Filter validation

[m0nkeyc0de]

Fixes: #695

Validation of filter() kwargs is optionnaly done in pynetbox:

• it can be enabled globally with strict_filters=True when instanciating the API object
• it can be enabled/disabled per request by setting strict_filters on the filter() or get() methods
• if enabled a ParameterValidationError is raised
• if disabled a line is printed in stdout

[jnovinger]

@m0nkeyc0de , looks like your changes have some linter errors.

[m0nkeyc0de]

@jnovinger, my bad, I'm too used to pre-commit hooks who do that automagically

Linting is done in commit 73d9821

[m0nkeyc0de]

@jnovinger code updated regarding your comments (except for the yet unresolved conversation)

I also mentionned the purpose of strict_filters in the README file.

[jnovinger]

This looks great, @m0nkeyc0de, thanks for the iteration and improvements!

I appreciate the separation of the strict validation tests in to a separate test class. Also, I did leave one comment on a completely optional approach to setting up mocks for the strict validation tests--happy to merge this with or without that.

[jnovinger]

This is totally optional, but you can also use patch as a function/method or class decorator, allowing you to reduce boilerplate and indentation in test code.

It would look something like this:

@patch("pynetbox.core.query.Request._make_call", return_value=Mock())
class StrictFilterTestCase(unittest.TestCase):
    def test_filter_strict_valid_kwargs(self, mock):
        # the rest of the test
        ...

Again, totally optional, but thought it was worth sharing.

[m0nkeyc0de]

Thanks for sharing this trick. I figured out that I wasn't using this mock variable as I'm only watching for the Exception beeing raised or not in my tests.

This unused boilerplate is now removed.