Inconsistent type definition and test suites for `mlflow`

[yan-foto]

tl;dr

The type definition for mlflow implicitly declares the name component to be case-sensitive. However, the test suite contains tests that expect names that are not in lowercase to be normalized to lowercase.

Before submitting a PR, I need feedback to know if the definition needs to be corrected or the tests.

Details

All PURL components for type mlflow are implicitly case-sensitive, because:

1. The type definition does not explicitly set case_sensitive for name, and version to false (namespace` is prohibited), and
2. The default value for case_sensitive is true according to the Type Definition schema

At the same time, a number tests explicitly expect that normalization converts names to lowercase, for example, the very first test case:

purl-spec/tests/types/mlflow-test.json

Lines 4 to 12 in 710767b

	{
	"description": "MLflow model tracked in Azure Databricks (case insensitive). Roundtrip an input purl to canonical.",
	"test_group": "recommended",
	"test_type": "validate",
	"input": "pkg:mlflow/CreditFraud@3?repository_url=https://adb-5245952564735461.0.azuredatabricks.net/api/2.0/mlflow",
	"expected_output": "pkg:mlflow/creditfraud@3?repository_url=https:%2F%2Fadb-5245952564735461.0.azuredatabricks.net%2Fapi%2F2.0%2Fmlflow",
	"expected_failure": false,
	"expected_message": null
	},