Immutable Resources fix

[Trashtalk217]

Objective

Fixes #24285.

Solution

Add UnsafeWorldCell::get_resource_assume_mutable, World::modify_resource and World::modify_resource_by_id and updated the migration guide.

Testing

The new methods are essentially identical to their component counterparts.

[Trashtalk217]

I don't know where all the migration guides have gone to, but for future reference: I'd like to add the following to the resources_as_components migration guide:

If it's not possible to add the bound, use either `UnsafeWorldCell::get_resource_mut_assume_mutable`, `UnsafeWorldCell::get_resource_by_id` and make sure that the safety conditions are satisfied. 
There's also `World::modify_resource` and `World::modify_resource_by_id`, that behave exactly like their component counterparts.
Lastly, it's also possible to use the `*_assume_mutable` component methods for this purpose, but you'd first have to retrieve the resource entity from `ResourceEntities`.

And also the PR number on the top.

[kfc35]

The migration guides are in the release-0.19.0 branch https://github.com/bevyengine/bevy/tree/release-0.19.0/_release-content/migration-guides

I would keep this PR separate for logic fixes that should also be in main and then to open a separate one to update the migration guide specifically on the release branch

[urben1680]

Is there a reason you did not add World::get_resource_assume_mutable(_by_id) like they exist for components at EntityWorldMut (1, 2)?

EDIT: ah you suggest to first retrieve the entity in the migration guide. Dunno if it would hurt much to just offer a resource method.

[chescock]

Yeah, having versions of these APIs for resources seems important!

[chescock]

Should we change get_resource_mut to delegate to this, like how UnsafeEntityCell::get_mut just calls get_mut_assume_mutable?

[kfc35]

looks good to me! i agree that get_resource_mut should delegate to the new method