Managed Inputs

Managed inputs are used to store and keep track of your inputs to use for an experiment or run. They can be created either by referencing a prior run within your application, or by uploading a new input. A major benefit of using an input as part of an input set is a potentially long run does not have to be completed in order to start a batch experiment or perform manual testing via the run endpoint. Using managed inputs for individual runs may also be helpful for the purposes of reproducibility.

There are several interfaces for creating managed inputs:

• Console: use the web interface.
• Cloud API: use the HTTP endpoints.
• Python SDK: use python

Console

Go to the Console web interface, and open your app. Go to the Experiments > Inputs tab. Click on New Input. Fill in the fields.

Cloud API

Define the desired input ID, name, and your input's format (csv-archive and multi-file for .tar.gz files, json, or text). After, create the input by:

• Referencing a run ID

{
    "run_id": "nextroute-12345",
    "name":"my new input",
    "description":"Input based on nextroute-12345",
    "format": {
        "input": {
            "type" : "json"
        }
    }
}

• Or, referencing an upload ID

Please see the large payloads page to learn how to upload your input to retrieve an upload ID.

{
    "id": "",
    "run_id": "9a2a2835-559d-496b-a732-e30ccc329a92",
    "name":"my new input",
    "description":"Input based on an upload",
    "format": {
        "input": {
            "type" : "json"
        }
    }
}