Skip to main content

esper is a lightweight Entity System (ECS) for Python, with a focus on performance.

Project description

[![pypi](https://badge.fury.io/py/esper.svg)](https://pypi-hypernode.com/pypi/esper) [![rtd](https://readthedocs.org/projects/esper/badge/?version=latest)](https://esper.readthedocs.io) [![PyTest](https://github.com/benmoran56/esper/actions/workflows/unit-tests.yml/badge.svg)](https://github.com/benmoran56/esper/actions/workflows/unit-tests.yml)

Esper is a lightweight Entity System module for Python, with a focus on performance

Esper is an MIT licensed Entity System, or, Entity Component System (ECS). The design is based on the Entity System concepts outlined by Adam Martin in his blog at http://t-machine.org/, and others. The primary focus is on keeping it as lightweight and performant as possible, while handling common use cases.

The following Wikipedia article provides a summary of the ECS pattern: https://en.wikipedia.org/wiki/Entity_component_system

API documentation is hosted at ReadTheDocs: https://esper.readthedocs.io Due to the small size of the project, this README currently serves as general usage documentation.

What’s new: As of Esper 2.0, there is support for basic event dispatching and handling. This is fairly minimal, in keeping with the scope of the project, but it should be robust enough to handle most common needs.

See also: Another Python ECS library https://github.com/ikvk/ecs_pattern

  • [Compatibility](#compatibility)

  • [Installation](#installation)

  • [Design](#design)

  • [Quick Start](#quick-start)

  • [General Usage](#general-usage) * [Adding and Removing Processors](#adding-and-removing-processors) * [Adding and Removing Components](#adding-and-removing-components) * [Querying Specific Components](#querying-specific-components) * [Boolean and Conditional Checks](#boolean-and-conditional-checks) * [More Examples](#more-examples)

  • [Event Dispatching](#event-dispatching)

  • [Contributing](#contributing)

Compatibility

Esper attempts to target all currently supported Python releases (not EOL). Because Esper is written in pure Python, any compliant interpreter should work, such as Pypy3. Continuous Integration (automated testing) is currently done for both CPython and PyPy3.

Installation

No installation is necessary. Esper is a Python package with no dependencies or extensions. Simply copy the esper folder into your project folder, and import esper.

If you prefer, Esper is also available on PyPI for easy installation via pip:

pip install --user --upgrade esper

Design

  • World

A World is the main point of interaction in Esper. After creating a World object, you will use that object to create Entities and assigning Components to them. A World is also assigned all of your Processor instances, and handles smoothly running everything with a single call per frame. Of course, Entities, Components and Processors can be created and assigned, or deleted while your application is running.

  • Entities

Entities are simple integer IDs (1, 2, 3, 4, etc.). Entities are “created”, but they are generally not used directly. Instead, they are simply used as IDs in the internal Component database to track collections of Components. Creating an Entity is done with the World.create_entity() method.

  • Components

Components are defined as simple Python classes. In keeping with a pure Entity System design philosophy, they should not contain any logic. They might have initialization code, but no processing logic whatsoever. A simple Component can be defined as:

class Position:
    def __init__(self, x=0.0, y=0.0):
        self.x = x
        self.y = y

In addition, the excellent dataclass decorator is available in Python 3.7+. https://docs.python.org/3/library/dataclasses.html#module-dataclasses This decorator simplifies defining your Component classes. The attribute names don’t need to be repeated, and you can still instantiate the Component with positional or keyword arguments:

from dataclasses import dataclass as component

@component
class Position:
    x: float = 0.0
    y: float = 0.0
  • Processors

Processors, also commonly known as “Systems”, are where all processing logic is defined and executed. All Processors must inherit from the esper.Processor class, and have a method called process. Other than that, there are no restrictions. All Processors will have access to the World instance, which is how you query Components to operate on. A simple Processor might look like:

class MovementProcessor(esper.Processor):

    def process(self):
        for ent, (vel, pos) in self.world.get_components(Velocity, Position):
            pos.x += vel.x
            pos.y += vel.y

In the above code, you can see the standard usage of the World.get_components() method. This method allows efficient iteration over all Entities that contain the specified Component types. This method can be used for querying two or more components at once. Note that tuple unpacking is necessary for the return component pairs: (vel, pos). In addition the Components, you also get a reference to the Entity ID (the ent object) for the current pair of Velocity/Position Components. This entity ID can be useful in a variety of cases. For example, if your Processor will need to delete certain Entites, you can call the self.world.delete_entity() method on this Entity ID. Another common use is if you wish to add or remove a Component on this Entity as a result of some condition being met.

Quick Start

The first step after importing Esper is to create a World instance. You can have a single World instance for your entire game, or you can have a separate instance for each of your game scenes. Whatever makes sense for your design. Create a World instance like this:

world = esper.World()

Create some Processor instances, and assign them to the World. You can specify an optional processing priority (higher numbers are processed first). All Processors are priority “0” by default:

movement_processor = MovementProcessor()
collision_processor = CollisionProcessor()
rendering_processor = RenderingProcessor()
world.add_processor(movement_processor, priority=2)
world.add_processor(collision_processor, priority=3)
world.add_processor(rendering_processor)
# or just add them in one line:
world.add_processor(SomeProcessor())

Create an Entity, and assign some Component instances to it:

player = world.create_entity()
world.add_component(player, Velocity(x=0.9, y=1.2))
world.add_component(player, Position(x=5, y=5))

Optionally, Component instances can be assigned directly to the Entity on creation:

player = world.create_entity(Velocity(x=0.9, y=1.2), Position(x=5, y=5))

Executing all Processors is done with a single call to world.process(). This will call the process method on all assigned Processors, in order of their priority. This is usually called once per frame update of your game.:

world.process()

Note: You can pass any args you need to world.process(), but you must also make sure to receive them properly in the process() methods of your Processors. For example, if you pass a delta time argument as world.process(dt), your Processor’s process() methods should all receive it as: def process(self, dt): This is appropriate for libraries such as pyglet, which automatically pass a delta time value into scheduled methods.

General Usage

Adding and Removing Processors

You have already seen examples of adding Processors in an earlier section. There is also a remove_processor method available:

  • World.add_processor(processor_instance)

  • World.remove_processor(ProcessorClass)

Depending on the structure of your game, you may want to add or remove certain Processors when changing scenes, etc.

Adding and Removing Components

In addition to adding Components to Entities when you’re creating them, it’s a common pattern to add or remove Components inside of your Processors. The following methods are available for this purpose:

  • World.add_component(entity_id, component_instance)

  • World.remove_component(entity_id, ComponentClass)

As an example of this, you could have a “Blink” component with a duration attribute. This can be used to make certain things blink for s specific period of time, then disappear. For example, the code below shows a simplified case of adding this Component to an Entity when it takes damage in one processor. A dedicated BlinkProcessor handles the effect, and then removes the Component after the duration expires:

class BlinkComponent:
    def __init__(self, duration):
        self.duration = duration


.....


class CollisionProcessor(esper.Processor):

    def process(self, dt):
        for ent, enemy in self.world.get_component(Enemy):
            ...
            is_damaged = self._some_method()
            if is_damaged:
                self.world.add_component(ent, BlinkComponent(duration=1))
            ...


class BlinkProcessor(esper.Processor):

    def process(self, dt):
        for ent, (rend, blink) in self.world.get_components(Renderable, BlinkComponent):
            if blink.duration < 0:
                # Times up. Remove the Component:
                rend.sprite.visible = True
                self.world.remove_component(ent, BlinkComponent)
            else:
                blink.duration -= dt
                # Toggle between visible and not visible each frame:
                rend.sprite.visible = not rend.sprite.visible

Querying Specific Components

If you have an Entity ID and wish to query one specific, or ALL Components that are assigned to it, the following methods are available:

  • World.component_for_entity

  • World.components_for_entity

The component_for_entity method is useful in a limited number of cases where you know a specific Entity ID, and wish to get a specific Component for it. An error is raised if the Component does not exist for the Entity ID, so it may be more useful when combined with the has_component method that is explained in the next section. For example:

if self.world.has_component(ent, SFX):
    sfx = self.world.component_for_entity(ent, SFX)
    sfx.play()

The components_for_entity method is a special method that returns ALL of the Components that are assigned to a specific Entity, as a tuple. This is a heavy operation, and not something you would want to do each frame or inside of your Processor.process method. It can be useful, however, if you wanted to transfer all of a specific Entity’s Components between two separate World instances (such as when changing Scenes, or Levels). For example:

player_components = old_world.components_for_entity(player_entity_id)
...
player_entity_id = new_world.create_entity(player_components)

Boolean and Conditional Checks

In some cases you may wish to check if an Entity has a specific Component before performing some action. The following methods are available for this task:

  • World.has_component(entity, ComponentType)

  • World.has_components(entity, ComponentTypeA, ComponentTypeB)

  • World.try_component(entity, ComponentType)

  • World.try_components(entity, ComponentTypeA, ComponentTypeB)

For example, you may want projectiles (and only projectiles) to disappear when hitting a wall in your game. We can do this by checking if the Entity has a Projectile Component. We don’t want to do anything to this Component, simply check if it’s there. Consider this example:

class CollisionProcessor(esper.Processor):

    def process(self, dt):
        for ent, body in self.world.get_component(PhysicsBody):
            ...
            colliding_with_wall = self._some_method(body):
            if colliding_with_wall and self.world.has_component(ent, Projectile):
                self.world.delete_entity(ent)
            ...

In a different scenario, we may want to perform some action on an Entity’s Component, if it has one. For example, a MovementProcessor that skips over Entities that have a Stun Component:

class MovementProcessor(esper.Processor):

    def process(self, dt):
        for ent, (body, vel) in self.world.get_components(PhysicsBody, Velocity):

            if self.world.has_component(ent, Stun):
                                stun = self.world.component_for_entity(ent, Stun)
                                stun.duration -= dt
                                    if stun.duration <= 0:
                                        self.world.remove_component(ent, Stun)
                                return      # Return without processing movement

                            movement_code_here()
            ...

Lets look at the core part of the code:

if self.world.has_component(ent, Stun):
    stun = self.world.component_for_entity(ent, Stun)
    stun.duration -= dt

This code works fine, but the try_component method can accomplish the same thing with one less call to World. The following example will get a specific Component if it exists, or return None if it does not:

stun = self.world.try_component(ent, Stun)
if stun:
    stun.duration -= dt

With Python 3.8+, the new “walrus” operator (:=) can also be used, making the try_component methods even more concise

if stun :=  self.world.try_component(ent, Stun):
    stun.duration -= dt

More Examples

See the /examples folder to get an idea of how a basic structure of a game might look.

Event Dispatching

Esper includes basic support for event dispatching and handling. This functionality is provided by three functions, exposed directly on the esper module. Events are not tied to a specific World instance. Minimal error checking is done, so it’s left up to the user to ensure correct naming and number of arguments.

Events are dispatched by name:

esper.dispatch_event('event_name', arg1, arg2)

In order to receive the above event, you must register handlers. An event handler can be a function or class method. Registering a handler is also done by name:

esper.set_handler('event_name', my_func)
# or
esper.set_handler('event_name', self.my_method)

Note: Only weak-references are kept to the registered handlers. If a handler is garbage collected, it will be automatically un-registered by an internal callback.

Handlers can also be removed at any time, if you no longer want them to receive events:

esper.remove_handler('event_name', my_func)
# or
esper.remove_handler('event_name', self.my_method)

Contributing

Contributions to Esper are always welcome, but there are some specific project goals to keep in mind:

  • Pure Python code only - no binary extensions, Cython, etc.

  • Try to target all currently supported Python versions. Exceptions can be made if there is a compelling reason.

  • Avoid bloat as much as possible. New features will be considered if they are commonly useful. Generally speaking, we don’t want to add functionality that is better handled in another module or library.

  • Performance is preferrable to readability.

If you have any questions before contributing, feel free to [open an issue].

[open an issue]: https://github.com/benmoran56/esper/issues

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

esper-2.5.zip (47.9 kB view details)

Uploaded Source

Built Distribution

esper-2.5-py3-none-any.whl (12.0 kB view details)

Uploaded Python 3

File details

Details for the file esper-2.5.zip.

File metadata

  • Download URL: esper-2.5.zip
  • Upload date:
  • Size: 47.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.3

File hashes

Hashes for esper-2.5.zip
Algorithm Hash digest
SHA256 2ec1e06ed18d1ee194074be4a1783cce29083e8a9a21049eaa24a9d11f5b73fd
MD5 ca5339a634f3539e169d849215e012b0
BLAKE2b-256 50e82abe41591bab80ceddb179cdefe41ca53ae230a18ad6dfd21ce289f4aae5

See more details on using hashes here.

File details

Details for the file esper-2.5-py3-none-any.whl.

File metadata

  • Download URL: esper-2.5-py3-none-any.whl
  • Upload date:
  • Size: 12.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.3

File hashes

Hashes for esper-2.5-py3-none-any.whl
Algorithm Hash digest
SHA256 e1046bab41f21289848728317678d4f56979e203b1df13a94859eaaa8fba2a2a
MD5 5e9b2a5abe917f59de9c91d99e2cdf98
BLAKE2b-256 465857f7b8806e223c0955d0ba2a2dbdb99b8e0cb6b3b0d69ce243fc6dca52d9

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page