Skip to content

Tenkiv/kuantify

Repository files navigation

Kuantify  License Matrix

Kuantify is usable in its current state but it's in a very early pre-alpha stage of development. Some things will not work and you should expect relatively frequent breaking changes.

Kuantify is a physical computing / data acquisition and control library written in Kotlin. Its purpose is to provide a common set of interfaces and abstractions for utilizing arbitrary collection and control endpoints in a safe, reasonable, and performant way. Kuantify is focused on enabling rapid prototyping of physical computing programs and a smooth transition from prototypes to production.

Kuantify currently only works on the JVM and Android but it is a core design principle for it to ultimately be multiplatform and multiplatform support will be added as the api starts to stabilize (the main thing we need to do is rewrite Physikal in pure Kotlin so it's usable on all Kotlin targets).

We also plan to create another repository where we centralize support for as many DAQC devices as possible but again, we actually want to keep the number of supported devices to a minimum until we stabilize the api so we don't end up with a maintenance nightmare. For now we only support the sensors and physical controls build into Android. Although Android phone sensors are not really a big intended use case for this library, they provide a great accessible way to start getting a feel for Kuantify and you can actually throw together some fairly interesting demos.

Artifacts

Core Android Core Learning
JVM Maven Central Maven Central Maven Central

Local (for adapting OS support for device sensors and controllers to Kuantify):

  • Android Local - Maven Central

Basics

First of all, in the name of type safety, Kuantify never passes around data as raw numbers. We use DaqcValue which is either a BinaryState (high/low) or a Quantity. A Quantity is a type safe dimensioned number (a number with a unit), Quantity support is done in a separate library called Physikal. You should probably play with Physikal a bit before attempting to use Kuantify.

The three most important types to understand in Kuantify are Input, Output, and Device. Here's a super bare bones usage example with an Android device.

val device = LocalAndroidDevice.get(this)
val lightSensor = device.lightSensors.first()

lightSensor.startSampling()
lightSensor.updateBroadcaster.consumeEach { measurement ->
    // Do something with every measurement
}

This example assumes your program is actually running in Android.

val device = LocalAndroidDevice.get(this)

Gets the device it's running on as a Kuantify device. Since different Android devices have different sensors and controllers available each type of sensor or controller is given as a list that may be empty if the device doesn't support a given type.

lightSensor.startSampling()

Starts sampling the light sensor; you won't receive any updates from it until you start it.

lightSensor.updateBroadcaster.consumeEach

Each Input has a BroadcastChannel called updateBroadcaster that broadcasts updates from that Input. Here we're just consuming all updates.

Using an Output is very similar.

val device = LocalAndroidDevice.get(this)
val torchController = device.torchControllers.first()

torchController.setOutput(BinaryState.High)

In the above examples we're using Kuantify on the device with the sensors and controllers. But Kuantify is primarily intended to be used on a device that is connected to other devices which actually have the sensors and controllers. So let's change our example to now run on a computer that is connected to an Android device. The Android device is now a RemoteDevice. Note: to do this we need to run a host utility on the Android Device we're connecting to. This utility is trivial to make but you can just install android-simple-host to get started with remote connections. More information in the "Supporting Devices" section.


val device = RemoteAndroidDeivce("192.168.1.5")
device.connect()
val lightSensor = device.lightSensors.first()

lightSensor.startSampling()

lightSensor.updateBroadcaster.consumeEach { measurement ->
    // Do something with every measurement
}

The only difference here is we now first connect to the device before using it. We're running everything in some suspending block now because both creating a RemoteAndroidDevice and connecting to one is a suspending operation.

In val device = RemoteAndroidDeivce("192.168.1.5:8080") the parameter passed to RemoteAndroidDevice() is the IP address of the device, in this example it's on the local network but it doesn't have to be. We have a Locator class for automating the process of finding and connecting to remote devices but it's not yet implemented for Android Devices.

Supporting Devices

There are two categories of device Kuantify can support.

  1. Devices that are intended to be controlled remotely and run some firmware facilitating remote connections. Examples include National Instruments / Agilent data acquisition systems and Arduino boards.
  2. Devices which are intended to have software installed directly on them. Examples include standard Android devices and Raspberry Pi.

Both categories utilize Kuantifies built in message routing system for communication, although in the case of full stack devices most of the routing should be handled automatically.

First Category - Remote Only Devices

Adding support for a remote only device is basically a matter of binding the functionality of the device to the Kuantify interfaces. So you'll create a class that extends Device and potentially some of the other Device interfaces depending on the functionality of the device you're adding support for. In most cases this process is a matter of establishing some connection, serializing / deserialzing commands and incoming data and sending / receiving these things to and from the device. Most of this process will be handled with a custom NetworkCommunicator. Kuantify already depends on kotlinx.serialization and parts of ktor so you should consider using them for these tasks. Ktor supports raw tcp and udp sockets which are often necessary for supporting a device. Both of these libraries are pure kotlin and multiplatform so you only need to make one set of common code to support the device on all platforms if you use them.

Second Category - Full Stack Devices

To support the second category we use Kuantify's "full stack" abstractions because we will be facilitating both the hosting / serving on device that has the sensors and controls and the remote connection and control on any device we may want to interact with the host device from. A full stack device can be used locally only or remotely only but support should be added for both, hence the name. There are three steps for adding support for a full stack device.

  1. Extend interface FSDevice for the device to define what Inputs, Outputs and potentially other functionality is has available.
  2. Extend LocalDevice to create Kuantify bindings for the devices functionality.
  3. Extend FSRemoteDevice so the device can be connected to remotely. This step should be very simple as in most cases Kuantify will automatically handle all the communication between the local and remote device so long as the local device is set up correctly.

This structure helps programs be more maintainable because the majority of the code can be written in a way where it doesn't care if it's running on the device itself or it is remotely connected to the device.