Added profiles from CAKE

[AreWeDreaming]

• Ripped out other references to profiles needs to be reviewed carefully

[orso82]

See the "options" section in the "*.json" files to see how options can be passed to mapping functions. For example, you should be able to do:

"__options__": {
  "EFIT_tree": "EFIT01",
  "PROFILES_tree": "ZIPFIT01",
  "analysis_type": "CERQUICK",
  "default_tree": "D3D",
  "fast_ece": false,
  "nref": 0,
  "revision": "BLESSED"
 },
...
"core_profiles.profiles_1d.:.electrons.density": {
  "PYTHON": "core_profiles_profile_1d(ods, {pulse}, {PROFILES_tree})"
 }

which can then be passed with

with ods.open("d3d", 133221, PROFILES_tree="mytree")
    plot(ods['core_profiles.profiles_1d.0.electrons.density'])

[orso82]

Should the ability to read ZIPFITS be retained? Are you sure there is nothing that depends on that?

[AreWeDreaming]

@orso82 Thanks for the example. I am not suggesting to removed anything, I was just confused about what to do and wanted to have an example for what I was struggling with when asking for help. I didn't realize that I need to move the zipfit mapping out of the .json file and into my new routine. I'll add an if-clause and the tree name as an argument. The Omas viewer should be able to show whatever Omas can load from zip fits.

[AreWeDreaming]

The last commit has several problems:

1. The "PROFILES_tree" extra argument has to include \', e.g. "\'ZIPFIT01\'" in order to not loose its string nature. This is not the case for "EFIT_tree" which is a very similar extra argument.
2. PROFILES_tree="\'OMFIT_PROFS\'"and PROFILES_tree="\'ZIPFIT01\'" do not provide the same fields. This seems to confuse the .json file that wants to have a fixed set of fields for each mapping function regardless of its input argument.
3. I also need uncertainties. I learned from Sean that the TDI function is error_of but I am not sure how to do this for all fields without explicitly looping. I am not sure if explicitly looping over each field might cause separate MDS+ requests for each quantity. I guess I could just create a query like this:

   query = {
           "electrons.density.data_error_upper": "error_of(\\TOP.n_e)",
           "electrons.temperature.data_error_upper": "error_of(\\TOP.T_e)",
           "ion[0].density.data_error_upper": "error_of(\\TOP.n_D)",
           "ion[0].temperature.data_error_upper": "error_of(\\TOP.T_i)",
           "ion[0].velocity.toroidal.data_error_upper": "error_of(\\TOP.V_tor_C)",
           "ion[0].density.data_error_upper": "error_of(\\TOP.n_C)"
       }

[orso82]

1. see how it's done for charge_exchange.channel for example
2. I need to see the error, if there is an error
3. Indeed you'll have to query the uncertainties separately and then assign them to the ODS

[AreWeDreaming]

Ok, so I managed to sort out all the problems I think. It seems that there were some minor errors that then became several weird problems. Now I can switch between ZIPFIT/OMFIT_PROFS without issues and the tree specification works without \". I also added the uncertainties. We'll have to revisit this when the fields for *_fit have data in MDS+. For now, I am happy with this.

[github-actions[bot]]

Stale pull request message

[smithsp]

@AreWeDreaming should remove the no-pr-activity label if this PR is still valid. If it needs further testing, let's work on it. The last automatic test failed.

[github-actions[bot]]

Stale pull request message