Address fourth part of 451: Make Key constructor easier to use by taking positional args.

[dhermes]

Make Key constructor easier to use by taking positional args.

Addresses fourth part of #451.

NOTE: This has #459 as a diffbase.

[coveralls]

Coverage remained the same when pulling c17800d on dhermes:fix-451-part4 into b343acf on GoogleCloudPlatform:master.

[tseaver]

c17800d do we need to expose flat_path as a public method (it isn't tested explicitly, for instance). It looks as though we only use it in _make_parent (where we could just use the attribute).

Other than that, LGTM.

[dhermes]

RE: flat_path, I initially added it because I thought we'd need it for things like

>>> partial_key = Key('Foo', 1, 'Bar')
>>> new_path = partial_key.flath_path + (1000,)
>>> entity = Entity(*new_path)

This is no longer the goal, but ISTM the contents of flat_path will be much more useful to users than the contents of path as a list of dicts.

(Through some discussions in #336 and elsewhere it seems it's best to just have Entity.__init__ take a key as a positional arg, since we made the Key constructor easier to use already.)

[coveralls]

Coverage remained the same when pulling 03f5dd8 on dhermes:fix-451-part4 into 8cb41f6 on GoogleCloudPlatform:master.

[dhermes]

@tseaver I added 16760fb since your review because I realized we had false positives from things like

>>> datastore.key.Key('Foo', None, dataset_id='foo')
<Key[{'kind': 'Foo'}], dataset=foo>
>>> datastore.key.Key('Foo', None, 'Bar', 10, dataset_id='foo')
<Key[{'kind': 'Foo'}, {'kind': 'Bar', 'id': 10}], dataset=foo>

[tseaver]

Ugh, the rebase ruins the commit hash in your comment.

If you don't think we need to pull the flat_path property, go ahead and merge.

[coveralls]

Coverage remained the same when pulling 676b20c on dhermes:fix-451-part4 into cfea8e9 on GoogleCloudPlatform:master.

[dhermes]

@tseaver I agree having flat_path and path is confusing, I just am not sure which will be more useful going forward.