[DOCS] MiniDataAPI Spec suggestions

**Issue Type**
- [x] Existing documentation (e.g. typos, errors, outdated information)
- [ ] Suggest new documentation that is currently missing

**Suggested Improvement**
@pydanny I just read through the new MiniDataAPI Spec docs. It was a great read, thanks for your efforts! I made some notes on a few things that occurred to me. Hope they are useful!\

_Edit by @pydanny: Added checkboxes so I can address each of these bullets in turn_

- [x] I noticed that in the doc and in some of Jeremy's videos both `db = Database('test.db')` and `db = database('test.db')` functions are used interchangeably to connect to the db. Are they both exactly the same function, and is either one preferred? In the doc should we use one of them only or clarify both are OK?
- [ ] Early on in the doc we learn how to extract a Dataclass version of tables. The method is clear but I wasn't exactly sure _why_ this might be beneficial?
- [ ] When we do a table transform (e.g. to add a new table field) is it also OK to change the `pk` to a different field? If so, do we use table transforms to update the primary key field, even if no other columns are added?
- [ ] Related to the above, what other transforms are available? Can we remove fields too for example?
- [x] In the 2nd paragraph of the **.xtra()** section I think there is a type. 'v ia' should be 'via'?
- [ ] Also in the **.xtra()** section, should `2 in todos` be formatted as code, and show an output of `False`, just like `1 in todos` is?
- [ ] Is it OK to set more than one field at a time with `.xtra()`? And can we call `.xtra()` again at any time to update the filtering?
- [ ] Is it possible to _unassign_ a filter set with `.xtra()`.

**Confirmation**
Please confirm the following:
- [x] I have checked the existing issues and pull requests to ensure this documentation issue hasn't been reported before.
- [x] I have read the project's documentation and am confident this issue or suggestion is valid.