Better Typing for Tool Execution Plumbing

[jmchilton]

The latest mypy upgrade (xref #18608 ) is going to force us to clean up some of the typing around tool execution and in particular conflicting type signatures around the execute method of ToolAction subclasses. That PR is already getting large and I think cleaning up the typing around tool execution is a bit intricate - so I was hoping to do it in its own PR ahead of that one.

I think SetMetadataToolAction.execute was doing the most mangling of the type execute signature. Adding a overwrite argument. I have tried to refactor access to the internals of that so that consumers can call them with an overwrite argument that isn't through the generic execute method. Additionally, I've tried to make set_output_hid a class variable instead of an argument since it seems we were mostly changing it via changing the default arguments to subclass execute methods - which is problematic. (This didn't work - I've reworked everything to put set_output_hid back into the argument list.)

How to test the changes?

(Select all options that apply)

• This is a refactoring of components with existing test coverage.

License

• I agree to license these and all my past contributions to the core galaxy codebase under the MIT license.

[jmchilton]

It looks like I need to undo the set_output_hid changes based on the failing tests. Everything is more verbose but I think I am much more confident in changing this code in the structured tool state branch with all this typing in place.

Done. I think it is in good shape now.