[ty] Implement patterns and typevars in the LSP

[Gankra]

Summary

This is the final goto-targets with missing goto-definition/declaration implementations!
You can now theoretically click on all the user-defined names in all the syntax. 🎉

This adds:

• goto definition/declaration on patterns/typevars
• find-references/rename on patterns/typevars
• fixes syntax highlighting of *rest patterns

This notably does not add:

• goto-type for patterns/typevars
• hover for patterns/typevars (because that's just goto-type for names)

Also I realized we were at the precipice of one of the great GotoTarget sins being resolved, and so I made import aliases also resolve to a ResolvedDefinition. This removes a ton of cruft and prevents further backsliding.

Note however that import aliases are, in general, completely jacked up when it comes to find-references/renames (both before and after this PR). Previously you could try to rename an import alias and it just wouldn't do anything. With this change we instead refuse to even let you try to rename it.

Sorting out why import aliases are jacked up is an ongoing thing I hope to handle in a followup.

Test Plan

You'll surely not regret checking in 86 snapshot tests

[astral-sh-bot]

Diagnostic diff on typing conformance tests

No changes detected when running ty on typing conformance tests ✅

[astral-sh-bot]

mypy_primer results

Changes were detected when running on open source projects

beartype (https://github.com/beartype/beartype)
- beartype/claw/_package/clawpkgtrie.py:66:29: warning[unsupported-base] Unsupported class base with type `<class 'dict[str, PackagesTrieBlacklist]'> | <class 'dict[str, Divergent]'>`
- beartype/claw/_package/clawpkgtrie.py:247:29: warning[unsupported-base] Unsupported class base with type `<class 'dict[str, PackagesTrieWhitelist]'> | <class 'dict[str, Divergent]'>`
- Found 500 diagnostics
+ Found 498 diagnostics

No memory usage changes detected ✅

[Gankra]

This one is actually useless I think (I added it while thrashing and trying to get imports rename/reference working... it did not work).

[MichaReiser]

Should we remove it then? Seems easy enough to add once we know why we need it

[MichaReiser]

Nice, this actually streamlines things :)

[MichaReiser]

Nit: You could reduce the boilerplate here, I think, by doing something like:

let definitions = match self {
    GotoTarget::Expression(expression) => {
        definitions_for_expression(model, *expression)
    },
    GotoTarget::FunctionDef(function) => {
        Some(vec![ResolvedDefinition::Definition(
            function.definition(model),
        )])
    },
    ...
};

definitions.map(Definitions)

[MichaReiser]

Should we add a record_name method or at least rename the struct (or update its documentation) to explain that, yes, it's mostly expressions but sometimes it's names too

[Gankra]

This is probably a good idea, I'll integrate it into some followup work I have planned that will involve more reckoning about ""expressions""

[MichaReiser]

Should we remove it then? Seems easy enough to add once we know why we need it