Currently the ertobs file is only containing the training observations, the yaml file always contains everything.
I suggest we, by default, create for both formats a training, test and complete file.
I also suggest to make a python test which checks that the data in both outputted formats is identical, as we now have had multiple times issues with this (see for example #222).
So the tasks:
[ ] Create a train, test and complete ertobs and yaml file by default
[ ] Create a test which verifies data written to a yaml and ertobs file is identifical
Currently the ertobs file is only containing the training observations, the yaml file always contains everything.
I suggest we, by default, create for both formats a training, test and complete file.
I also suggest to make a python test which checks that the data in both outputted formats is identical, as we now have had multiple times issues with this (see for example #222).
So the tasks: