Official SiteAPI Reference

Interfaces and Handlers

Introduction to Interfaces

Cement builds upon a standard interface and handler system that is used extensively to break up pieces of the framework into relatable chunks, and allow customization of everything from logging to config file parsing, and almost every operation in between.

In Cement, an interface is what defines some functionality, and a handler is what implements that functionality.

We call the implementation of an interface a handler, and provide the ability to easily register and retrieve them via the app.handler object. Cement interfaces are defined as Python Abstract Base Classes, and handlers implement them by sub-classing and overriding the defined abstract methods required to make the implementation legit.

API References

Builtin Interfaces

The following interfaces are builtin to Cement's core foundation:

Interface

Description

Framework extension loading.

Messaging to console, and/or file via common log facilities (INFO, WARNING, ERROR, FATAL, DEBUG).

Merging of application configuration defaults, configuration files, and environment settings into a single config object.

Remote message sending (email, smtp, etc).

Application plugin loading.

Rendering of template data (content, files, etc).

Rendering of data/content to end-user output (console text from template, JSON, Yaml, etc). Often uses an associated template handler backend.

Command line argument/option parsing.

Command dispatch (sub-commands, arguments, etc)

Key/Value data store (memcached, redis, etc)

Working With Interfaces

The InterfaceManager (app.interface) provides quick mechanisms to list, get, and verify interfaces:

from cement import App

with App('myapp') as app:
    # list defined interfaces
    app.interface.list()

    # get an interface class
    i = app.interface.get('output')

    # validate if an interface is defined
    app.interface.defined('mail')

Defining an Interface

Defining interfaces is more of an advanced topic, and is not required to fully grasp for new developers or those new to the framework.

Cement uses interfaces and handlers extensively to manage the framework, however developers can also make use of this system to provide a clean, and standardized way of allowing other developers to customize their application (generally via application plugins).

The following defines a basic interface we'll call greeting:

from abc import abstractmethod
from cement import App, Interface

​class GreetingInterface(Interface):
    class Meta:
        interface = 'greeting'

    @abstractmethod
    def _get_greeting(self):
        """
        Get a greeting message for the end-user.

        Returns:
            greeting (str): The greeting string to present to
                the end-user.
        """
        pass

    @abstractmethod
    def greet(self):
        """
        Display a greeting message for the end-user.

        Returns: None
        """
        pass

class MyApp(App):
    label = 'myapp'
    interfaces = [
        GreetingInterface,
    ]

The above example defines the greeting interface, by providing abstract methods that any handlers implementing this interface must provide. It does not implement any functionality on its own (though it could), but rather defines and documents its purpose and its expected implementation. The interface is easily defined with the framework by listing it in App.Meta.interfaces, but you can also define interfaces directly with app.interface.define().

Implementing an Interface

In order to implement the above greeting interface, we first want to provide a handler base class that will be the starting point for all implementations that sub-class from it:

from cement import Handler

class GreetingHandler(GreetingInterface, Handler):

    def greet(self):
        self.app.log.debug('about to greet end-user')
        msg = self._get_greeting()
        assert isinstance(str, msg), "The msg is not a string!"
        print(msg)

Handler base classes sub-class from both the interface they are implementing (GreetingInterface), and also the Cement Handler base class (Handler). The application developer defining the interface should always provide a handler base class for the implementation, even if the base class does not fully satisfy the interface.

In the above example, the developer would call the greet() method that will do all of the common operations like logging, exception handling, etc for all implementations, leaving only the minimal _get_greeting() method to be provided by the final implementation sub-classes. This follows the common Don't Repeat Yourself (DRY) principle's best practice (all-be-it a tediously simple example), where all of the reusable logic can live in one place, and sub-classes only focus on their unique means of implementing an interface.

In order to complete our implementation of the above greeting interface, we can now sub-class from the provided GreetingHandler base class, and fill in the missing pieces:

class Hello(GreetingHandler):
    class Meta:
        label = 'hello'

    def _get_greeting(self):
        return 'Hello!'


class Goodbye(GreetingHandler):
    class Meta:
        label = 'goodbye'

    def _get_greeting(self):
        return 'Goodbye!'

An interface defines itself to the framework via the Interface.Meta.interface string, and a handler defines itself via the Handler.Meta.label string. Collectively, all implementation handlers are referred to by both as <interface>.<label> or in the above examples greeting.hello and greeting.goodbye.

Putting It All Together

The following is an MCVE of defining an interface, providing an implementation handler base class, and registering multiple handlers that implement the interface differently:

from abc import abstractmethod
from cement import App, Interface, Handler

class GreetingInterface(Interface):
    class Meta:
        interface = 'greeting'

    @abstractmethod
    def _get_greeting(self):
        """
        Get a greeting message for the end-user.

        Returns:
            greeting (str): The greeting string to present to
                the end-user.
        """
        pass

    @abstractmethod
    def greet(self):
        """
        Display a greeting message for the end-user.

        Returns: None
        """
        pass


class GreetingHandler(GreetingInterface, Handler):

    def greet(self):
        self.app.log.debug('about to greet end-user')
        msg = self._get_greeting()
        assert isinstance(msg, str), "The msg is not a string!"
        print(msg)


class Hello(GreetingHandler):
    class Meta:
        label = 'hello'

    def _get_greeting(self):
        return 'Hello!'


class Goodbye(GreetingHandler):
    class Meta:
        label = 'goodbye'

    def _get_greeting(self):
        return 'Goodbye!'


class MyApp(App):
    class Meta:
        label = 'myapp'

        interfaces = [
            GreetingInterface,
        ]

        handlers = [
            Hello,
            Goodbye,
        ]

with MyApp() as app:
    app.run()
    g = app.handler.get('greeting', 'hello', setup=True)
    g.greet()

Introduction to Handlers

Where an interface defines what an implementation is expected to provide, a handler is what actually implements the interface.

Working with Handlers

The HandlerManager (app.handler) provides quick mechanisms to list, get, resolve and verify handlers. The following are a few examples of working with handlers:

from cement import App

​with App('myapp') as app:
    # get a log handler called logging
    lh = app.handler.get('log', 'logging', setup=True)​

    # List all handlers registered to the config interface
    app.handler.list('config')​

    # check if the handler argparse is registered
    # to the argument interface
    app.handler.registered('argument', 'argparse')

    # resolve a handler by string, class, or object
    app.handler.resolve('log', 'logging')
    app.handler.resolve('log', LoggingLogHandler)
    app.handler.resolve('log', LoggingLogHandler())

It is important to note that handlers are stored within the application as uninstantiated objects. Meaning you must instantiate them after retrieval (call _setup(app)) or simply pass setup=True to the app.handler.get() method.

Registering Handlers to an Interface

An interface simply defines what an implementation is expected to provide, whereas a handler actually implements the interface.

The following is a simple example of sub-classing an existing handler, then registering that with the framework.

from cement import App
from cement.ext.ext_configparser import ConfigParserConfigHandler

class MyConfigHandler(ConfigParserConfigHandler):
    class Meta:
        label = 'my_config_handler'

    # do something to sub-class ConfigParserConfigHandler

class MyApp(App):
    class Meta:
        label = 'myapp'
        handlers = [MyConfigHandler]

Overriding Default Handlers

Cement sets up a number of default handlers for logging, config parsing, etc. These can be overridden in a number of ways, however the primary method is to override their associated setting in App.Meta.

Using the above MyConfigHandler example, we simply need to register a handler to an interface, and then override the App.Meta.config_handler option:

class MyApp(App):
    class Meta:
        label = 'myapp'
        handlers = [MyConfigHandler]
        config_handler = 'my_config_handler'

All builtin core interfaces have an associated App.Meta.x_handler option, allowing complete customization of every aspect of the framework. See the reference documentation of App.Meta for more detail.

Multiple Registered Handlers

All handlers and interfaces are unique. In most cases, where the framework is concerned, only one handler is used. For example, whatever is configured for the App.Meta.log_handler will be used and setup as app.log. However, take for example an Output Handler. You might have a default App.Meta.output_handler of mustache (a text templating language) but may also want to override that handler with the json output handler when -o json is passed at command line. In order to allow this functionality, both the mustache and json output handlers must be registered.

Any number of handlers can be registered to an interface. You might have a use case for an Interface/Handler that may provide different compatibility based on the operating system, or perhaps based on simply how the application is called. A good example would be an application that automates building packages for Linux distributions. An interface would define what a build handler needs to provide, but the build handler would be different based on the OS. The application might have an rpm build handler, or a dpkg build handler to perform the build process differently.

Customizing Handler Configuration and Meta

Depending on the handler, you will have different ways of customizing its functionality. Some handlers honor application configuration setting, while others may only rely on meta-options. In either case, both can be modified at the top level of your application meta.

In the following example, we modify the configuration defaults of the log handler, and also the meta-options to enable an optional log level command line argument feature it supports.

Note that configuration settings are always overridable via configuration files, however this example is modifying the builtin default setting of the log handler.

from cement import App, init_defaults

CONFIG = init_defaults('myapp', 'log.logging')
CONFIG['log.logging']['level'] = 'warning'

META = init_defaults('log.logging')
META['log.logging']['log_level_argument'] = ['-l', '--level']

class MyApp(App):
    class Meta:
        label = 'myapp'
        config_defaults = CONFIG
        meta_defaults = META

If modifying the configuration or meta options isn't enough, you can always sub-class an existing handler and register your own in its place.

Overriding Handlers via Command Line

In some use cases, you will want the end user to have access to override the default handler of a particular interface. For example, Cement ships with multiple Output Handlers including json, yaml, and mustache. A typical application might default to using mustache to render console output from text templates. That said, without changing any code in the application, the end user could simply pass the -o json command line option and output the same data that is rendered to template, out in pure JSON.

Output hander overrides are not enabled by default, but can be enabled by setting the OutputHandler.Meta.overridable option to True for the output handlers that you want overridable by the -o option at command line. See the documentation on Output Rendering for more details and examples.

The only built-in handler override that Cement includes is for the above mentioned JSON example, but you can add any that your application requires.

myapp.py
from cement import App, init_defaults

META = init_defaults('output.json', 'output.yaml')
META['output.json']['overridable'] = True
META['output.yaml']['overridable'] = True

class MyApp(App):
    class Meta:
        label = 'myapp'
        extensions = ['json', 'yaml', 'mustache']
        meta_defaults = META
        output_handler = 'mustache'
        template_dir = './templates'

with MyApp() as app:
    app.run()

    data = {'foo': 'bar'}
    app.render(data, 'example.m')
templates/example.m
Foo: {{ foo }}

Notice the -o command line option, that includes the choices: yaml and json. This feature will include all Output Handlers that have the overridable meta-data option set to True. We did not set this option for the mustache handler, therefore it did not show up as a choice for the -o option at command-line.

PreviousCore FoundationNextHooks

Last updated