Improve position coordinate transformation and add json reader / writer

[GernotMaier]

This PR adds a couple of improvements to the position coordinate transformations:

• add possibility to do a coordinate conversion of single array element positions in the simtools-db-json format (application will understand the generic json format and write it also; see tests/integration_tests/config/convert_geo_coordinates_of_array_elements_ground_to_utm_json.yml
• generalized the writing of simtools-db-json format-style files and move the function from simtel_config_reader to model_data_writer (simple move; no change in functionality
• renamed application to convert coordinates (from misleading print_array_elements to convert_geo_coordinates_of_array_elements
• improvement of docstrings to follow numpydoc standards

Add a small improvement to the integration tests to the function of compare_json_files: for simtools-db-type json files where a list of float is encoded as string (similar to sim_telarray; e.g., 5.5 3.2 7.6). Used np.all_close to compare those floats.

Closes #924

[tobiaskleiner]

Maybe you can add here that the z coordinate (altitude) is the relative coordinate with the additional corsika sphere center added. (not the ground level)

[GernotMaier]

Good point. I've added one sentence to make this clearer:

Altitude is the height of the elevation rotation axis (plus some possible mirror offset).

[VictorBarbosaMartins]

Hi @GernotMaier. I provide some comments here. In general I think it is already in a good state.

[VictorBarbosaMartins]

Not passing any --input leads to an error that is not yet dealt with in the code, I suggest including a check to avoid this:

simtools-convert-geo-coordinates-of-array-elements --export ground --select_assets LSTN --site South
Traceback (most recent call last):
  File "/workdir/env/bin/simtools-convert-geo-coordinates-of-array-elements", line 8, in <module>
    sys.exit(main())
             ^^^^^^
  File "/workdir/external/simtools/simtools/applications/convert_geo_coordinates_of_array_elements.py", line 149, in main
    json_type = args_dict.get("input", "").endswith(".json")
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'endswith'

[GernotMaier]

I am actually the opinion that it is dealt with: you get a an error message telling you that --input should not be None.

I am aware that we have to discuss on the command line configuration of the tools and decide if we need to work the "required" flags or not. Most tools are in the meanwhile quite complex, and parameters interfere. Either we put a effort in getting this right (which is very hard) or we put some trust in the user. This is a follow up item.

[VictorBarbosaMartins]

We actually do not get an error message telling you that --input should not be None. On the contrary, we get an error that a method we tried to use does not exist, which is an issue. These messages are not the same. A simple try and except statement with the message "--input should not be None" replacing the error message would already correct this.

[VictorBarbosaMartins]

It is not clear to me what "plus some possible mirror offset" means. I think this refers to a possible shift due to the mirror shape, but it does not inform whether the shift is already included in the model, or how much is the shift? Should the user worry about this?

[GernotMaier]

There is a parameter called "mirror_offset", this is what is mean it. I think "mirror offset" is a good description for the offset of the mirror. Do you have a better suggestion?

[VictorBarbosaMartins]

My better suggestion is to remove (plus some possible mirror offset). If this is an information the user needs, then I would pass the complete information, otherwise I think this piece of information is not needed here. This confuses more than adds to the docstring.

[GernotMaier]

@VictorBarbosaMartins - thanks a lot for the review. I have picked the items related to this PR.

[GernotMaier]

There is a parameter called "mirror_offset", this is what is mean it. I think "mirror offset" is a good description for the offset of the mirror. Do you have a better suggestion?

[GernotMaier]

I am actually the opinion that it is dealt with: you get a an error message telling you that --input should not be None.

I am aware that we have to discuss on the command line configuration of the tools and decide if we need to work the "required" flags or not. Most tools are in the meanwhile quite complex, and parameters interfere. Either we put a effort in getting this right (which is very hard) or we put some trust in the user. This is a follow up item.

[VictorBarbosaMartins]

Ok!

[ctao-dpps-sonarqube]

• 72.80% Coverage on New Code (is less than 80.00%)

Analysis Details

2 Issues

• 0 Bugs
• 0 Vulnerabilities
• 2 Code Smells

Coverage and Duplications

• 72.80% Coverage (91.80% Estimated after merge)
• 0.00% Duplicated Code (0.10% Estimated after merge)

Project ID: gammasim_simtools_AY_ssha9WiFxsX-2oy_w

View in SonarQube