This repository is a quick way to skip the boilerplate and get started writing an iTwin.js Agent.
Simply clone, npm install, setup your config, and you're ready to go!
The provided agent app almost works right out of the box, but you'll need to configure it with your project, iModel, and OIDC information.
To do this, just create a .env file at the project root with the following:
###############################################################################
# This file contains secrets - don't commit or share it!
###############################################################################
# Specify an iModel
CONTEXT_ID=
IMODEL_ID=
# OIDC configuration
# Don't forget to add <CLIENT_ID>@apps.imsoidc.bentley.com to your Project too!
CLIENT_ID=
CLIENT_SECRET=
imjs_itwin_platform_authority=https://ims.bentley.comThe CLIENT_ID and CLIENT_SECRET can be obtained by registering an app from the iTwin Platform Portal - be sure to create a "Service" app and add the "Visualization" API!
Once you've created your .env file, you can build your agent via npm run build.
Or - even better - start TypeScript in watch/incremental rebuild mode: npm run build -- --watch
To run your agent, you can simply do npm start (or node . if you're in the project root dir). This will continuously poll for new changesets pushed to the iModelHub.
Once the agent is listening for changesets, you can use the iTwin Synchronizer to synchronize a change.
For testing, it can often also be useful to skip the event listening and just run against a specific changeset. To do that, either run npm start -- --latest to use the latest changeset, or npm start -- --changeset=<CHANGESETID> to use any specific changeset.
The entry point of for both single-changeset test runs and handling individual iModelHub change events is the MyAgent.run() method. There you can implement your agent's behavior after the iModel is downloaded.