Overview
Consult REPO.md for task guidelines.
We are not implementing a full new resource, but we are adding a method to the existing PipelineResource that allows us to deploy a pipeline.
You should also add a corresponding tool and the mcp tool.
Make sure to also add tests (add them to existing test files for the pipeline resource and for the pipeline tools).
This is how to deploy a pipeline:
https://api.cloud.deepset.ai/api/v1/workspaces/home-electronics/pipelines/RAG-Qdrant/deploy
It is a POST request without payload (you need to set the right pipeline name and workspace name).
We get a 200, a 400, a 404 or a 422 (validation error). In case of the validation error, for the tool only add the details for the error as return value (same as for the validate tool).
Overview
Consult REPO.md for task guidelines.
We are not implementing a full new resource, but we are adding a method to the existing PipelineResource that allows us to deploy a pipeline.
You should also add a corresponding tool and the mcp tool.
Make sure to also add tests (add them to existing test files for the pipeline resource and for the pipeline tools).
This is how to deploy a pipeline:
It is a POST request without payload (you need to set the right pipeline name and workspace name).
We get a 200, a 400, a 404 or a 422 (validation error). In case of the validation error, for the tool only add the details for the error as return value (same as for the validate tool).