We were using metadata client by legacy. Architecture is not great for
Core engine: workflows are available both in data and metadata client.
It makes more sense to use the data client since workflows are part of
standard objects
After hitting use as draft, we redirect the user to the workflow page.
If the user already accessed that page, the workflow and its current
version will be stored in cache. So we also need to update that cache.
I tried to update it manually but it was more complex than expected.
Steps and trigger are not fully defined objects.
I ended with a simple refetch query that I wanted to avoid but that is
at least fully working with minimum logic.
When a workflow was activate, the previous activated version was still
available in cmd+k
This is because we were only updating entities in cache after
activation.
We also need to update the queries stored in cache, because this is
where the cmd+k look to know which workflow is active.
https://github.com/user-attachments/assets/750cc24a-7583-4641-856c-0f9892d26331
- In the `formatFieldMetadataValue` function, allow people to call
TypeORM's `save()` method with unserialized JSON data.
- Create an `overrideWorkflowDraftVersion` mutation that takes a
workflow id and the id of the workflow version to use as the new draft
- If no draft exists yet, create one
- If a draft already exists, deactivate its serverless functions
- Duplicate every step. For serverless function steps, it includes
duplicating the functions
- Save the data of the step in DB
- Call the `overrideWorkflowDraftVersion` mutation in the old workflow
header and in the new Cmd+K actions
- I chose to not update the Apollo cache manually as the information of
the new draft are going to be automatically fetched once the user lands
on the workflow's show page. Note that we redirect the user to this page
after overriding the draft version.
Implemented the following actions for the workflows:
- Test Workflow
- Discard Draft
- Activate Draft
- Activate Workflow Last Published Version
- Deactivate Workflow
- See Workflow Executions
- See Workflow Active Version
- See Workflow Versions History
And the following actions for the workflow versions:
- Use As Draft
- See Workflow Versions History
- See Workflow Executions
Video example:
https://github.com/user-attachments/assets/016956a6-6f2e-4de5-9605-d6e14526cbb3
A few of these actions are links to the relations of the workflow object
(link to a filtered table). To generalize this behavior, I will create
an hook named `useSeeRelationsActionSingleRecordAction` to automatically
generate links to a showPage if the relation is a Many To One or to a
filtered table if the relation is a One to Many for all the record
types.
I will also create actions which will allow to create a relation.
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
- create a serverless function when creating a new workflow code step
- add code editor in workflow code step
- move workflowVersion steps management from frontend to backend
- add a custom resolver for workflow-version management
- fix optimistic rendering on frontend
- fix css
- delete serverless function when deleting workflow code step
TODO
- Don't update serverlessFunction if no code change
- Factorize what can be between crud trigger and crud step
- Publish serverless version when activating workflow
- delete serverless functions when deleting workflow or workflowVersion
- fix optimistic rendering for code updates
- Unify CRUD types
<img width="1279" alt="image"
src="https://github.com/user-attachments/assets/3d97ee9f-4b96-4abc-9d36-5c0280058be4">
In this PR:
- Refactored components to clarify their behavior. For example, I
renamed the `Workflow` component to `WorkflowVisualizer`. This moved
forward the issue #7010.
- Create two variants of several workflow-related components: one
version for editing and another for viewing. For instance, there is
`WorkflowDiagramCanvasEditable.tsx` and
`WorkflowDiagramCanvasReadonly.tsx`
- Implement the show page for workflow versions. On this page, we
display a readonly workflow visualizer. Users can click on nodes and it
will expand the right drawer.
- I added buttons in the header of the RecordShowPage for workflow
versions: users can activate, deactivate or use the currently viewed
version as the next draft.
**There are many cache desynchronisation and I'll fix them really
soon.**
## Demo
(Turn sound on)
https://github.com/user-attachments/assets/97fafa48-8902-4dab-8b39-f40848bf041e
- Allows the deletion of triggers and steps in workflows. If the
workflow can not be edited right now, we create a new draft version.
- The workflow right drawer can now render nothing. It's necessary to
behave that way because a deleted step will still be displayed for a
short amount of time in the drawer. The drawer will be filled with blank
content when it disappears.
https://github.com/user-attachments/assets/abd5184e-d3db-4fe7-8870-ccc78ff23d41Closes#7057
- Add the SAVE_EMAIL action. This action requires more setting
parameters than the Serverless Function action.
- Changed the way we computed the workflow diagram. It now preserves
some properties, like the `selected` property. That's necessary to not
close the right drawer when the workflow back-end data change.
- Added the possibility to set a label to a TextArea. This uses a
`<label>` HTML element and the `useId()` hook to create an id linking
the label with the input.
## Setup
This PR can be tested only if some feature flags have specific values:
- `IsWorkflowEnabled` equals `true`
- `IsQueryRunnerTwentyORMEnabled` equals `false`
These feature flags weren't committed to don't break other branches.
## What this PR brings
- Display buttons to activate and deactivate a workflow version and a
button to discard the current draft version. I also scaffolded a "Test"
button, which doesn't do anything for now.
- Wired the activate, deactivate and discard draft buttons to the
backend.
- Made it possible to "edit" active and deactivated versions by
automatically creating a new draft version when the user tries to edit
the version.
- Hide the "Discard Draft", button if the current version is not a draft
or is the first version ever created.
- On the backend, don't consider discarded drafts when checking if a new
draft version can be created.
- On the backend, disallow deleting the first created workflow version.
Otherwise, we will end up with a blank canvas in the front end, and it
will be impossible to recover from it.
- On the backend, disallow running deactivation steps if the workflow
version is not currently active. Previously, we were throwing, which is
unnecessary as it's a valid case.
## Spotted bugs that we must dive into
### Duplicate workflow versions in Apollo cache
https://github.com/user-attachments/assets/7cfffd06-11e0-417a-8da0-f9a5f43b84e2
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
- Create a workflow version when the user visits an empty workflow.
- If the trigger is not defined yet and the user selects either the
standard object type or the event type first, we automatically select
the first option of the other value. Indeed, every state update is
automatically saved on the backend and we need both standard object and
event types to save the event name.
- Introduces a change in the backend. I removed the assertions that
throw when a workflow version is not complete, that is, when it doesn't
have a defined trigger, which is the case when scaffolding a new
workflow with a first empty workflow version.
- We should keep validating the workflow versions, at least when we
publish them. That should be done in a second step.
- Improve the design of the right drawer
- Allow to update the trigger of the workflow: the object and the event
listened to
- Allow to update the selected serverless function that a code action
should execute
- Change how we determine which workflow version to display in the
visualizer. We fetch the selected workflow's data, including whether it
has a draft or a published version. If the workflow has a draft version,
it gets displayed; otherwise, we display the last published version.
- I used the type `WorkflowWithCurrentVersion` to forward the currently
edited workflow with its _current_ version embedded across the app.
- I created single-responsibility hooks like
`useFindWorkflowWithCurrentVersion`, `useFindShowPageWorkflow`,
`useUpdateWorkflowVersionTrigger` or `useUpdateWorkflowVersionStep`.
- I updated the types for workflow related objects, like `Workflow` and
`WorkflowVersion`. See
`packages/twenty-front/src/modules/workflow/types/Workflow.ts`.
- This introduced the possibility to have `null` values for triggers and
steps. I made the according changes in the codebase and in the tests.
- I created a utility function to extract both parts of object-event
format (`company.created`):
`packages/twenty-front/src/modules/workflow/utils/splitWorkflowTriggerEventName.ts`
This PR adds the possibility of creating new steps. For now, only
actions are available. The steps are stored on the server, and the
visualizer is reloaded to include them.
Selecting a step opens the right drawer and shows its details. For now,
it's only the id of the step, but in the future, it will be the
parameters of the step.
In the future we'll want to let users add steps at any point in the
diagram. As a consequence, it's crucial to be able to walk in the tree
that make the steps to find the correct place where to put the new step.
I wrote a function that returns where the new step should be inserted.
This function will become recursive once we get branching implemented.
Things to mention:
- Reactflow needs every node and edge to have a unique identifier. In
this PR, I chose to use steps' id as nodes' id. That way, it's easy to
move from a node to a step, which helps make operations on a step
without resolving the step's id from the node's id.
