All supported message transforms types.
| JSON representation |
|---|
{ "enabled" : boolean , "disabled" : boolean , // Union field |
enabled
(deprecated)
boolean
Optional. This field is deprecated, use the disabled
field to disable transforms.
disabled
boolean
Optional. If true, the transform is disabled and will not be applied to messages. Defaults to false
.
transform
. The type of transform to apply to messages. transform
can be only one of the following:javascriptUdf
object (
JavaScriptUDF
)
Optional. JavaScript User Defined Function. If multiple JavaScriptUDF's are specified on a resource, each must have a unique functionName
.
aiInference
object (
AIInference
)
Optional. AI Inference. Specifies the Vertex AI endpoint that inference requests built from the Pub/Sub message data and provided parameters will be sent to.
JavaScriptUDF
User-defined JavaScript function that can transform or filter a Pub/Sub message.
| JSON representation |
|---|
{ "functionName" : string , "code" : string } |
| Fields | |
|---|---|
functionName
|
Required. Name of the JavasScript function that should applied to Pub/Sub messages. |
code
|
Required. JavaScript code that contains a function
|
AIInference
Configuration for making inference requests against Vertex AI models.
| JSON representation |
|---|
{ "endpoint" : string , "serviceAccountEmail" : string , // Union field |
endpoint
string
Required. An endpoint to a Vertex AI model of the form projects/{project}/locations/{location}/endpoints/{endpoint}
or projects/{project}/locations/{location}/publishers/{publisher}/models/{model}
. Vertex AI API requests will be sent to this endpoint.
serviceAccountEmail
string
Optional. The service account to use to make prediction requests against endpoints. The resource creator or updater that specifies this field must have iam.serviceAccounts.actAs
permission on the service account. If not specified, the Pub/Sub service agent
, service-{project_number}@gcp-sa-pubsub.iam.gserviceaccount.com, is used.
inference_mode
. The format of inference requests made to the endpoint. inference_mode
can be only one of the following:unstructuredInference
object (
UnstructuredInference
)
Optional. Requests and responses can be any arbitrary JSON object.
UnstructuredInference
Configuration for making inferences using arbitrary JSON payloads.
| JSON representation |
|---|
{ "parameters" : { object } } |
| Fields | |
|---|---|
parameters
|
Optional. A parameters object to be included in each inference request. The parameters object is combined with the data field of the Pub/Sub message to form the inference request. |

