assert the timeseries-twa algorithm more leniently

[kristjanvalur]

Pull Request check-list

Please make sure to review and check all of these items:

• Does $ tox pass with this change (including linting)?
• Do the CI tests pass with this change (enable it first in your forked repo and wait for the github action build to finish)?
• Is the new or changed code fully tested?
• Is a documentation update included (if this change modifies existing APIs, or introduces new ones)?
• Is there an example added to the examples folder (if applicable)?
• Was the change added to CHANGES file?

NOTE: these things are not required to open a PR and can be done
afterwards / while the PR is open.

Description of change

The unittests of the timeseries twa algorithm were assuming too much knowledge of edge cases. In particular, the value
of the right edge of the interval was badly predictable. The unittest should really not be testing the implementation of the
server, rather ensure that the python api invokes it correctly. Changed the test to be more forgiving, since the unit tests
were failing here.

[kristjanvalur]

No idea why redismod, test_alias_basic now fails. These tests sure are flakey sometimes.

[kristjanvalur]

Closing, #2358 is a duplicate which was merged.