Fixes #219

[sanguinariojoe]

As promised, I am coming with a verification code. Some notes:

• In contrast to your code @RyanDavies19 , this is running the OpenFAST tests (just the moordyn ones), and then it is automagically collecting and running them with MoorDynC.
• The tolerances cannot be so strict. Instead of rtol=2 and atol=1.9, I had to set rtol=1.5 and atol=1.5
• I am directly importing the pass_fail.py from openfast instead of vendoring it.
• It is running right now on https://github.com/core-marine-dev/MoorDyn/actions/runs/9596912401/job/26464823608. I hope it will finish just fine
• I tried that with the dev branch of openfast... Oh boy, our lines failure stuff seems to be very broken...

[RyanDavies19]

@sanguinariojoe thanks for putting this together and for the rewrite of the test approach, this looks really solid. I will work on getting more regression test cases uploaded that really test our capabilities, and I will make sure that the options sections are compatible with both versions. If you have any suggestions for those send them to me, I am planning to get them in next week. For the line failures in MD-C, that is something I need to fix. I just rewrote and fixed the MD-F ones (https://github.com/OpenFAST/openfast/pull/2214) but didn't have the time to do MD-C yet, so don't worry about that one for now.

Noted on the tolerances as well, I had just taken them from the OpenFAST defaults.

[sanguinariojoe]

I'll do that, sure thing! But it has to wait until next week. Tomorrow we celebrate midsummer! Partyyyyyyyyy

On Thu, 20 Jun 2024, 18:11 RyanDavies19, @.***> wrote:

@.**** commented on this pull request.

In tests/.mdf_verification/verify.py https://github.com/FloatingArrayDesign/MoorDyn/pull/224#discussion_r1647825242 :

• rod_state = moordyn.GetRodState(rod)
• state += np.array(rod_state).flatten().tolist()
• names += [f"rod_{rod_id}",] * len(rod_state)
• n_points = moordyn.GetNumberPoints(system)
• for i in range(n_points):
• point = moordyn.GetPoint(system, i + 1)
• if moordyn.GetPointType(point) not in (moordyn.POINT_TYPE_COUPLED, ):
• continue
• point_id = moordyn.GetPointID(point)
• point_state = moordyn.GetPointPos(point)
• state += point_state
• names += [f"point_{point_id}",] * len(point_state)
• return state, names

• +def read_outs(fpath, skiplines=2):

Add a check here for *** akin to: https://github.com/RyanDavies19/MoorDynC_ryan/blob/c572051566b119b401c623fe6c3f42d8b085dcc3/verification/verify.py#L51

OpenFAST/MD-F will sometimes have an overflow error for numbers very close to 0 (e-100) where it prints *** instead of the value

— Reply to this email directly, view it on GitHub https://github.com/FloatingArrayDesign/MoorDyn/pull/224#pullrequestreview-2130853055, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAMXKKGK7HITLKCPL6J26OLZIL5M7AVCNFSM6AAAAABJTZTF2SVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDCMZQHA2TGMBVGU . You are receiving this because you were mentioned.Message ID: @.***>

[sanguinariojoe]

BTW, I am idiot and pointed the PR against master. Can you change that to point to dev?

Otherwise I'll do that next week

On Thu, 20 Jun 2024, 18:06 RyanDavies19, @.***> wrote:

@sanguinariojoe https://github.com/sanguinariojoe thanks for putting this together and for the rewrite of the test approach, this looks really solid. I will work on getting more regression test cases uploaded that really test our capabilities, and I will make sure that the options sections are compatible with both versions. For the line failures in MD-C, that is something I need to fix. I just rewrote and fixed the MD-F ones but didn't have the time to do MD-C yet, so don't worry about that one for now.

— Reply to this email directly, view it on GitHub https://github.com/FloatingArrayDesign/MoorDyn/pull/224#issuecomment-2181056962, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAMXKKDB4ZST64USB5JG4FLZIL4YTAVCNFSM6AAAAABJTZTF2SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBRGA2TMOJWGI . You are receiving this because you were mentioned.Message ID: @.***>

[RyanDavies19]

BTW, I am idiot and pointed the PR against master. Can you change that to point to dev? Otherwise I'll do that next week … On Thu, 20 Jun 2024, 18:06 RyanDavies19, @.> wrote: @sanguinariojoe https://github.com/sanguinariojoe thanks for putting this together and for the rewrite of the test approach, this looks really solid. I will work on getting more regression test cases uploaded that really test our capabilities, and I will make sure that the options sections are compatible with both versions. For the line failures in MD-C, that is something I need to fix. I just rewrote and fixed the MD-F ones but didn't have the time to do MD-C yet, so don't worry about that one for now. — Reply to this email directly, view it on GitHub <#224 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAMXKKDB4ZST64USB5JG4FLZIL4YTAVCNFSM6AAAAABJTZTF2SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBRGA2TMOJWGI . You are receiving this because you were mentioned.Message ID: @.>

Done

Ojala que pasa bien las fiestas!