Signal name when logging signals to csv file

[tomarnepedersen]

When signals are logged to csv file, the naming convention seem to be: [signal name] [[signal id] [signal type] e.g. azimuth_thrust [6 real output]

When new signals are added to the model, the id may change, even though the name is the same. So after the model has been updated and compiled to fmu, the signal could have gotten a new id, which gives e.g. the following signal name in the csv file: azimuth_thrust [8 real output]

Then, when reading the csv file, and plotting signals, the code for plotting will need to be updated since the ID has changed. I would suggest to not include the ID in the signal name when logging to file.

[eidekrist]

Moved to libcosim and added "discussion needed" label.

Has anyone ever used the additional metadata ([valuereference type causality]) for anything?

[markaren]

In the very specific cases where this would be an issue, one could just ignore the metadata in the plotter. This sounds like an easy to fix issue in the plotting code.

Altough, there probably isn't that much value in keeping valuereference as you'd probably search using the name anyway.

[kyllingstad]

I think we should drop the additional metadata from the header line and just have the variable name there. In the few cases where such information is needed, it can be gotten from elsewhere.

In the very specific cases where this would be an issue, one could just ignore the metadata in the plotter.

This presumes that the user has programmatic control of the plotting. But I'd argue that's not the typical case; most will use point-and-click solutions (e.g. Excel).

[restenb]

This has (hopefully) been covered by https://github.com/open-simulation-platform/libcosim/pull/725 in the latest release, so I'm closing this.