Hyperfast and lightweight templating for the Python platform.
Home   |   Community   |   Documentation   |   Download

Mako 1.0.1 Documentation

Release: 1.0.1

Caching

Any template or component can be cached using the cache argument to the <%page>, <%def> or <%block> directives:

<%page cached="True"/>

template text

The above template, after being executed the first time, will store its content within a cache that by default is scoped within memory. Subsequent calls to the template’s render() method will return content directly from the cache. When the Template object itself falls out of scope, its corresponding cache is garbage collected along with the template.

By default, caching requires that the Beaker package be installed on the system, however the mechanism of caching can be customized to use any third party or user defined system – see Cache Plugins.

In addition to being available on the <%page> tag, the caching flag and all its options can be used with the <%def> tag as well:

<%def name="mycomp" cached="True" cache_timeout="60">
    other text
</%def>

... and equivalently with the <%block> tag, anonymous or named:

<%block cached="True" cache_timeout="60">
    other text
</%block>

Cache Arguments

Mako has two cache arguments available on tags that are available in all cases. The rest of the arguments available are specific to a backend.

The two generic tags arguments are:

  • cached="True" - enable caching for this <%page>, <%def>, or <%block>.

  • cache_key - the “key” used to uniquely identify this content in the cache. Usually, this key is chosen automatically based on the name of the rendering callable (i.e. body when used in <%page>, the name of the def when using <%def>, the explicit or internally-generated name when using <%block>). Using the cache_key parameter, the key can be overridden using a fixed or programmatically generated value.

    For example, here’s a page that caches any page which inherits from it, based on the filename of the calling template:

    <%page cached="True" cache_key="${self.filename}"/>
    
    ${next.body()}
    
    ## rest of template
    

On a Template or TemplateLookup, the caching can be configured using these arguments:

  • cache_enabled - Setting this to False will disable all caching functionality when the template renders. Defaults to True. e.g.:

    lookup = TemplateLookup(
                    directories='/path/to/templates',
                    cache_enabled = False
                    )
    
  • cache_impl - The string name of the cache backend to use. This defaults to 'beaker', which has historically been the only cache backend supported by Mako.

    New in version 0.6.0.

    For example, here’s how to use the upcoming dogpile.cache backend:

    lookup = TemplateLookup(
                    directories='/path/to/templates',
                    cache_impl = 'dogpile.cache',
                    cache_args = {'regions':my_dogpile_regions}
                    )
    
  • cache_args - A dictionary of cache parameters that will be consumed by the cache backend. See Using the Beaker Cache Backend for examples.

    New in version 0.6.0.

Backend-Specific Cache Arguments

The <%page>, <%def>, and <%block> tags accept any named argument that starts with the prefix "cache_". Those arguments are then packaged up and passed along to the underlying caching implementation, minus the "cache_" prefix.

The actual arguments understood are determined by the backend.

Using the Beaker Cache Backend

When using Beaker, new implementations will want to make usage of cache regions so that cache configurations can be maintained externally to templates. These configurations live under named “regions” that can be referred to within templates themselves.

New in version 0.6.0: Support for Beaker cache regions.

For example, suppose we would like two regions. One is a “short term” region that will store content in a memory-based dictionary, expiring after 60 seconds. The other is a Memcached region, where values should expire in five minutes. To configure our TemplateLookup, first we get a handle to a beaker.cache.CacheManager:

from beaker.cache import CacheManager

manager = CacheManager(cache_regions={
    'short_term':{
        'type': 'memory',
        'expire': 60
    },
    'long_term':{
        'type': 'ext:memcached',
        'url': '127.0.0.1:11211',
        'expire': 300
    }
})

lookup = TemplateLookup(
                directories=['/path/to/templates'],
                module_directory='/path/to/modules',
                cache_impl='beaker',
                cache_args={
                    'manager':manager
                }
        )

Our templates can then opt to cache data in one of either region, using the cache_region argument. Such as using short_term at the <%page> level:

<%page cached="True" cache_region="short_term">

## ...

Or, long_term at the <%block> level:

<%block name="header" cached="True" cache_region="long_term">
    other text
</%block>

The Beaker backend also works without regions. There are a variety of arguments that can be passed to the cache_args dictionary, which are also allowable in templates via the <%page>, <%block>, and <%def> tags specific to those sections. The values given override those specified at the TemplateLookup or Template level.

With the possible exception of cache_timeout, these arguments are probably better off staying at the template configuration level. Each argument specified as cache_XYZ in a template tag is specified without the cache_ prefix in the cache_args dictionary:

  • cache_timeout - number of seconds in which to invalidate the cached data. After this timeout, the content is re-generated on the next call. Available as timeout in the cache_args dictionary.
  • cache_type - type of caching. 'memory', 'file', 'dbm', or 'ext:memcached' (note that the string memcached is also accepted by the dogpile.cache Mako plugin, though not by Beaker itself). Available as type in the cache_args dictionary.
  • cache_url - (only used for memcached but required) a single IP address or a semi-colon separated list of IP address of memcache servers to use. Available as url in the cache_args dictionary.
  • cache_dir - in the case of the 'file' and 'dbm' cache types, this is the filesystem directory with which to store data files. If this option is not present, the value of module_directory is used (i.e. the directory where compiled template modules are stored). If neither option is available an exception is thrown. Available as dir in the cache_args dictionary.

Using the dogpile.cache Backend

dogpile.cache is a new replacement for Beaker. It provides a modernized, slimmed down interface and is generally easier to use than Beaker. As of this writing it has not yet been released. dogpile.cache includes its own Mako cache plugin – see dogpile.cache.plugins.mako_cache in the dogpile.cache documentation.

Programmatic Cache Access

The Template, as well as any template-derived Namespace, has an accessor called cache which returns the Cache object for that template. This object is a facade on top of the underlying CacheImpl object, and provides some very rudimental capabilities, such as the ability to get and put arbitrary values:

<%
    local.cache.set("somekey", type="memory", "somevalue")
%>

Above, the cache associated with the local namespace is accessed and a key is placed within a memory cache.

More commonly, the cache object is used to invalidate cached sections programmatically:

template = lookup.get_template('/sometemplate.html')

# invalidate the "body" of the template
template.cache.invalidate_body()

# invalidate an individual def
template.cache.invalidate_def('somedef')

# invalidate an arbitrary key
template.cache.invalidate('somekey')

You can access any special method or attribute of the CacheImpl itself using the impl attribute:

template.cache.impl.do_something_special()

Note that using implementation-specific methods will mean you can’t swap in a different kind of CacheImpl implementation at a later time.

Cache Plugins

The mechanism used by caching can be plugged in using a CacheImpl subclass. This class implements the rudimental methods Mako needs to implement the caching API. Mako includes the BeakerCacheImpl class to provide the default implementation. A CacheImpl class is acquired by Mako using a pkg_resources entrypoint, using the name given as the cache_impl argument to Template or TemplateLookup. This entry point can be installed via the standard setuptools/setup() procedure, underneath the EntryPoint group named "mako.cache". It can also be installed at runtime via a convenience installer register_plugin() which accomplishes essentially the same task.

An example plugin that implements a local dictionary cache:

from mako.cache import Cacheimpl, register_plugin

class SimpleCacheImpl(CacheImpl):
    def __init__(self, cache):
        super(SimpleCacheImpl, self).__init__(cache)
        self._cache = {}

    def get_or_create(self, key, creation_function, **kw):
        if key in self._cache:
            return self._cache[key]
        else:
            self._cache[key] = value = creation_function()
            return value

    def set(self, key, value, **kwargs):
        self._cache[key] = value

    def get(self, key, **kwargs):
        return self._cache.get(key)

    def invalidate(self, key, **kwargs):
        self._cache.pop(key, None)

# optional - register the class locally
register_plugin("simple", __name__, "SimpleCacheImpl")

Enabling the above plugin in a template would look like:

t = Template("mytemplate",
             file="mytemplate.html",
             cache_impl='simple')

Guidelines for Writing Cache Plugins

  • The CacheImpl is created on a per-Template basis. The class should ensure that only data for the parent Template is persisted or returned by the cache methods. The actual Template is available via the self.cache.template attribute. The self.cache.id attribute, which is essentially the unique modulename of the template, is a good value to use in order to represent a unique namespace of keys specific to the template.
  • Templates only use the CacheImpl.get_or_create() method in an implicit fashion. The CacheImpl.set(), CacheImpl.get(), and CacheImpl.invalidate() methods are only used in response to direct programmatic access to the corresponding methods on the Cache object.
  • CacheImpl will be accessed in a multithreaded fashion if the Template itself is used multithreaded. Care should be taken to ensure caching implementations are threadsafe.
  • A library like Dogpile, which is a minimal locking system derived from Beaker, can be used to help implement the CacheImpl.get_or_create() method in a threadsafe way that can maximize effectiveness across multiple threads as well as processes. CacheImpl.get_or_create() is the key method used by templates.
  • All arguments passed to **kw come directly from the parameters inside the <%def>, <%block>, or <%page> tags directly, minus the "cache_" prefix, as strings, with the exception of the argument cache_timeout, which is passed to the plugin as the name timeout with the value converted to an integer. Arguments present in cache_args on Template or TemplateLookup are passed directly, but are superseded by those present in the most specific template tag.
  • The directory where Template places module files can be acquired using the accessor self.cache.template.module_directory. This directory can be a good place to throw cache-related work files, underneath a prefix like _my_cache_work so that name conflicts with generated modules don’t occur.

API Reference

class mako.cache.Cache(template, *args)

Bases: object

Represents a data content cache made available to the module space of a specific Template object.

New in version 0.6: Cache by itself is mostly a container for a CacheImpl object, which implements a fixed API to provide caching services; specific subclasses exist to implement different caching strategies. Mako includes a backend that works with the Beaker caching system. Beaker itself then supports a number of backends (i.e. file, memory, memcached, etc.)

The construction of a Cache is part of the mechanics of a Template, and programmatic access to this cache is typically via the Template.cache attribute.

get(key, **kw)

Retrieve a value from the cache.

Parameters:
  • key – the value’s key.
  • **kw – cache configuration arguments. The backend is configured using these arguments upon first request. Subsequent requests that use the same series of configuration values will use that same backend.
get_or_create(key, creation_function, **kw)

Retrieve a value from the cache, using the given creation function to generate a new value.

id = None

Return the ‘id’ that identifies this cache.

This is a value that should be globally unique to the Template associated with this cache, and can be used by a caching system to name a local container for data specific to this template.

impl = None

Provide the CacheImpl in use by this Cache.

This accessor allows a CacheImpl with additional methods beyond that of Cache to be used programmatically.

invalidate(key, **kw)

Invalidate a value in the cache.

Parameters:
  • key – the value’s key.
  • **kw – cache configuration arguments. The backend is configured using these arguments upon first request. Subsequent requests that use the same series of configuration values will use that same backend.
invalidate_body()

Invalidate the cached content of the “body” method for this template.

invalidate_closure(name)

Invalidate a nested <%def> within this template.

Caching of nested defs is a blunt tool as there is no management of scope – nested defs that use cache tags need to have names unique of all other nested defs in the template, else their content will be overwritten by each other.

invalidate_def(name)

Invalidate the cached content of a particular <%def> within this template.

put(key, value, **kw)

A synonym for Cache.set().

This is here for backwards compatibility.

set(key, value, **kw)

Place a value in the cache.

Parameters:
  • key – the value’s key.
  • value – the value.
  • **kw – cache configuration arguments.
starttime = None

Epochal time value for when the owning Template was first compiled.

A cache implementation may wish to invalidate data earlier than this timestamp; this has the effect of the cache for a specific Template starting clean any time the Template is recompiled, such as when the original template file changed on the filesystem.

class mako.cache.CacheImpl(cache)

Bases: object

Provide a cache implementation for use by Cache.

get(key, **kw)

Retrieve a value from the cache.

Parameters:
  • key – the value’s key.
  • **kw – cache configuration arguments.
get_or_create(key, creation_function, **kw)

Retrieve a value from the cache, using the given creation function to generate a new value.

This function must return a value, either from the cache, or via the given creation function. If the creation function is called, the newly created value should be populated into the cache under the given key before being returned.

Parameters:
  • key – the value’s key.
  • creation_function – function that when called generates a new value.
  • **kw – cache configuration arguments.
invalidate(key, **kw)

Invalidate a value in the cache.

Parameters:
  • key – the value’s key.
  • **kw – cache configuration arguments.
pass_context = False

If True, the Context will be passed to get_or_create as the name 'context'.

set(key, value, **kw)

Place a value in the cache.

Parameters:
  • key – the value’s key.
  • value – the value.
  • **kw – cache configuration arguments.
mako.cache.register_plugin(self, name, modulepath, objname)
class mako.ext.beaker_cache.BeakerCacheImpl(cache)

Bases: mako.cache.CacheImpl

A CacheImpl provided for the Beaker caching system.

This plugin is used by default, based on the default value of 'beaker' for the cache_impl parameter of the Template or TemplateLookup classes.