/**
 * A container for data that is specific to a session.
 * All components may use this storage to associate arbitrary data with a session.
 * <p>
 * Unlike a cache, this session data is not subject to purging. For this same reason, session data should also not be
 * abused as a cache (i.e. for storing values that can be re-calculated) to avoid memory exhaustion.
 * <p>
 * <strong>Note:</strong> Actual implementations must be thread-safe.
 *

So, you can use `user_agent` as you normally would in Python code, instead of needing to capitalize the first letters as `User_Agent` or something similar.

If for some reason you need to disable automatic conversion of underscores to hyphens, set the parameter `convert_underscores` of `Header` to `False`:

//// tab | Python 3.10+

```Python hl_lines="10"

test:pip_venv --python_path="/bazel_pip/bin/python3"
test:pip_venv --define=no_tensorflow_py_deps=true
test:pip --config=pip_venv
# Yes, we don't exclude the gpu tests on pip for some reason.
test:pip_filters --test_tag_filters=-nopip,-no_pip,-no_oss,-oss_serial,-benchmark-test,-v1only,-no_aarch64,-no_oss_py38,-no_oss_py39,-no_oss_py310

	disksToHealCount := 0
	for i := range availableDisks {
		yes, reason := shouldHealObjectOnDisk(errs[i], dataErrsByDisk[i], partsMetadata[i], latestMeta)
		if yes {
			outDatedDisks[i] = storageDisks[i]
			disksToHealCount++
		}

		driveState := ""
		switch {
		case reason == nil:
			driveState = madmin.DriveStateOk
		case IsErr(reason, errDiskNotFound):
			driveState = madmin.DriveStateOffline

			t.Errorf("Test %d: %s: Expected to fail with error \"%s\", but passed instead.", i, instanceType, testCase.expectedError.Error())
			continue
		}
		// Failed as expected, but does it fail for the expected reason.
		if actualErr != nil && actualErr != testCase.expectedError {
			t.Errorf("Test %d: %s: Expected to fail with error \"%v\", but instead failed with error \"%v\" instead.", i, instanceType, testCase.expectedError, actualErr)

  // one status to another
  // +optional
  optional k8s.io.apimachinery.pkg.apis.meta.v1.Time lastTransitionTime = 3;

  // reason is the reason for the condition's last transition.
  // +optional
  optional string reason = 4;

  // message is a human-readable explanation containing details about
  // the transition
  // +optional
  optional string message = 5;
}

  // one status to another
  // +optional
  optional k8s.io.apimachinery.pkg.apis.meta.v1.Time lastTransitionTime = 3;

  // reason is the reason for the condition's last transition.
  // +optional
  optional string reason = 4;

  // message is a human-readable explanation containing details about
  // the transition
  // +optional
  optional string message = 5;
}

And you don't need to declare anything else for that parameter, so you don't really need to use `Query`.

But you still need to use `Path` for the `item_id` path parameter. And you don't want to use `Annotated` for some reason.

Python will complain if you put a value with a "default" before a value that doesn't have a "default".

But you can re-order them, and have the value without a default (the query parameter `q`) first.

        return APP_TYPE;
    }

    @Override
    protected OptionalThing<String> myUserType() { // for framework
        return OptionalThing.of(USER_TYPE); // same reason as getUserBean()
    }

    // ===================================================================================
    //                                                                          Validation

    public void injectTransformedArtifacts(RepositorySystemSession session, MavenProject project) throws IOException {
        if (project.getFile() == null) {
            // If there is no build POM there is no reason to inject artifacts for the consumer POM.
            return;
        }
        if (Features.consumerPom(session.getUserProperties())) {
            Path buildDir =