Add PipelineTemplateResource

[mathislucka]

Overview

We want to add a PipelineTemplateResource to our AsyncDeepsetClient that allows us to interact with pipeline templates on the deepset platform. Specifically, we will allow to get a specific template by name as well as allowing to list all templates.

Roughly:

class PipelineTemplateResource:
  def __init__(self, client, workspace):
    ...
  async def get_template(template_name):
    ...
  async def list_templates(limit = 100):
    ...

API Spec
This is how to list templates:

curl --request GET \
     --url 'https://api.cloud.deepset.ai/api/v1/workspaces/workspace_name/pipeline_templates?limit=100&page_number=1&field=created_at&order=DESC' \
     --header 'accept: application/json'

This is a response example:

{
  "data": [
    {
      "author": "string",
      "available_to_all_organization_types": false,
      "best_for": [
        "string"
      ],
      "deepset_cloud_version": "v2",
      "description": "string",
      "expected_output": [
        "string"
      ],
      "indexing_yaml": "string",
      "name": "string",
      "pipeline_name": "string",
      "pipeline_template_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6",
      "pipeline_type": "indexing",
      "potential_applications": [
        "string"
      ],
      "query_yaml": "string",
      "recommended_dataset": [
        "string"
      ],
      "tags": [
        {
          "name": "string",
          "tag_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6"
        }
      ]
    },
    {
      "additionalProp": {}
    }
  ],
  "has_more": true,
  "total": 0
}

This is how to get a single template:

curl --request GET \
     --url https://api.cloud.deepset.ai/api/v1/workspaces/some_workspace/pipeline_templates/template_name \
     --header 'accept: application/json'

This is an example response:

{
  "author": "string",
  "available_to_all_organization_types": false,
  "best_for": [
    "string"
  ],
  "deepset_cloud_version": "v2",
  "description": "string",
  "expected_output": [
    "string"
  ],
  "indexing_yaml": "string",
  "name": "string",
  "pipeline_name": "string",
  "pipeline_template_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6",
  "pipeline_type": "indexing",
  "potential_applications": [
    "string"
  ],
  "query_yaml": "string",
  "recommended_dataset": [
    "string"
  ],
  "tags": [
    {
      "name": "string",
      "tag_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6"
    }
  ]
}

Note: We dropped the indexing_yaml. When creating a model for the PipelineTemplate you should only consider the query_yaml (same as for DeepsetPipeline).

Note: map pipeline_name to template_name in the response model

Steps

• reference src/deepset_mcp/api/pipeline as a point of comparison to check out how to configure the PipelineTemplateResource and the corresponding models
• add a protocol for the resource in src/deepset_mcp/api/protocols.py
• add the resource as a method both to the AsyncClientProtocol as well as on the actual client: src/deepset_mcp/api/client.py
• refer to test/unit/api/pipeline/test_pipeline_resource.py to see how to test a resource, test the PipelineTemplateResource using the same approach. Use the same BaseFakeClient and add the template resource
• don't forget to add the resource to the BaseFakeClient in test/unit/conftest.py

Use the same coding style and conventions as throughout the project. Make sure to use proper type annotations. The PR should pass our CI!

[mathislucka]

I can see that you already implemented most of the changes but we need a few key changes:

• you created a new directory called tests although all tests should go into the test directory
• delete tests/unit/conftest.py we already have the FakeClient implemented in test
• move your resource tests to test/unit/api/pipeline_template
• revisit the unit tests you created in tests/unit/api/pipeline_template/test_pipeline_template_resource.py (after moving them) -> we are missing some sample PipelineTemplates that you can use for testing, as it stands, your unit tests don't do much