Name a dataclip

[taylordowns2000]

User story

As an admin that wants to run a job with an arbitrary dataclip (#238) I want to be able to select that input dataclip from a list of saved and named dataclips.

Consider the two types of NAMED dataclips:

Type 1: There are inputs I want to save to test my jobs. These, when saved, form a “test suite” of cases that we’d like to validate for a particular job in a flow. They are results from previous runs (save as a test input for job3 in the flow) and saved webhook requests.

Type 2: There are constants, or blobs of JSON data, that non-technical users want to build via a simple UI and, most commonly, store “mapping” data that links, e.g. a clinic code to it’s underlying clinic name.

Details

Currently, there is a picklist which allows the user to change the type of a dataclip. This picklist should dissappear. Instead, they should have a button which says "SAVE AS" and if they do so, they should be prompted to give that dataclip a name which follows the same naming pattern as projects (i.e., it's a slug.) and be able to edit the body of the dataclip.

Once it's saved as a named_input, they can not convert it back to a run_result or a http_request.

Once it's saved as a named_input, they can set the name to null, effectively removing it from the "named globals" list.

Implementation notes

• [ ] add a nullable column on the dataclips table called name
• [ ] add this nullable name to the dataclip model
• [ ] disable the body editor when type is run_result or http_request
• [ ] remove the type picklist
• [ ] add a "Save as Global" button which, when clicked makes the body editable and displays the name field... from this altered view of the form the user may still "Cancel" or "Save".

Release notes

Tests

QA

• [ ] This has been QA’d by Amber, branch can be merged after review

[stuartc]

re a conversation with @amberrignell:

Marking cloned dataclips with 'associated jobs'

I brought up my thoughts around a suggested feature relating to 'tagging' the associated job/s for a cloned dataclip; I'd like to try implementing this without that feature - the thought is that the list of available dataclips for running (when coming from a job) will be ordered by relevance, and in the (less common) case where a cloned dataclip was created without running - it will be associated on the first run. My hope is to avoid having extra columns which in the context of datclips will be 99% empty. The approach outlined here shares almost the entire implementation requirement of the initial one. So if we feel that having a user editable 'associated job' linked to a dataclip we can add that on top.

It do see that this issue doesn't make any mention of what I'm referring to, so this note could be a non-issue and can be ignored.