CODEWITHSUNDEEP
restgoogle-cloud-vertex-aigoogle-ai-platform
Taylor F Saraiva
Taylor F Saraiva

Reputation: 11

How to create PipelineJob with pipeline spec stored in GCS

Trying to create a PipelineJob following the Vertex AI API docs.

According to the docs, if a templateUri is given then it will use it to download the pipelineSpec

A template uri from where the PipelineJob.pipeline_spec, if empty, will be downloaded.

But when calling the API, it is complaining about missing required field.

REQUEST

{
    "displayName": "pipeline test",
    "templateUri": "gs://bucket-name/3a279d63-f965-4db2-b54c-ad933b04d278.json"
}

POST - POST https://{service-endpoint}/v1/{parent}/pipelineJobs

RESPONSE

{
    "error": {
        "code": 400,
        "message": "Required field pipelineInfo is not specified. ",
        "status": "INVALID_ARGUMENT"
    }
}

EXPECTED

PipelineInfo downloaded from templateUri

Upvotes: 1

Views: 162

Answers (1)

krishr2d2
krishr2d2

Reputation: 64

The closest I got to resolving this issue was when I found that the pipelineInfo has a field called "name" from google. enter image description here

After I provide some random name, it throws the following error. I couldn't find reference for the "tasks" field either from the documentation.

"error": {
"code": 400,
"message": "Required field tasks is not specified. ",
"status": "INVALID_ARGUMENT"}

Upvotes: 0

Related Questions