Kajiki XML Templates

Kajiki provides a full-featured XML-based template engine that guarantees well-formed output when generating HTML and XML. This document describes that language. Templates are XML files that include template directives that control how the template is rendered and expressions that are substituted into the generated text at render time.

Please see Kajiki Templating Basics for general information on embedding Python code in templates.

Output Modes

Although Kajiki XML templates must be well-formed XML documents, Kajiki is capable of rendering HTML or XML. By default, Kajiki will inspect the doctype of the template to determine how to render:

>>> tpl_text = '''<!DOCTYPE html>
... <html>
...     <head><!-- Some stuff here --></head>
...     <body>
...         <form>
...             <input type="checkbox" checked="checked"/>
...             <select>
...                 <option selected="selected">One</option>
...                 <option>Two</option>
...                 <option>Three</option>
...             </select>
...         </form>
...     </body>
... </html>'''
>>> import kajiki
>>> Template = kajiki.XMLTemplate(tpl_text)
>>> print(Template().render().strip())
<!DOCTYPE html>
<html>
    <head><!--  Some stuff here  --></head>
    <body>
        <form>
            <input checked type="checkbox">
            <select>
                <option selected>One</option>
                <option>Two</option>
                <option>Three</option>
            </select>
        </form>
    </body>
</html>
>>> # If we want to override the detected type, we can pass a 'mode' param
>>> Template = kajiki.XMLTemplate(tpl_text, mode='xml')
>>> print(Template().render().strip())
<!DOCTYPE html>
<html>
    <head><!--  Some stuff here  --></head>
    <body>
        <form>
            <input checked="checked" type="checkbox"/>
            <select>
                <option selected="selected">One</option>
                <option>Two</option>
                <option>Three</option>
            </select>
        </form>
    </body>
</html>
>>> # We can also omit the generated DOCTYPE by specifying the template
>>> # is a fragment
>>> Template = kajiki.XMLTemplate(tpl_text, mode='xml', is_fragment=True)
>>> print(Template().render().strip())
<html>
    <head><!--  Some stuff here  --></head>
    <body>
        <form>
            <input checked="checked" type="checkbox"/>
            <select>
                <option selected="selected">One</option>
                <option>Two</option>
                <option>Three</option>
            </select>
        </form>
    </body>
</html>

Note

In Kajiki, you can use normal XML comments, and the comments will exist in the generated markup. If you wish the comments to be removed before rendering the document, you can begin the comment with the syntax <!–! instead of <!–:

>>> Template = kajiki.XMLTemplate('''<div>
... <!-- This comment is preserved.
... --><!--! This comment is stripped. -->
... </div>''')
>>> print(Template().render())
<div>
<!--  This comment is preserved.
 -->
</div>

Basic Expressions

Let’s start with a hello world template:

>>> Template = kajiki.XMLTemplate('<div>Hello, $name!</div>')
>>> print(Template(dict(name='world')).render())
<div>Hello, world!</div>

By default, the $-syntax picks up any identifiers following it, as well as any periods. If you want something more explicit, use the extended expression form as follows:

>>> Template = kajiki.XMLTemplate('<div>Hello, 2+2 is ${2+2}</div>')
>>> print(Template().render())
<div>Hello, 2+2 is 4</div>

If you wish to include a literal $, simply double it:

>>> Template = kajiki.XMLTemplate('<div>The price is $$${price}</div>')
>>> print(Template(dict(price='5.00')).render())
<div>The price is $5.00</div>

You can also include expressions in template attributes:

>>> Template = kajiki.XMLTemplate('<div id="$foo">Bar</div>')
>>> print(Template(dict(foo='baz')).render())
<div id="baz">Bar</div>

Control Flow

Kajiki provides several directives that affect the rendering of a template. This section describes the various directives. Directives in text templates can either appear as special attributes on tags prefixed by py: or as standalone tags whose tagname is prefixed by py:.

py:if, py:else

Only render the enclosed content if the expression evaluates to a truthy value:

>>> Template = kajiki.XMLTemplate('<div><py:if test="foo">bar</py:if><py:else>baz</py:else></div>')
>>> print(Template(dict(foo=True)).render())
<div>bar</div>
>>> print(Template(dict(foo=False)).render())
<div>baz</div>
>>> Template = kajiki.XMLTemplate('<div><span py:if="foo">bar</span></div>')
>>> print(Template(dict(foo=True)).render())
<div><span>bar</span></div>
>>> print(Template(dict(foo=False)).render())
<div></div>

py:switch, py:case, py:else

Perform multiple tests to render one of several alternatives. The first matching case is rendered, and if no case matches, the else branch is rendered:

>>> Template = kajiki.XMLTemplate('''<div>
... $i is <py:switch test="i % 2">
... <py:case value="0">even</py:case>
... <py:else>odd</py:else>
... </py:switch></div>''')
>>> print(Template(dict(i=4)).render())
<div>
4 is even</div>
>>> print(Template(dict(i=3)).render())
<div>
3 is odd</div>

py:match, py:case

Similar to py:switch this makes use of PEP622 Structural Pattern Matching:

>>> import sys, pytest
>>> if sys.version_info < (3, 10): pytest.skip('pep622 unsupported')
>>> Template = kajiki.XMLTemplate('''<div>
... $i is <py:match on="i % 2">
... <py:case match="0">even</py:case>
... <py:case match="_">odd</py:case>
... </py:match></div>''')
>>> print(Template(dict(i=4)).render())
<div>
4 is even</div>
>>> print(Template(dict(i=3)).render())
<div>
3 is odd</div>

Note

py:match compiles directly to Python’s match syntax, and will therefore not work on versions less than 3.10. Only use this syntax if your project targets Python 3.10 or newer.

py:for

Repeatedly render the content for each item in an iterable:

>>> Template = kajiki.XMLTemplate('''<ul>
... <li py:for="x in range(sz)">$x</li>
... </ul>''')
>>> print(Template(dict(sz=3)).render())
<ul>
<li>0</li><li>1</li><li>2</li>
</ul>

py:def

Defines a function that can be used elsewhere in the template:

>>> Template = kajiki.XMLTemplate('''<div
... ><py:def function="evenness(n)"
... ><py:if test="n%2==0">even</py:if><py:else>odd</py:else></py:def
... ><ul>
... <li py:for="x in range(sz)">$x is ${evenness(x)}</li>
... </ul></div>''')
>>> print(Template(dict(sz=3)).render())
<div><ul>
<li>0 is even</li><li>1 is odd</li><li>2 is even</li>
</ul></div>

py:call

Call a function, passing a block of template code as a ‘lambda’ parameter. Note that this is a special case of calling when you wish to insert some templated text in the expansion of a function call. In normal circumstances, you would just use ${my_function(args)}.

>>> Template = kajiki.XMLTemplate('''<div
... ><py:def function="quote(caller, speaker)"
... ><ul>
...    <li py:for="i in range(sz)">Quoth $speaker, ${caller(i)}</li>
... </ul></py:def
... ><py:call args="n" function="quote(%caller, 'the raven')"
... >Nevermore $n</py:call></div>''')
>>> print(Template(dict(sz=3)).render())
<div><ul>
   <li>Quoth the raven, Nevermore 0</li><li>Quoth the raven, Nevermore 1</li><li>Quoth the raven, Nevermore 2</li>
</ul></div>

py:include

Includes the text of another template verbatim. The precise semantics of this tag depend on the TemplateLoader being used, as the TemplateLoader is used to parse the name of the template being included and render its contents into the current template. For instance, with the FileLoader, you might use the following:

<py:include href="path/to/base.txt"/>

whereas in the PackageLoader you would use

<py:include href="package1.package2.base"/>

py:import

With py:import, you can make the functions defined in another template available without expanding the full template in-place. Suppose that we saved the following template in a file lib.xml:

<py:def function="evenness(n)">
   <py:if test="n%2==0">even</py:if><py:else>odd</py:else>
</py:def>

Then (using the FileLoader) we could write a template using the evenness function as follows:

<div>
   <py:import href="lib.xml" alias="lib"/>
   <ul>
      <li py:for="i in range(sz)">$i is ${lib.evenness(i)}</li>
   </ul>
</div>

py:with

Using py:with, you can temporarily assign variables values for the extent of the block:

>>> Template = kajiki.XMLTemplate('''<div py:with="a='foo'">
... <div>$a</div>
... <div py:with="a=5">$a</div>
... <div>$a</div>
... </div>''')
>>> print(Template().render())
<div>
<div>foo</div>
<div>5</div>
<div>foo</div>
</div>

Content Generation

py:attrs

With the py:attrs custom attribute, you can include dynamic attributes in an xml/html tag by passing a either a Python dict or a list of pairs:

>>> Template = kajiki.XMLTemplate('<div py:attrs="attrs"/>')
>>> print(Template(dict(attrs={'id':'foo', 'class':'bar'})).render())
<div class="bar" id="foo"/>
>>> print(Template(dict(attrs=[('id', 'foo'), ('class', 'bar')])).render())
<div class="bar" id="foo"/>

Any attribute values that evaluate to None will not be emitted in the generated markup:

>>> Template = kajiki.XMLTemplate('<div py:attrs="attrs"/>')
>>> print(Template(dict(attrs={'id':'foo', 'class':None})).render())
<div id="foo"/>

py:strip

With py:strip, you can remove the tag to which the attribute is attached without removing the content of the tag:

>>> Template = kajiki.XMLTemplate('<div><div py:strip="True">Foo</div></div>')
>>> print(Template().render())
<div>Foo</div>

As a shorthand, if the value of the py:strip attribute is empty, that has the same effect as using a truth value (i.e. the element is stripped).

py:content

With py:content, you can remove the tag to which the attribute is attached without removing the content of the tag:

>>> Template = kajiki.XMLTemplate('<div py:content="content"/>')
>>> print(Template(dict(content="Foo")).render())
<div>Foo</div>

py:replace

With py:replace, you can replace the entire tag to which the document is attached and its children:

>>> Template = kajiki.XMLTemplate('<div py:replace="content"/>')
>>> print(Template(dict(content="Foo")).render())
Foo

Inheritance (py:extends, py:block)

Kajiki supports a concept of inheritance whereby child templates can extend parent templates, replacing their “methods” (functions) and “blocks” (to be defined below). For instance, consider the following template “parent.xml”:

<div>
   <py:def function="greet(name)"
      >Hello, $name!</py:def>
   <py:def function="sign(name)"
      >Sincerely,<br/>
      <em>$name</em></py:def>
   ${greet(to)}

   <p py:block="body">It was good seeing you last Friday.
   Thanks for the gift!</p>

   ${sign(from_)}
</div>

This would render to something similar to the following (assuming a context of dict(to=Mark, from_=Rick):

<div>
   Hello, Mark!

   <p>It was good seeing you last friday.
   Thanks for the gift!</p>

   Sincerely, <br/>
   Rick
</div>

Now we can extend “parent.xml” with “child.xml”:

<py:extends href="parent.xml">
   <py:def function="greet(name)"
   >Dear $name:</py:def>
   <py:block name="body">${parent_block()}
   <p>And don't forget you owe me money!</p>
   </py:block>
</py:extends>

Rendering this template would then give us:

<div>
   Dear, Mark:

   <p>It was good seeing you last friday.
   Thanks for the gift!</p>
   <p>And don't forget you owe me money!</p>

   Sincerely, <br/>
   Rick
</div>

Notice how in the child block, we have overridden both the block “body” and the function “greet.” When overriding a block, we always have access to the parent template’s block of the same name via the parent_block() function.

If you ever need to access the parent template itself (perhaps to call another function), kajiki provides access to a special variable in child templates parent. Likewise, if a template is being extended, the variable child is available. Kajiki also provides the special variables local (the template currently being defined) and self (the child-most template of an inheritance chain). The following example illustrates these variables in a 3-level inheritance hierarchy:

>>> parent = kajiki.XMLTemplate('''<div
... ><h1 py:def="header()">Header name=$name</h1
... ><h6 py:def="footer()">Footer</h6
... ><div py:def="body()">
... id() = ${id()}
... local.id() = ${local.id()}
... self.id() = ${self.id()}
... child.id() = ${child.id()}
... </div><span py:def="id()">parent</span>
... ${header()}
... ${body()}
... ${footer()}
... </div>''')
>>> mid = kajiki.XMLTemplate('''<py:extends href="parent.html"
... ><span py:def="id()">mid</span
... ></py:extends>''')
>>> child=kajiki.XMLTemplate('''<py:extends href="mid.html"
... ><span py:def="id()">child</span
... ><div py:def="body()">
... <h2>Child Body</h2>
... ${parent.body()}
... </div></py:extends>''')
>>> loader = kajiki.MockLoader({
... 'parent.html':parent,
... 'mid.html':mid,
... 'child.html':child})
>>> Template = loader.import_('child.html')
>>> print(Template(dict(name='Rick')).render())
<div>
<h1>Header name=Rick</h1>
<div>
<h2>Child Body</h2>
<div>
id() = <span>child</span>
local.id() = <span>parent</span>
self.id() = <span>child</span>
child.id() = <span>mid</span>
</div>
</div>
<h6>Footer</h6>
</div>

Summary of Directives

Directive

Usable as an attribute

Usable as a separate element

When used as a separate element, requires attributes named

py:if

test

py:else

py:switch

test

py:match

on

py:case

value or match (for usage with py:switch or py:match)

py:for

each

py:def

function

py:call

args, function

py:include

href

py:import

href

py:with

vars

py:attrs

py:strip

py:content

py:replace

value

py:extends

href

py:block

name

Built-in functions

The following functions are available by default in template code, in addition to the standard built-ins that are available to all Python code.

defined(name)

This function determines whether a variable of the specified name exists in the context data, and returns True if it does.

When would you use it? Well, suppose you tried the following template snippet:

<h3 py:if=’user’>$user.name</h3>

If you don’t pass, from your python code, a “user” variable to the template, the above code will fail with this exception: NameError: global name 'user' is not defined. This is undesired!

Following Genshi, Kajiki offers the defined() function to make that condition possible, so you can write this:

<h3 py:if=”defined(‘user’)”>$user.name</h3>

literal(text)

All good templating languages escape text by default to avoid certain attacks. But sometimes you have an HTML snippet that you wish to include in a page, and you know the HTML is safe.

The literal() function marks a given string as being safe for inclusion, meaning it will not be escaped in the serialization stage. Use this with care, as not escaping a user-provided string may allow malicious users to open your web site to cross-site scripting attacks. Example:

${literal(name_of_the_variable_containing_the_html)}

Kajiki is a reimplementation of most of Genshi and, since Genshi has a Markup() function for the same purpose, we provide Markup() as a synonym, too.

value_of(name, default=None)

This function returns the value of the variable with the specified name if such a variable is defined, and returns the value of the default parameter if no such variable is defined.

Genshi has this, too. Example:

<div py:if="value_of('explanation')">${explanation}</div>

In the above example, the div will only appear in the output if the explanation variable exists in the context and has a truish value. (Remember in Python, None and the empty string are not truish, they are evaluated as False.)

Tips on writing your templates

Kajiki takes XML as input, with the exception that it recognizes HTML entities in addition to XML entities. (HTML has many more entities than XML; for instance, XML does not define &nbsp).

If your template contains complex content in <style> or <script> tags, you should either:

  1. Externalize these contents onto their own files, or

  2. Remembering that Kajiki takes XML as input, use CDATA sections inside these tags in order to escape characters that are illegal in XML, such as <, > and &. Here is an example:

    <style><![CDATA[
        html > body { display: none; }
    ]]></style>
    <script><![CDATA[
        if (1 < 2) { document.write("<p>Albatross!!!</p>"); }
    ]]></script>
    

This is not necessary when you are writing HTML because HTML defines that the content of <style> and <script> tags is CDATA by default. However, Kajiki takes XML as input.