Call Endpoint
Synopsis
Allow to call Rest APIs or Python Models inside existing workflows.
Description
The operator enables the call of deployments. For deployed workflows the input is at most one table, because an input may not be necessary. In the case of deployed models, the operator accepts arbitrary number of input tables, as some cases pre-processing and post-processing will be needed. After applying the model to the input(s), the output will be the modified table(s). Besides selecting AI Cloud endpoints from accessible projects, it is also possible to input them manually.
The output format value only affects manually entered deployments. When selecting a deployment from the available options, the system automatically applies the default output format. However, for manually entered inputs, users must ensure the selected deployment uses the correct output format. Failing to do so may result in errors or inaccurate outputs.
Input
input
The Call Endpoint expects arbitrary number of inputs, but for deployed workflows at most one table is allowed.
Output
output
The Call Endpoint return arbitrary number of outputs, but for deployed workflows at most one table is returned.
Parameters
accessible project
A toggle button to decide, whether the desired deployment is in an accessible project or not.
deployment
The deployment, which can be found in the accessible projects. It will be applied on the input table.
deployment type
Whether the type of the deployment is Rest API or Python Model.
url
The url of the AI Cloud deployment, that will be applied on the input table.
auth token
The token is needed to access the deployment in case it has a secured endpoint.