use uuids with pycadf

I'd like to see a fix for this land before Ocata-RC. This can be moved from the milestone if it doesn't land before then, but making tests less chatty with non-useful data is great!

Fix proposed to branch: master
Review: https://review.openstack.org/426411

bumping from RC, not necessary

Patch to pycadf to loosen uuid validation - https://review.openstack.org/#/c/428543/

Reviewed: https://review.openstack.org/428543
Committed: https://git.openstack.org/cgit/openstack/pycadf/commit/?id=a4a8f46248ab1d1a68d422276df21f75b24be84c
Submitter: Jenkins
Branch: master

commit a4a8f46248ab1d1a68d422276df21f75b24be84c
Author: Gage Hugo <email address hidden>
Date: Thu Feb 2 20:01:53 2017 -0600

    Make `is_valid` more flexible with uuid validation

    In keystone, multiple domain ids are handled by concatenating
    two valid uuids together. This causes issues with pycadf validation
    and causes the following warning to be thrown:

    warnings.warn('Invalid uuid. To ensure interoperability, identifiers '
    'should be a valid uuid.')

    This appears throughout the testing logs while running keystone tests
    and there is a current attempt to remove most/all invalid uuids to
    eliminate this warning [0]. However due to the multiple domain id
    issue, this cannot be solved in keystone alone.

    The idea to allow multiple uuids that were joined together was
    mentioned before in a previous attempt to solve this issue [1].

    This change:
    - Allows 2 or more concatenated uuids to be considered valid without
    emitting a warning
    - Cleaned up the list of words that are exceptions for validation
    - Added 'default' to list of valid words in exception list
    - Added offending value to printed warning
    - Broke up test_identifiers for better clarity about which tests
    were valid. This also solves the issue of `warning_mock.called` always
    being `True` once an invalid uuid of type string was checked within
    test_identifier
    - Added test for valid exception words and extra characters that
    can be present in a valid uuid according to [2]

    [0] https://bugs.launchpad.net/keystone/+bug/1659053
    [1] https://bugs.launchpad.net/keystone/+bug/1521844
    [2] https://docs.python.org/2/library/uuid.html

    Co-Authored-By: Tin Lam <email address hidden>

    Partial-Bug: #1659053
    Change-Id: I58bba04c21c2d24fd37850c9ecc6fac99deb3fc4

Reviewed: https://review.openstack.org/426411
Committed: https://git.openstack.org/cgit/openstack/keystone/commit/?id=2be615ea9303109f11033e293c229df586e3b1e1
Submitter: Jenkins
Branch: master

commit 2be615ea9303109f11033e293c229df586e3b1e1
Author: Gage Hugo <email address hidden>
Date: Fri Jan 27 15:52:02 2017 -0600

    Fix multiple uuid warnings with pycadf

    There are multiple pycadf warnings about invalid uuids when
    running keystone tests:

    To ensure interoperability, identifiersshould be a valid uuid.
      warnings.warn('Invalid uuid. To ensure interoperability, identifiers'

    This changes multiple fixtures within default_fixtures to use valid
    uuids for their 'id' values. Also changed load_fixtures to build
    the fixtures based on the default_fixtures' 'name' value, rather
    than 'id'. Replaced many instances of invalid hard-coded ids to use
    random uuids, default_fixture ids, or 'default'.

    Co-Authored-By: Tin Lam <email address hidden>

    Change-Id: Ic4fff30c306561b71288712480f073aba1fccbde
    Closes-Bug: #1659053
    Depends-On: I58bba04c21c2d24fd37850c9ecc6fac99deb3fc4

This issue was fixed in the openstack/keystone 12.0.0.0b1 development milestone.