- Websocket example
- GraphQL endpoint
- Asynchronous SQLAlchemy usage instead of raw connection.
This is the entry point of the application and is also the highest level of orchestration for your
API Star app. This is where we import the libraries and app
specific modules that we'll use to construct the App instance.
In this application, we have several custom components we need to register with the App and
our settings module for use by the App. We also setup a basic root logger here. If debug/dev mode
is detected then we'll log the app specific routes, settings and components at start up.
The components we build and use in this app are:
- aioclient for performing HTTP client call asynchronously
This is fairly simple, mostly it just maps environmental values into our settings.
We add and override some default settings for template searching and override the
RENDERERS setting to use our own src/local_utils/renderer.py for
our endpoints.
These backends are API Star Components and it's important to note that components are registered and preloaded(initialized) before the App's IOLoop is started. This will cause problems if you try get an IOLoop and create connections on component Initialization. For instance, if the Redis component gets the IOLoop and uses it to create a connection pool at initialization. It's connection pool will be on a different IOLoop than what the Redis client calls will come from, which is an error. This is because components are register and preloaded(by default) before the App's IOLoop is created and started. Because of this, components that need connection pools or connections to other services make those connections lazily as needed by client requests. This is not as elegant as beging able to initialize components after the App's IOLoop is created and then have access to that IOLoop, but is necessary for now.
The redis module defines a Component and a Command
The command available is redis_cli and will open the Redis CLI using the App's settings.
$ ./app.py redis_cli
[::1]:6379>The asyncpg module defines a Component and a Command
The command available is dbshell and will run psql with the App's connection settings.
$ ./app.py dbshell
psql (10.1)
Type "help" for help.
cryptoprice=#