Avoid volume mode in hybrid tracings

[MichaelBuessemeyer]

This PR defaults to orthogonal mode when the active tracing is hybrid and the URL says to go into volume mode.

URL of deployed dev instance (used for testing):

• https://___.webknossos.xyz

Steps to test:

• Create a volume tracing and convert it into a hybrid tracing (Dataset info tab).
• The opened tracing should be a hybrid tracing in orthogonal mode.

Issues:

• fixes "Convert to Hybrid" tracing leads to page with "volume" mode activated #4460

---

• Updated changelog
• [ ] Updated migration guide if applicable
• [ ] Updated documentation if applicable
• [ ] Adapted wk-connect if datastore API changes
• [ ] Needs datastore update after deployment
• Ready for review

[MichaelBuessemeyer]

@daniel-wer I would like to open up the login popover (top right corner) when the dataset cloud not be loaded. The problem is that the dataset is fetched in the model-initialization. Can you think of a slightly elegant way on how to get the information, that the dataset could not be loaded, to the navbar to show the user login popover?
Using a new field in the store for this seems not appropriate in my opinion.

[philippotto]

@MichaelBuessemeyer I'm taking over for @daniel-wer, since he's currently on vacation :)

[philippotto]

@daniel-wer I would like to open up the login popover (top right corner) when the dataset cloud not be loaded. The problem is that the dataset is fetched in the model-initialization. Can you think of a slightly elegant way on how to get the information, that the dataset could not be loaded, to the navbar to show the user login popover?
Using a new field in the store for this seems not appropriate in my opinion.

I think, using a store field is the easiest solution. Something like state.uiInformation.showLoginMask which the navbar would read (and write to) as well as the model-initialization. Alternatively, only overkill solutions come to my mind (e.g., having an event emitter and listener system).

[MichaelBuessemeyer]

By just looking at the code I came across this code. This might be useful to trigger re-fetching the tracing again without reloading it 🤔

[MichaelBuessemeyer]

@philippotto Do you know a way how to know whether currently a tracing is opened (but not fully loaded), thus when the user logs in the view should be reloaded?
I cannot find any difference between the states of the store during tracing and while the dashboard is open that seems to be sufficient.

[philippotto]

@philippotto Do you know a way how to know whether currently a tracing is opened (but not fully loaded), thus when the user logs in the view should be reloaded?
I cannot find any difference between the states of the store during tracing and while the dashboard is open that seems to be sufficient.

Hm, I can think of multiple options:

• in controller.js, there is a this.state.ready flag which could be moved into the store to check whether a dataset or tracing was loaded completely.
• check whether the dataset in the store is the default dataset with which the store was initialized. I think, this could be helpful in other places, too (e.g., a sharing token request is send for the "dummy dataset" which always fails).
• set a new store attribute for "loadingFailed" in the store if the model initialization crashes.

Ok, now that I'm thinking about it a bit, I suggest the following which might be the easiest solution:
Change the ready attribute in Controller to a state attribute which can have the following values: loading, loaded and loadingFailed. Setting the value to loadingFailed can probably be done by extending the catch-handler for Model.fetch in componentDidMount. Then, you simply use the state flag to either display the brain-spinner, the tracing view or the log-in form. Consequently, the log-in form would be displayed in the center of the page (instead of opening the popover in the navbar).

What do you think?

[MichaelBuessemeyer]

@philippotto Do you think it is helpful to additionally add an error message below the loading brain when the user has logged in but the model initialization still fails or not?

with additional information

without additional text

The rest is working perfectly 🚀

[philippotto]

I'd show the text with the brain logo, but without the loading bar (since loading has finished in that case).

[MichaelBuessemeyer]

@philippotto I cleaned up this PR. So it is ready for your first review 😄

[philippotto]

Awesome stuff! I'd just tune the "either the dataset is missing..." message a bit. (1) I'd center the text and (2) I would add a "Return to dashboard" link to that message so that the user isn't "lost" in this view (the navbar is still there, but that might be not obvious enough).

[MichaelBuessemeyer]

@philippotto What do you mean by

I'd center the text and

Is this what you meant with centering?

As you can see, I also added a button to navigate back to the dashboard.

With this, I finished all your requested changes 😄. Could you please review this PR again? 🧐

[philippotto]

Awesome! Looking forward to seeing this merged.