command-system

Parts of this readme do not render correctly on pypi, see the repository for a better experience

Command System

A type-safe implementation of the command pattern in Python, designed to handle command execution with a lifecycle that includes deferral, cancellation, and dependency management.

With this package, you can create commands that can be queued and manage their own lifecycle, allowing for interacting with external systems, or building your own systems with complex interactions in a safe, clean, and maintainable way.

Example Usage

Simply subclass Command, CommandArgs, and optionally CommandResponse to create a custom command for your use case.

# SayHelloCommand.py
from command_system import (
    Command,
    CommandArgs,
    CommandResponse,
    ExecutionResponse,
)
from dataclasses import dataclass

@dataclass
class SayHelloArgs(CommandArgs):
    name: str

@dataclass
class SayHelloResponse(CommandResponse):
    message: str = ""

class SayHelloCommand(Command[SayHelloArgs, SayHelloResponse]):
    ARGS = SayHelloArgs
    _response_type = SayHelloResponse

    def execute(self) -> ExecutionResponse:
        if not self.args.name:
            return ExecutionResponse.failure("Name cannot be empty.")
        self.response.message = f"Hello, {self.args.name}!"
        return ExecutionResponse.success()

You can then use the CommandQueue to submit commands, and then queue.process_once() or queue.process_all() to execute them.

# main.py
from command_system import CommandQueue
from SayHelloCommand import SayHelloCommand, SayHelloArgs

queue = CommandQueue()
response = queue.submit(SayHelloCommand(SayHelloArgs(name="Alice")))
print(response.status) # Pending
queue.process_once()
print(response.status) # Completed
print(response.message) # Hello, Alice!

Command Lifecycle

flowchart TD
    A[create command] -->|"queue.submit(command)"| B[responseStatus.CREATED]
    B --> C{"command.should_defer()"}
    C -->|"DeferResponse.defer()"| W["responseStatus.PENDING"]
    W --> C
    C -->|"DeferResponse.proceed()"| D{"command.should_cancel()"}
    D -->|"CancelResponse.cancel()"| E[ResponseStatus.CANCELED]
    D -->|"CancelResponse.proceed()"| F["command.execute()"]
    F -->|"ExecutionResponse.success()"| G["ResponseStatus.COMPLETED"]
    F -->|"ExecutionResponse.failure()"| H["ResponseStatus.FAILED"]

Chaining Commands

You can chain commands together using the CommandChainBuilder. By doing this, you can create a sequence of commands that will be executed in order, passing the output of one command as the input to the next (with customizable transformations).

from 
from my_command import MyCommand, MyCommandArgs
from my_other_command import MyOtherCommand, MyOtherCommandArgs

queue = CommandQueue()
chain = CommandChainBuilder[str, str].start(
    "my initial input",
    args_factory=lambda x: MyCommandArgs(input_data=x),
    command_class=MyCommand,
    result_extractor=lambda response: response.output_data,
).then(
    args_factory=lambda x: MyOtherCommandArgs(input_data=x, other_data="another value"),
    command_class=MyOtherCommand,
    result_extractor=lambda response: response.output_data,
).build(queue)

queue.submit(chain)
queue.process_all()

Dependency Management

Commands can now define dependencies on other commands. Dependencies are evaluated before each lifecycle check begins, and they can preemptively defer or cancel the command based on their statuses.

Dependencies only have an effect when using the CommandQueue to process commands. If you write your own command processing logic, you will need to handle dependencies manually.

It is your responsibility to ensure that dependencies do not create circular references that will cause an infinite loop.

Adding Dependencies

You can add dependencies to a command by passing them during initialization or using the add_dependency method. Dependencies can be other commands or wrapped in DependencyEntry for more control.

from command_system import Command, CommandArgs, DependencyEntry
from MyCommand import MyCommand

# Example usage
dependency_command = MyCommand(MyCommand.ARGS())
main_command = MyCommand(MyCommand.ARGS(), dependencies=[dependency_command])

Writing Dependency Rules

When defining dependencies, you can specify actions for different statuses of the dependent command:

• on_pending: What to do if the dependency is pending/created (default: "defer").
• on_canceled: What to do if the dependency is canceled (default: "cancel").
• on_failed: What to do if the dependency fails (default: "cancel").
• on_completed: What to do if the dependency is completed (default: "proceed").

Example

from command_system import CommandQueue, DependencyEntry
from MyCommand import MyCommand

queue = CommandQueue()
dependency_command = MyCommand(MyCommand.ARGS(...))
main_command = MyCommand(MyCommand.ARGS(...), dependencies=[DependencyEntry(dependency_command, on_pending="cancel")])

queue.submit(dependency_command)
queue.submit(main_command)
queue.process_once()

In this example, main_command will cancel if dependency_command is still pending.

Creating a command

Subclass CommandArgs and add any arguments your command needs. This class will be used to pass parameters to your command.

[!CAUTION] Don't add your arguments directly to the Command class. The args class is required for command chaining to work in a type-safe manner.

Subclass Command and set the ARGS class attribute to the subclass of CommandArgs you created. Implement the execute method to define the command's behavior.

• You can also override should_defer and should_cancel methods to control the command's lifecycle.

Optionally, you can create a custom response class by subclassing CommandResponse so that your command can return specific type-safe data. If you do this, set the _response_type class attribute of your Command subclass to your custom response class.

Writing your execute method

[!WARNING]
Your execute method should not return your custom response class directly. The self.response attribute is automatically set to an instance of your custom response class, which you should modify instead. Then, return an ExecutionResponse instance to indicate the command's success or failure.

Writing should_defer and should_cancel methods

These methods can be overridden to control the command's lifecycle. They must return a DeferResponse or CancelResponse instance, respectively. You can use them to set conditions for deferring or canceling the command.

Complex Command example

For an example of deferring and canceling commands, see the tests/test_defer_cancel.py file.