Blender: Dynamic workflow using linked libraries and overrides

[kaamaurice]

Brief description

• Use collections and overrides for the instances.

Description

• Change Blender plugins to work with a collection instead of an empty to store custom properties of the Avalon instance.

• Discussion : https://github.com/pypeclub/OpenPype/discussions/2598

• Based on the work of jojoba106 https://github.com/pypeclub/OpenPype/pull/2771

• Squashed and rebase from : https://github.com/pypeclub/OpenPype/pull/3171

Resolve #2890

[ynbot]

Task linked: OP-2598 Demo generator

[Tilix4]

I'm wondering if we shouldn't take advantage of families for the layout/anim issue we have. Maybe the layout family should be updatable in the Animation task and it'll define how things are placed. It should be dissociated from the Layout task which is only meant to create the firsts versions of the layout family instance. anim family instances must always contain only actions from Blender, and are most of the time created in the Animation task but could be updated in the Rendering one for example.

Staging information are stored into the layout family and can be updated also in the Animation task.
To avoid confusion between the task and the family, maybe the layout family should be renamed to staging.

[iLLiCiTiT]

PR adds OpenColorIO-Configs submodule to deprecated place.

[Tilix4]

The PR cannot be merged because of changes in integrator plugin which breaks publish workflow conversation here. We can't afford to add ability to define representation id before integrator plugin due to many reasons. Also the branch need update from develop branch.

We would like to have that changed so we could merge the PR.

@iLLiCiTiT I'm working on the seamless integration #3902 which includes the changes you require and a completely universal way to handle datablocks. It deeply refactors the plugin also. I'm not far from having it finished, therefore we'll be able to merge the current branch once my work have been merged into it!

[iLLiCiTiT]

I'm not far from having it finished

Happy to hear :)

[BigRoy]

I've seen steady commits to this PR but I am not sure what the actual status is. Any brief status update of where it's at?

[Tilix4]

I've seen steady commits to this PR but I am not sure what the actual status is. Any brief status update of where it's at?

@BigRoy, we are closer and closer. The stack of remaining things I'd like to make clean is very thin, but as you know it, polishing takes more time as it is not about functionality and gets constantly put behind emergencies... Remaining:

• [ ] Refactor for the main Loader class for Blender to share as much code as possible between loaders (fbx, abc, blend...)
• [ ] Consistent operators and variables naming
• [ ] OP user documentation writing

Therefore, I don't know what you're expecting, but if for testing, you may start blindly, if for implementing loaders, I'd recommend to wait for the refactor to be achieved.
Do you have any specific question?

[BigRoy]

Thanks for the thorough reply. The question came forth from currently not using Blender much in production and wanting to adopt it more. The work done on this PR seemed massive and looked like one 'better to wait for' with regards to potential backwards compatibility later.

It also was just a general interest question for an ETA of when this might be ready for review.

[BigRoy]

Just looping back to your reply from a little over a month ago @Tilix4 here.

we are closer and closer. The stack of remaining things I'd like to make clean is very thin, but as you know it, polishing takes more time as it is not about functionality and gets constantly put behind emergencies... Remaining:

• Refactor for the main Loader class for Blender to share as much code as possible between loaders (fbx, abc, blend...)
• Consistent operators and variables naming
• OP user documentation writing

Where are we now? If this is close to a mergable state I'd say we should Resolve conflicts, mark it Ready for review and maybe have someone demo it briefly (or do a screen recording with some explanation). Once that's there I'd say some testers maybe can put effort into helping out with OP user documentation writing maybe?

Would also be good to know to mention what expected issues we might face and give us context on what you're expecting should work (e.g. what Blender versions have you tested with, etc.). This PR is missing a short explanation on what should be tested for this PR. It'd be great to get some few bullet points on what to investigate as an outside tester and we can go from there.

[Minkiu]

Hey there @kaamaurice is this PR still valid?

[Tilix4]

Hey there @kaamaurice is this PR still valid?

Hey @Minkiu, Kevin will take a long break, he is no longer in charge for this dev. I still have some refactor to do and take @iLLiCiTiT's review into account, and then I'll likely propose a new PR.

[mkolar]

@kaamaurice we're closing this PR as it is now stale. If there is any functionality from this you'd like to contribute again, feel free to re-open this again, targeting AYON