Dataiku + Stream Deck

Start a run of a DSS scenario — the usual way to kick off an orchestrated pipeline (a sequence of builds, checks and reporters) as opposed to building a single dataset, which Build Dataset does. Use List Scenarios to find a valid scenario ID. A successful call only means the run was accepted, and the response carries no run identifier, so poll List Scenario Runs to follow the outcome. Requires the RUN_JOBS privilege on the project. View docs

List the scenarios of a DSS project, with each scenario's id, whether it is currently running, and whether it is active (i.e. responding to its own triggers). Call this before Run Scenario to find a valid scenario ID, or to check whether a scenario is already in flight before starting another run. Requires the MONITOR_JOBS privilege on the project. View docs

Retrieve the last runs of a DSS scenario. Use this after Run Scenario, which returns no run identifier of its own, to follow the outcome: each entry carries a runId, start/end timestamps and a result object reporting outcome (e.g. SUCCESS) and type (e.g. SCENARIO_DONE). This tool cannot tell you which entry corresponds to a run you started: the response is just the scenario's recent runs, so a concurrent run, or one already in progress before you called Run Scenario, looks no different. Correlate deliberately — match on the start timestamp (the runId, e.g. 2016-04-15-16-57-37-759, is derived from it) against the moment you triggered the run, rather than assuming any particular entry is yours. Stop polling once the run you are tracking reports a result, whatever its outcome. Requires the RUN_JOBS privilege on the project. View docs

List the projects on the DSS instance. Start here when you only know a project by its display name: every other Dataiku tool is addressed by projectKey (e.g. MYPROJECT), which this tool returns. Only projects the connected API key holds the READ_CONF privilege on are listed, so an empty result usually means a permissions gap rather than an empty instance. Note that the free edition of DSS does not include Public API access — the API key must come from a trial or licensed instance. View docs

Retrieve the latest build jobs of a DSS project, each with its jobId and state. Use this to check what a project has been building recently, or to recover a jobId you no longer have before calling Get Job Status. Requires the READ_CONF privilege on the project. View docs

List the datasets of a DSS project. Use this to discover a dataset's name — the identifier Build Dataset needs to build it — along with its type (e.g. Filesystem) and connection parameters. Use List Projects first if you do not know the project key. Requires the READ_CONF privilege on the project. View docs

Check the status of a DSS build job, returned as baseStatus.status. Poll this after Build Dataset using the id it returned. NOT_STARTED and RUNNING mean the job is still in flight; DONE, FAILED and ABORTED are terminal, so stop polling on any of them and treat FAILED/ABORTED as an unsuccessful build. Use List Jobs to recover a job ID you no longer have. Requires the MONITOR_JOBS privilege on the project. View docs

Start a job that builds one or more outputs (typically datasets) in a DSS project. Use this to rebuild specific outputs directly; use Run Scenario instead when the pipeline is already orchestrated as a scenario. Use List Datasets to find valid output names. A successful call only means the job was accepted — the response's id is the job ID, which you pass to Get Job Status to follow it to completion. Requires the RUN_JOBS privilege on the project. View docs