feat: Planar Clarity Video Wall

[GabFitzgerald]

First attempt
Resolves #59

some confusion with options / **options

[kimburgess]

Having a quick read now. To provide some further reading on **options, that's a use of splatting.

TL;DR it can be used as a continent way to forward an arbitrary set of keyword arguments around. It was used a lot in the previous Ruby drivers, but would recommend avoiding it unless absolutely needed here as it can lead to some sloppy type signatures.

[kimburgess]

When writing specs, you also want to mock mock the device responses so that you test full request -> response behaviour. For some info on this, see: https://github.com/PlaceOS/drivers/blob/master/docs/writing-a-spec.md#testing-streaming-io.

[dukenguyenxyz]

Having a quick read now. To provide some further reading on **options, that's a use of splatting.

TL;DR it can be used as a continent way to forward an arbitrary set of keyword arguments around. It was used a lot in the previous Ruby drivers, but would recommend avoiding it unless absolutely needed here as it can lead to some sloppy type signatures.

Hi @kimburgess , upon going through this comment I thought it'd be a good idea to implement Hash instead of splat. However, this is not allowed since splatting in drivers is mostly used in

queue(**({
      name: command,
    }.merge(options))) do
...
end

and queue's method is defined in placeos-driver/src/placeos-driver.cr as follows (line 114-117)

  # Queuing
  def queue(**opts, &block : Task -> Nil)
    @__queue__.add(**opts, &block)
  end

which disallows anything but splats and NamedTuple in the argument.

[pkheav]

I've had a look through and have picked up some things on style/feedback that @kimburgess has given me previously. There are probably other things I have missed or not knowledgable enough to give feedback on so best to reach out to @kimburgess for further feedback but this should hopefully be a good start of some things to look out for :)

[pkheav]

self[:power] will be of type JSON::Any so this should be changed to:
input_status if self[:power]?&.try as_bool

I've noticed this is necessary in order to use self values that are of type Bool properly in conditional statements.

Explanation of why this is needed is here #19 (comment) and is probably to do with how crystal deals with truthy/falsy values

[pkheav]

self[:key] will return an error if the key does not exist while self[:key]? will return nil so is safer to use generally.

.try is a safe way to deal with objects that may be nil https://crystal-lang.org/api/0.35.1/Object.html#try(&)-instance-method

as_bool https://crystal-lang.org/api/0.19.4/JSON/Any.html#as_bool%3ABool-instance-method is necessary in order to convert self[:power] which is of type JSON::Any to type Bool so it behaves as expected in conditional statements

[pkheav]

I think this should be renamed to input? to keep it consistent with power? but @kimburgess can probably better comment on this