Better pipeline outputs

[qgerome]

In pipelines, users can call current_run.add_file_output & current_run.add_database_output. We transform those two types of output to their correct representation of BucketObject & DatabaseTable.

Example of return when querying a pipeline run

{
  "data": {
    "workspace": {
      "bucket": {
        "objects": {
          "hasNextPage": false,
          "hasPreviousPage": false,
          "items": [
            {
              "name": ".hidden_files",
              "key": ".hidden_files/"
            },
            {
              "name": "all",
              "key": "all/"
            },
            {
              "name": "README.MD",
              "key": "README.MD"
            },
            {
              "name": "covid_data.csv",
              "key": "covid_data.csv"
            },
            {
              "name": "demo.ipynb",
              "key": "demo.ipynb"
            }
          ]
        }
      }
    },
    "pipelineRun": {
      "id": "91efe9c7-277d-45e2-9d7a-3b9dbadf49fc",
      "outputs": [
        {
          "__typename": "GenericOutput",
          "name": "dhis2_extract.csv",
          "uri": "httsp://google.com"
        },
        {
          "__typename": "BucketObject",
          "key": "all/ChatGPT.log",
          "objectName": "ChatGPT.log",
          "path": "hexa-test-quentin-s-workspace-823946/all/ChatGPT.log",
          "size": 5319,
          "type": "FILE"
        }
      ]
    }
  }
}

[pvanliefland]

LGTM, but 2 remarks:

• Shouldn't we do the same for table outputs?
• The frontend needs a PR as well

[qgerome]

LGTM, but 2 remarks:

• Shouldn't we do the same for table outputs?
• The frontend needs a PR as well

What do you mean by "table outputs" ?
Frontend is coming.

[pvanliefland]

LGTM!