reset currentmodule to traits.has_traits so Property Factory function is correctly documented

[aaronayres35]

fixes #149

This PR simply makes a 1 line addition to the documentation by resetting currentmodule to traits.has_traits instead of traits.adaptation.api so that the Property Factory Function is correctly documented. After building the docs this is what you now see:

Excuse the highlighted "Property" in the screenshot.

Checklist
- [ ] Tests
- [ ] Update API reference (docs/source/traits_api_reference)
- [ ] Update User manual (docs/source/traits_user_manual)
- [ ] Update type annotation hints in traits-stubs

[mdickinson]

Thanks for the PR. I'm not sure that this is the right fix: the use of currentmodule in the first place feels a little fragile, in that if we move documentation sections to a different file, we have to remember to check and possibly alter those uses. Is there any sane way to avoid that use in the first place? E.g., in the .. function declaration, can we specify the full path to the function somehow?

[mdickinson]

Ah, apparently there's a canonical argument to the .. function directive, but it's not available until Sphinx 4.0 (which is not yet released).

Okay, so until we have that available, maybe the only sane way is to set currentmodule immediately before any and all .. function directives (at least outside the API documentation). Or we find a way to rewrite so that we don't need those directives.

[mdickinson]

LGTM as a quick fix, though I'd still be interested to know whether there are cleaner solutions.

However, I think that's a question for a separate issue/PR, and affects more than just this one case - there are various places where we're using the .. method, .. function and .. class directives within the user documentation, which feels a bit wrong when those functions, classes and methods are already being documented in the API documentation.