Ondřej Samohel

We should move with this forward. So, what about one class for hooks (`PrePostLoadHook`) with `pre_load` and `post_load` methods - hook it on loader class name and monkey-patch loader `.load()`...

good work, something I had in mind for some time. Rollback should be done by patch executor code for whole db. And it should be locked by first Pype instances...

Maybe as part of the anatomy. Something like `search_path_fragment` 🤔 ```yaml # default.yaml search_part_fragment : "1:2" work: folder: "{root[work]}/baz/{project[name]}/{hierarchy}/{asset}/work/{task}" ``` now in python: ```python >>> search_path = "[{}]{}".format(anatomy.root_environments().keys()[0], anatomy.get_search_fragment()) >>>...

> @antirotor I'm not sure this is actually still required. Haven't hear anything related to it for a long time. What do you think? I think it is still valid,...

I would add to this that during that refactor, we should cleanup legacy code and switch to better handling of render setup. It seems we are hitting legacy api limits.

My idea was to split it into more files to make it more managable even if they will at the end link back to the lib. The file is huge...

> My main issue with the lib is that unless you know what is available, you end up writing duplicate code while developing, then trying to refactor to use the...

That's because now it is all over the place. I think the clearest and most "pythonic" way would be to move things related to specific workflows to separate files -...

> I am trying to create one representation for the outputs from multiple operators, but it seems the `integrate.py` deosn't allow me to and error out. This is indeed the...

adding screenshot of the behaviour: