Currently the schema_str parts in the unit tests of the analyzer always gets formatting errors before it's actually parsed correctly.
Consider switching the schema_str's from rawstrings to https://github.com/dtolnay/indoc for a less whitespace sensitive test setup.