getComputeResult

Get a download URL for a finished job's result: by default its output, outputs.tar.

Usage

const url = await nautilus.getComputeResult({ jobId: 'a-job-id' }) 
 
if (url) {
  const results = await fetch(url)
}

Returns

• Type: string | undefined

undefined — with a log line explaining which — when the node does not know the job, the job has not finished, or it has no output result and no resultIndex was given.

A job has finished once the node sets its dateFinished, whatever its status: a failed job (41, 11, …) has finished too, usually without an output, and its other results are readable through resultIndex. The node sets 71 (JobSettle) before it writes outputs.tar, so a job at 71 without an output result yet is reported as not ready (logged, not warned) rather than as having none; try again.

The URL is signed with your address, a nonce and a signature, and fetches the result for whoever holds it: do not log or share it.

The output result is outputs.tar, a tar archive of what the algorithm wrote to its outputs folder. It is listed only when the results stay on the node, not when the job was started with output storage or an outputBucketId.

To stream a large result instead of downloading it through a URL, use streamComputeResult.

Parameters

config

• Type: ComputeResultConfig

Properties

jobId

• Type: string

<environmentHash>-<jobId>, as compute returns it.

resultIndex (optional)

• Type: number

The index of the result to fetch, from the job's results; a non-negative integer, or a RangeError is thrown. Defaults to its output result. On ocean-node 4.2 a finished job lists imageLog (image.log), configurationLog (configuration.log), algorithmLog (algorithm.log) and output (outputs.tar), in that order, and publishLog (publish.log) when the node wrote one.

nodeUri (optional)

• Type: string