Make MaybeTree the main Python AST entrypoint for constructing the syntax tree

[JCZuurmond]

Changes

Make MaybeTree the main Python AST entrypoint for constructing the syntax tree:

• Move the @classmethods and @staticmethods that construct a MaybeTree from the Tree to the MaybeTree class as this uses the @classmethod properly by returning the initialization of the for cls argument
• Make the @classmethod that normalizes the source code before parsing the only entrypoint by removing the other @classmethod that does not normalize the source code before parsing to enforce normalization (and resolve the linked issues below)
• Rename normalized_parse to from_source_code to match the commonly used naming for classmethods within UCX
• Remove MaybeTree's methods walk and first_statement as those are repetitions from Tree's methods

Linked issues

Resolves #3457
Resolves #3213

Functionality

• modified Python linting related code

Tests

• added unit tests

[JCZuurmond]

Test is removed as we now always parse the source code before constructing the tree

[github-actions]

✅ 175/175 passed, 1 flaky, 16 skipped, 3h57m54s total

Flaky tests:

• 🤪 test_hiveserde_table_in_place_migration_job[migrate-external-tables-ctas] (4m39.483s)

_{Running from acceptance #8104}

[JCZuurmond]

Test is moved to test_python_ast.py

[asnare]

LGTM, nothing that prevents merging. A few comments, but more food for thought and definitely not blocking.

[asnare]

I think this can remain a class method?

[JCZuurmond]

yes, maybe, it does not make sense to me though. This is a protected method thus it should not be reached from outside the class, thus losing the benefit of calling the method without initializing the class, like Tree._walk. Furthermore, it is only called from the Tree.walk method, which is a regular method already, so why should this not be a regular method?