Tidal power performance

[jmcvey3]

Tidal power performance capabilities specified in IEC/TS 6200-200.

TODOs left:

• [x] Add "assertions" to input arguments
• [x] Add tests
• [x] Flesh out jupyter notebook example
• [x] Extend profiles to seafloor and surface

I would appreciate a second set of eyes and would be happy to discuss interpretation of the standards. The mathematical section is not well written, and I was relying on information conveyed in the text and plots in the appendix to figure out what was being asked. (My TS copy is heavily loaded with comments and corrections to keep track of things).

Also includes bugfixes, docstring clarifications, and useability edits for the Tidal module in general that I came across from a users standpoint. Changelog as follows:

• Bugfixes:

  • Standardized function inputs to accept both numpy.ndarrays and pandas.Series (some functions already have this implemented).
  • Fixed assertions so that if a numpy array is supplied, the following assertions that rely on pandas don't fail (some functions already have this implemented).
  • Added a (temporary) fix so that assertions can handle nans
  • Fixed river.performance capture area calculation the multiplies by instead of dividing by 4.
  • Fixed a misnamed variable in tidal.performance.principal_flow_directions.
  • Corrected the "Returns" statement in the tidal.performance.principal_flow_directions docstring and added a "Notes" section.

• API/useability

  • Added 'ax' argument passthrough to tidal.graphics.plot_rose and tidal.graphics.plot_joint_prob_distribution to enable more complex matplotlib capabilities outside of the function.
  • Added check and updated docstring in tidal.performance.principal_flow_directions to make sure directions are between 0 and 360 degrees, which the function requires.

[jmcvey3]

@ssolson This should be ready for review

Also, the tests are failing because of something I thought it an error in river.performance.py capture area functions: the area of a circle is 1/4 pi d^2, not sure if the original 4 pi d^2 is an error or not

[jmcvey3]

@ssolson I just remembered there are some changes in the dolfyn PR (namely to the function/shortcut dolfyn.velocity.U_dir) that will affect this one. I'll try to wrap the dolfyn PR soon here so we can merge these back to back.

[jmcvey3]

@ssolson This one should also be good to merge in if you want to give it one more review

[ssolson]

@jmcvey3 merging #232 created a conflict with the adcp example.

I created PR https://github.com/jmcvey3/MHKiT-Python/pull/5 on your branch which resolves the issue.

[jmcvey3]

Thanks @ssolson - made some last fixes to the notebooks

[ssolson]

All tests pass excluding hindcast. Hindcast functionality was not changed in this PR.