Create schema for storing ensemble transition data
Currently we are saving some data when doing updates, but we are not always saving data indicating how an ensemble is created. We should create a data model for saving ensemble transitions, with information about how an ensemble is created, for example, for es-mda we should store:
- Weights
- Observation scaling factors
- Type of transformation
- Observation report
- Update settings (outlier thresholds, localization settings, etc)
For Everest this would be slightly different, but for most cases it is probably some tables with some metadata.
For EnIF we need to store some matrices, some of which are sparse (may be boiled down to non-sparse more concise datasets which can be used to create the same plots, remains to be seen)
Blocked by https://github.com/equinor/ert/issues/11477 . Moving some info to runmodels will alleviate us from having to store it as "ensemble transition"