Feedback from working group 3

[TeresiaOlsson]

The examples didn't work so well for us so we focused on looking at the code and compare to features in MML that we would like to have.

Configurations:

• A long yaml file is not easy to use. In the examples there is also data which is repeated in several files which makes it difficult to maintain and keep in sync. It would be better to separate the config in several files.
• At the moment different parts of the configuration is mixed in the same file. For example, device information which doesn't change often is in the same file as measurement settings which might change more often.
• We also need a good option to handle configuration for different optional modes. This is not only different optics but some labs also use this feature for different operational runs during the year.
• We need a way for applications to update when the configuration is changed so all parts are kept in sync.
• We also would like to have some functionality to browse and modify the configuration after it has been loaded.
• The yaml file is difficult to generate and modify for new users. It would be nice with an utility GUI which helps you to visualise and change the yaml file without having to do it directly in the file.

Conversions:

• We need some way to define the native view (hardware/physics) for the devices in the configuration so if you are using pyAML in that mode for backends that only need that mode there is no need for configuring conversions.
• We want the possibility to be able to use hardware or physics units in measurements.
• We not only need conversions when setting a device but also utility functions that can be used without setting a device, for example to easily calculate which RF frequency change corresponds to which energy change.

Saving data:

• We need a standardized data model for saving data. This could consist of a default part which is always included and some application dependent part which is different.
• For data that is saved as hdf5 files it would be nice with hdf5 viewer which is better than the official one which allows you to click on a data file, see which data is inside and visualize it without having to start python, load into pandas, plot yourself etc.