Skip to main content

Asyncio plugins, components, dependency injection and configs

Project description

This is heavily inspired by Pyramid and my daily needs to fastly create and maintain microservice like applications.

a plugin mechanic

  • plugin may depend on other plugins
  • plugins yield tasks to run
  • a registry serves as a store for application components created by plugins
  • a dependency injection creates intermediate components
  • a config source is mapped to plugin specific needs
  • structlog boilerplate for json/tty logging

You bootstrap like following:

from buvar import plugin

plugin.stage("some.module.with.plugin.function")
# some.module.with.plugin.function
from buvar import context, plugin

class Foo:
    ...


async def task():
    asyncio.sleep(1)


async def server():
    my_component = context.get(Foo)
    await asyncio.Future()


# you may omit include in arguments
async def prepare(load: plugin.Loader):
    await load('.another.plugin')

    # create some long lasting components
    my_component = context.add(Foo())

    # you may run simple tasks
    yield task()

    # you may run server tasks
    yield server()

a components and dependency injection solution

Dependency injection relies on registered adapters, which may be a function, a method, a class, a classmethod or a generic classmthod.

Dependencies are looked up in components or may be provided via, arguments.

from buvar import di

class Bar:
    pass

class Foo:
    def __init__(self, bar: Bar = None):
        self.bar = bar

    @classmethod
    async def adapt(cls, baz: str) -> Foo:
        return Foo()

async def adapt(bar: Bar) -> Foo
    foo = Foo(bar)
    return foo


async def task():
    foo = await di.nject(Foo, baz="baz")
    assert foo.bar is None

    bar = Bar()
    foo = await di.nject(Foo, bar=bar)
    assert foo.bar is bar

async def prepare():
    di.register(Foo.adapt)
    di.register(adapt)

a config source

buvar.config.ConfigSource is just a dict, which merges arbitrary dicts into one. It serves a the single source of truth for application variability.

You can load a section of config values into your custom attrs class instance. ConfigSource will override values by environment variables if present.

config.toml

log_level = "DEBUG"
show_warnings = "yes"

[foobar]
some = "value"
export APP_FOOBAR_SOME=thing
import attr
import toml

from buvar import config

@attr.s(auto_attribs=True)
class GeneralConfig:
    log_level: str = "INFO"
    show_warnings: bool = config.bool_var(False)


@attr.s(auto_attribs=True)
class FoobarConfig:
   some: str


source = config.ConfigSource(toml.load('config.toml'), env_prefix="APP")

general_config = source.load(GeneralConfig)
assert general_config == GeneralConfig(log_level="DEBUG", show_warnings=True)

foobar_config = source.load(FoobarConfig, 'foobar')
assert foobar_config.some == "thing"

There is a shortcut to the above approach provided by buvar.config.Config, which requires to be subclassed from it with a distinct section attribute. If one adds a buvar.config.ConfigSource component, he will receive the mapped config in one call.

from buvar import config, plugin


@attr.s(auto_attribs=True)
class GeneralConfig(config.Config):
    log_level: str = "INFO"
    show_warnings: bool = config.bool_var(False)


@attr.s(auto_attribs=True)
class FoobarConfig(config.Config, section="foobar"):
    some: str


async def prepare(load: plugin.Loader):
    # this would by typically placed in the main entry point
    source = context.add(config.ConfigSource(toml.load('config.toml'), env_prefix="APP"))

    # to provide the adapter to di, which could also be done inthe main entry point
    await load(config)
    foobar_config = await di.nject(FoobarConfig)

a structlog

Just structlog boilerplate.

import sys

from buvar import log

log.setup_logging(sys.stdout.isatty(), general_config.log_level)

Project details


Download files

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

Files for buvar, version 0.29.0
Filename, size File type Python version Upload date Hashes
Filename, size buvar-0.29.0-cp37-cp37m-manylinux1_x86_64.whl (376.3 kB) File type Wheel Python version cp37 Upload date Hashes View
Filename, size buvar-0.29.0-cp38-cp38-manylinux1_x86_64.whl (438.8 kB) File type Wheel Python version cp38 Upload date Hashes View
Filename, size buvar-0.29.0.tar.gz (129.6 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page