sbreatnach / fabpowertasks Goto Github PK

Offers an alternative to the standard way of generating Fabric tasks. Instead of random functions or Task declarations, instead create Commands classes that use a simple configuration flow and allow for dependency injection to share functionality between Tasks.

Additionally, some simple tasks are auto-generated and act as examples of how to write Commands.

Here's a simple fabfile.py:

import fabpowertasks
project_dir = os.path.dirname(__file__)
fabpowertasks.initialize(globals(), project_dir)

This automatically generates the Fabric tasks that come with fabpowertasks. They can be run as standard using Fabric with fab <command>

But that's not all - here's a more complex example fabfile/__init__.py that includes custom tasks:

import os
import fabpowertasks
from fabfile.config import ProjectPaths
from fabfile.tasks import CustomCommands

fabfile_dir = os.path.dirname(__file__)
project_dir = os.path.dirname(fabfile_dir)
properties = fabpowertasks.Properties.load_from_file(
    os.path.join(fabfile_dir, 'project.json')
)
properties.base_dir = project_dir
local_project_paths = ProjectPaths(project_dir, properties)
injection_objs = {
    'local_project_paths': local_project_paths,
    'remote_project_paths': fabpowertasks.SystemPaths(properties.remote_base_dir,
                                                      properties),
    'custom_commands': CustomCommands()
}
fabpowertasks.initialize(globals(), os.path.dirname(fabfile_dir),
                         project_properties=properties,
                         project_inject_objs=injection_objs)

Some explanation of this example is needed:

• CustomCommands is a class that extends the utility class fabpowertasks.commands.BaseCommands. This class comes with some additional functions that are used to generate Fabric tasks from itself.
• An instance of the Properties class is used extensively throughout the library and contains all project properties. At the very least, it requires a project directory location, which is why this is a required option to initialize(). But the default instance may be overridden, as shown here.
• fabpowertasks inspects the variables for each Commands instance and attempts to inject any matching values as described in injection_objs. Any variables that don't start with _ and are None as treated as potential injection points. By default, an attempt to inject local_project_paths into every Commands instance is made but more injections can be added using the project_injection_objs argument.

The concept of Commands is simple - they are an advanced form of a Fabric Task. Instead of being limited to the run() function, multiple functions of the Command instance can be treated as Tasks.

• Extend fabpowertasks.commands.BaseCommands
• Write the functions for your resulting Task
• __init__(): Update self._task_functions with the functions
• __init__(): Optionally, set instance variables for injection
• Import your new class into fabfile/__init__.py and pass into the project_injection_objs dict where the resulting Tasks are auto-generated

Note that arguments to the function mirror what a standard Fabric task expects.