Refactor Tests

[sco1]

Some aspects of the test suite are written in a way that makes them fairly fragile and difficult to both maintain and add additional cases to.

As an example, test_column_line_numbers relies on explicitly coded locations matching the formatting of column_line_numbers.py, making updates a manually intensive, error-prone process across multiple files.

This issue proposes investigating potential alternatives for specifying test cases & their expected returns to provide a more robust & maintainable test suite. One thought is to split monolithic source code blocks into more discrete components, allowing them to be more easily colocated with the test parameters.

Using the above example, our source file would be split into something along the lines of the following:

a_test_case = """def bar(x):
    pass
"""
# Expected yielded code definition

and

another_test_case = """@property
@some_decorator
@some_other_decorator
def foo(
    x,
    y
):
    pass
"""
# Expected yielded code definition

While this would potentially make test case specification more verbose overall, I believe it's significantly outweighed by the ease of making changes to the test suite; rather than having to adjusted every column & line number in the file, only the specific case needs to be adjusted.

With this in mind, here first-glance rundown of what's proposed to be looked at:

To be investigated:

• test_column_line_numbers
• test_fully_annotated
• test_parser
• test_type_comments
• test_variable_formatting

Seems ok(ish):

• test_classifier
• test_flake8_format
• test_object_formatting
  • The contents of printed_object_formatting.py could potentially be moved into the test file itself

[sco1]

This has been actioned & added to #40