Cognite Scala SDK

Under development, not recommended for production use cases

This is the Cognite Scala SDK for developers working with Cognite Data Fusion.

Visit Maven Repository to see the available versions, and how to include it as a dependency for sbt, Maven, Gradle, and other build tools.

Authentication is specified using an implicit Auth parameter, which by default looks for an API key to use in the COGNITE_API_KEY environment variable.

The GenericClient requires a backend for sttp, which can use any effects wrapper (for example, cats-effect, but users who do not want to use an effect wrapper can use Client to easily create a client using an identity effect.

Examples

Create a simple client, specifying an application name:

import com.cognite.sdk.scala.v1._

val c = Client("scala-sdk-examples", "https://api.cognitedata.com")

Create a client using the cats-effect IO:

import com.cognite.sdk.scala.v1._
import java.util.concurrent.Executors
import scala.concurrent._
import cats.effect.IO
import com.softwaremill.sttp.asynchttpclient.cats.AsyncHttpClientCatsBackend

implicit val cs = IO.contextShift(ExecutionContext.fromExecutor(Executors.newCachedThreadPool()))
implicit val sttpBackend = AsyncHttpClientCatsBackend[cats.effect.IO]()
val c = new GenericClient[IO, Nothing]("scala-sdk-examples", "https://api.cognitedata.com")

The following examples will use Client with the identity effect. Add .unsafeRunSync or similar to the end of the examples if you use IO or another effect wrapper.

Getting started with Ammonite

You can use Ammonite for a better interactive Scala environment, and load the SDK directly like this:

import $ivy.`com.cognite::cognite-sdk-scala:0.1.2`
import com.cognite.sdk.scala.v1._

Discover time series

List and filter endpoints use Stream from fs2, which loads more data as required. You can use .compile.toList to convert it to a list, but note that this could end up fetching a lot of data unless you limit it, for example by using .take(25) as we do here.

For the next examples, you will need to supply ids for the time series that you want to retrieve. You can find some ids by listing the available time series:

val timeSeriesList = c.timeSeries.list().take(25).compile.toList
val timeSeriesId = timeSeriesList.head.id
print(c.timeSeries.retrieveById(timeSeriesId))

Retrieve data points

If you have a time series ID you can retrieve its data points:

c.dataPoints.queryById(timeSeriesId, 0, Instant.now())

It is also possible to query aggregate values using a time series ID. Possible aggregate values can be found in the API documentation You must also specify a granularity.

val aggregates = Seq("count, average, max")
c.dataPoints.queryAggregatesById(timeSeriesId, Instant.ofEpochMilli(0L), Instant.now(), "1d", aggregates)

If you need only the last data point for a time series or group of timeseries, you can retrieve these using:

val latestPoints: Map[Long: Option[DataPoint]] = c.dataPoints.getLatestDataPointsByIds(Seq(timeSeriesId))

This returns a map from each of the time series IDs specified in the function call to the latest data point for that time series, if it exists, or None if it does not.

There are analogous functions available to execute these queries using external IDs instead of IDs and returning string valued data points rather than numeric valued data points.

Insert and Delete Data

It is possible to insert and delete numeric data points for specified time series. To insert data, you must specify the time series for which to insert data and the data points to insert:

val testDataPoints = (startTime to endTime by 1000).map(t => DataPoint(Instant.ofEpochMilli(t), math.random))
c.dataPoints.insertById(timeSeriesId, testDataPoints)

To delete data, you must specify the time series and the range of times for which to delete data:

c.dataPoints.deleteRangeById(timeSeriesId, Instant.ofEpochMilli(0L), Instant.now())

Create an asset hierarchy

val root = AssetCreate(name = "root", externalId = Some("1"))
val child = AssetCreate(name = "child", externalId = Some("2"), parentExternalId = Some("1"))
val descendant = AssetCreate(name = "descendant", externalId = Some("3"), parentExternalId = Some("2"))
val createdAssets = c.assets.create(Seq(root, child, descendant))

// clean up the assets we created
c.assets.deleteByExternalIds(Seq("1", "2", "3"))

Print an asset subtree using filter

def children(parents: Seq[Asset]): List[Asset] =
  c.assets.filter(AssetsFilter(parentIds = Some(parents.map(_.id)))).compile.toList

def printSubTree(parents: Seq[Asset], prefix: String = ""): Unit = {
  if (parents.nonEmpty) {
    val assetsUnderThisLevel = children(parents)
    parents.foreach { p =>
      println(prefix + p.name)
      val assetsUnderThisAsset = assetsUnderThisLevel.filter(_.parentId.exists(pid => pid == p.id))
      printSubTree(assetsUnderThisAsset, prefix + "  ")
    }
  }
}

val rootId = 2780934754068396L
printSubTree(c.assets.retrieveByIds(Seq(rootId)))

Find all events in an asset subtree

def eventsForAssets(parents: Seq[Asset]): List[Event] =
  c.events.filter(EventsFilter(assetIds = Some(parents.map(_.id)))).compile.toList

def eventsInSubTree(parents: Seq[Asset]): Seq[Event] = {
  if (parents.nonEmpty) {
    val assetsUnderThisLevel = children(parents)
    eventsForAssets(parents) ++ parents.flatMap { p =>
      val assetsUnderThisAsset = assetsUnderThisLevel.filter(_.parentId.exists(pid => pid == p.id))
      eventsInSubTree(assetsUnderThisAsset)
    }
  } else {
    Seq.empty[Event]
  }
}

eventsInSubTree(c.assets.retrieveByIds(Seq(rootId))).foreach(event => println(event.`type`))

3D

To list 3D models:

val models = c.threeDModels.list().take(10).compile.toList

You can create, update, delete, and retrieve models based on their ID. To access model revisions for a specific model:

val modelId = modelIds.head.id
val revisions = c.threeDRevisions(modelId).list().take(10).compile.toList

You can also create, update, delete, and retrieve revisions based on their IDs. You can also list all the nodes in a particular revision, as well as the ancestors of a particular node (including the node itself):

val revisionId = revisionIds.head.id
val nodes = c.threeDNodes(modelId, revisionId).list().listWithLimit(10).compile.toList
val nodeId = nodeIds.head.id
val ancestorNodes = c.threeDNodes(modelId, revisionId).ancestors(nodeId).compile.toList