Fix examples by re-running with up-to-date sonata-generator

[joni-herttuainen]

Notes:

• kept the originals in v2 directory of GPFS
  • copied the directory to v3 and overwrite what was updated
• bunch of data seems to change in the re-run, due to random and things being changed since it was originally run
  • re-running again now does not change the data in the files

closes #49 and #50

[jdcourcol]

I am not sure why this has changed (in usecase 1) too.

[joni-herttuainen]

I believe that this is due to samples being drawn randomly. In the usecase config we have

# from usecase_examples/config/nodes_configuration.yaml
etype:
  type: text
  values: [ "dNAC", "dSTUT" ]

and we only draw 2 samples, and since both of them were "dSTUT" only that value is represented in the @library.

[mgeplf]

nothing jumps out at me, but there's a lot to check.

[joni-herttuainen]

Yeah, unfortunately the fields added after the original run messes up the predicted results of the pseudo-randomness even though the seed hasn't changed.

So, basically, everytime a circuit is created, the first node file that is created has the same values, until we create a field that wasn't in the original run. All the random values after that (for that circuit) will be different for rest of the fields and files.

What we could do, before this goes in, is to instantiate each of the file generators with a with a seed to ensure that if we add fields in the future, not everything is changed (such as in this PR). But that would only make sense if we foresee that we're going to have to add new fields in the future.