This library provides convenient access to the Wandelbots API from frontend JavaScript applications. Currently this contains typed methods for each API endpoint. We are also working on providing some higher level abstractions to manage the websocket connection state for tracking robot movement and handling jogging and Wandelscript program execution, which will be included in this package.
npm install @wandelbots/wandelbots-js
If you develop an react application we also provide a set of react components which you can use together with this library.
The core of this package is the NovaClient
, which represents a connection to a configured robot cell on a given Nova instance:
import { NovaClient } from "@wandelbots/wandelbots-js"
const nova = new NovaClient({
instanceUrl: "https://example.instance.wandelbots.io",
cellId: "cell",
// Auth details come from the developer portal when you create an instance
username: "wb",
password: "SOME_PASSWORD",
})
You can make calls to the REST API via nova.api
, which contains a bunch of namespaced methods for each endpoint generated from the OpenAPI spec and documentation.
For example, to list the devices configured in your cell:
const devices = await nova.api.deviceConfig.listDevices()
// -> e.g. [{ type: 'controller', identifier: 'abb_irb1200_7', ... }, ...]
Documentation for the various API endpoints is available on your Nova instance at /api/v1/ui
(public documentation site is in the works)
NovaClient
has various convenience features for websocket handling in general. Use openReconnectingWebsocket
to get a persistent socket for a given Nova streaming endpoint that will handle unexpected closes with exponential backoff:
const programStateSocket = nova.openReconnectingWebsocket(`/programs/state`)
this.programStateSocket.addEventListener("message", (ev) => {
console.log(ev.data)
})
Websockets on a given Nova client are deduplicated by path, so if you call openReconnectingWebsocket
twice with the same path you'll get the same object. The exception is if you called dispose
, which you may do to permanently clean up a reconnecting websocket and free its resources:
programStateSocket.dispose()
The reconnecting websocket interface is fairly low-level and you won't get type safety on the messages. So when available, you'll likely want to use one of the following endpoint-specific abstractions instead which are built on top!
The library provides an easy to use way to access properties of a motion group.
activeRobot = await this.nova.connectMotionGroup(`some-motion-group-id`)
This connected motion group opens a websocket and listens to changes of the current joints and the TCP pose. You can read out those values by using the rapidlyChangingMotionState
of the object. Along other properties it also provides the current safetySetup
and tcps
.
const newJoints = activeRobot.rapidlyChangingMotionState.state.joint_position.joints
To render a visual representation, you can use the robot
component of the react components.
The ProgramStateConnection
provides an object which allows to execute and stop a given Wandelscript.
import script from "./example.ws"
...
programRunner.executeProgram(script)
You can stop
the current execution or listen to state updates of your wandelscript code by observing the programRunner.executionState
.
Jogging in a robotics context generally refers to the manual movement of the robot via direct human input. The Wandelbots platform provides websocket-based jogging methods which can be used to build similar jogging interfaces to those found on teach pendants.
const jogger = await nova.connectJogger(`some-motion-group-id`)
The jogger has two mutually exclusive modes. You must set the appropriate jogging mode before starting a jogging motion; this ensures that the motion is ready to start immediately when called with minimal delay.
// Set the jogger to "joint" mode, enabling continuous joint rotations.
await jogger.setJoggingMode("joint")
// Set the jogger to "tcp" mode, enabling continuous translation
// and rotation movements of the tool center point.
await jogger.setJoggingMode("tcp", {
tcpId: "flange",
coordSystemId: "world",
})
For safety purposes, let's first consider how to stop the jogger. Calling stop will stop all motion types regardless of mode:
await jogger.stop()
As a failsafe, the server will also stop any jogging motions when it detects the relevant websocket has been closed. This means that if e.g. the network connection drops out or the browser tab is closed in the middle of a motion, it will stop automatically.
However, you should never totally rely on any software being able to stop the robot: always have the hardware emergency stop button within reach just in case!
Requires joint
mode. This example starts joint 0 of the robot rotating in a positive direction at 1 radian per second:
await jogger.startJointRotation({
joint: 0,
direction: "+",
velocityRadsPerSec: 1,
})
Requires tcp
mode. This example starts moving a TCP in a positive direction along the X axis of the specified coordinate system, at a velocity of 10 millimeters per second:
await jogger.startTCPTranslation({
axis: "x",
direction: "+",
velocityMmPerSec: 10,
})
Requires tcp
mode. This example starts rotating the TCP in a positive direction around the X axis of the specified coordinate system, at a velocity of 1 radians per second:
await jogger.startTCPRotation({
axis: "x",
direction: "+",
velocityRadsPerSec: 1,
})
When you are done with a jogger, make sure to call dispose:
await jogger.dispose()
This will close any open websockets and ensure things are left in a good state.
You can use the Jogging Panel from the react components library to get a easy to use visualization component.
To set up wandelbots-js for development, first clone the repo and run:
npm install
Then you can run the tests which will use a mocked Nova API and socket connections:
npm run test