[ty] Add support for using the test command emitted when a mdtest fails (#19794)

## Summary

When seeing a failed test like 

```bash
is_subtype_of.md - Subtype relation - Callable - Class literals - Classes with `__new_… (1e9782853227c019)

  crates/ty_python_semantic/resources/mdtest/type_properties/is_subtype_of.md:1810 unexpected error: [unresolved-reference] "Name `Aa` used when not defined"

To rerun this specific test, set the environment variable: MDTEST_TEST_FILTER='is_subtype_of.md - Subtype relation - Callable - Class literals - Classes with `__new_… (1e9782853227c019)'
MDTEST_TEST_FILTER='is_subtype_of.md - Subtype relation - Callable - Class literals - Classes with `__new_… (1e9782853227c019)' cargo test -p ty_python_semantic --test mdtest -- mdtest__type_properties_is_subtype_of
```

running the following now works

```bash
MDTEST_TEST_FILTER='is_subtype_of.md - Subtype relation - Callable - Class literals - Classes with `__new_… (1e9782853227c019)' cargo test -p ty_python_semantic --test mdtest -- mdtest__type_properties_is_subtype_of
```


## Test Plan

Do we have tests for the test runner? :)

crates/ty_test/src/lib.rs

@@ -58,7 +58,7 @@ pub fn run(
     for test in suite.tests() {
         if filter
             .as_ref()
-            .is_some_and(|f| !test.uncontracted_name().contains(f))
+            .is_some_and(|f| !(test.uncontracted_name().contains(f) || test.name() == *f))
         {
             continue;
         }