qsys-qrc-client

A type-safe control client for QSC Q-SYS cores, written in Node.js.

Install

$ npm install qsys-qrc-client zen-observable zod

Zod is technically an optional peer dependency, but it is required for type hints. If you don't want the Zod Validator in production, you should at least install it as a dev dependency.

You can replace zen-observable with whichever Observable implementation you want. See any-observable.

Usage

import QrcClient, {commands, ZodValidator} from 'qsys-qrc-client';

const client = new QrcClient({validator: new ZodValidator()});
client.connect({
  port: 1710,
  host: '192.168.1.10'
});

async function checkStatus() {
  // The send function returns Promises, use the await syntax for synchronous-like code flow.
  const status = await client.send(commands.getStatus());
  console.log(`${status.DesignName} is running on a ${status.Platform}`);
  //=> "My Test Design is running on a Emulator"
}

API

client.send(command)

command

Type: Any valid QRC command.

You should generate most commands using the built-in commands object, as described below.

Note: It is not necessary to insert the "jsonrpc": "2.0" version number in your command. The client will insert it automatically if it is not present.

returns

Type: Promise<Object>

See the documentation for specific return types.

commands.getStatus() / commands.login(username, password) / etc.

The commands object has a number of helper functions to create specific QRC commands. They include type information for the parameters, and the expected return values. Modern IDE's should provide type hints out of the box.

For this reason, it is always recommended you use the commands object to generate the command objects. But you can create commands by hand if needed (i.e. a new command is not yet implemented here).

validation

The library includes two different validators for validating both commands and responses:

• A singleton noopValidator instance that bypasses any validation (relying on errors returned from the Q-Sys core to identify mistakes). This is the default as it removes the runtime dependency on zod. (You will still need zod as a dev dependency for type inference though).
• A ZodValidator that is capable of validating both commands and responses. It can be configured for loose or strict validation in either direction, and it can be configured to throw or just log the error. See the ParseOptions export.

Note: It's unlikely you want strict mode for response parsing on the ZodValidator. That mode is used during development so our tests throw errors if our validation types don't match the response.

client.pollGroup(groupId, options)

A convenience class for sending poll group commands. It implements observable, so you can just listen to changes. You must send the autoPoll command for change events to occur, however.

const group = client.pollGroup('groupId');
group.addControl('controlName1', 'controlName2');

group.subscribe(data => {
   for (const change of data.Changes) {
     if(change.Name === 'controlName1') {
       // do something with change.Value / change.String / change.Position
     } else if (change.Name === 'controlName2') {
       // do something with change.Value / change.String / change.Position
     }
   }
});

// Start receiving events, updated 10x a second:
group.autoPoll(0.1);

groupId

Type: string

The groupId of the group of values you want to change. It should already have been setup.

options.rate

Type: number Default: 0.2

The maximum rate (in seconds) at which to send updates.

options.autoDestroy

Type: boolean Default: false

If true, then when the last observer unsubscribes from the Observable, the polling group will be automatically destroyed. There are perils with either choice. If you enable autoDestroy, then you must rebuild the polling group to observe those values again, if you do not, then you must remember to destroy the group yourself with client.send(commands.destroyGroup('groupId'))

returns

Type: Observable<AutoPollUpdate>

Observables are a specification for observing changing data. You can find more information at the following resources:

• T39 Spec
• RxJS A popular, full-featured implementation.
• zen-observable A lightweight implementation.
• any-observable Does not provide an actual observable implementation. Use it to specify which Observable implementation you want to use (if zen or RxJs are installed, they will be found automatically).

Convenience Imports

Rather than importing the entire commands object, you can import just the specific commands you need, and use a slightly less verbose syntax:

import QrcClient, {getStatus} from 'qsys-qrc-client';

// ...

client.send(getStatus());

The typescript declarations also export a handful of useful type definitions (used to provide type hints in your IDE), that you can import and reuse.

import QrcClient, {commands, EngineStatus} from 'qsys-qrc-client';

let status: EngineStatus;

// ...

status = await client.send(commands.getStatus());

Contributing

Bug Reports and Pull Requests are always welcome. To get started.

1. Install Node.js and git
2. Clone the repository.
3. Run npm install in the cloned directory to pull down all the dependencies.
4. Run npm test to test any changes.

A very simple design file is included in the integration-tests folder, and it is used to test the client against an actual design running in Emulator. It defaults to connecting to localhost but you can change it by putting the connection information in integration-tests/design-location.json. This can be useful if developing on a non-Windows machine.

License

MIT © James Talmage