they are located in tests/ directory. Docx test templates are in tests/templates/. To generate final docx files:

cd tests/
python runtests.py

Generated files are located in tests/output directory.

If you are not sure about your python environment, python-docx-template provides Pipfiles for that:

pip install pipenv (if not already done)
cd python-docx-template (where Pipfiles are)
pipenv install --python 3.6 -d
pipenv shell
cd tests/
python runtests.py

Share

If you like this project, please rate and share it here : http://rate.re/github/elapouya/python-docx-template

Functions index

.. currentmodule:: docxtpl

Functions documentation

.. automodule:: docxtpl
   :members:

Indices and tables

:ref:`genindex`
:ref:`modindex`
:ref:`search`

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

index.rst

index.rst

Welcome to python-docx-template's documentation!

Introduction

Jinja2-like syntax

Restrictions

Extensions

Tags

Split and merge text

Display variables

Cell color

Column spanning

Escaping

RichText

Hyperlink with RichText

Inline image

Sub-documents

Escaping, newline, new paragraph, Listing

Replace docx pictures

Replace docx medias

Replace embedded objects

Microsoft Word 2016 special cases

Tables

Jinja custom filters

Examples

Share

Indices and tables

Files

index.rst

Latest commit

History

index.rst

File metadata and controls

Welcome to python-docx-template's documentation!

Introduction

Jinja2-like syntax

Restrictions

Extensions

Tags

Split and merge text

Display variables

Cell color

Column spanning

Escaping

RichText

Hyperlink with RichText

Inline image

Sub-documents

Escaping, newline, new paragraph, Listing

Replace docx pictures

Replace docx medias

Replace embedded objects

Microsoft Word 2016 special cases

Tables

Jinja custom filters

Examples

Share

Indices and tables