This duplicate the analysis file in two folders. We probably should only have one, maybe with a rename.
Not sure if it worth the effort but.
Is it feasible to either?

• remove the one in output, rename the original one in run_dir (with a suffix like _input, _user or _origin)
• point the process that consume the setting after to the output one

ditto comment above ^

This duplicate the analysis file in two folders. We probably should only have one, maybe with a rename. Not sure if it worth the effort but. Is it feasible to either?

• remove the one in output, rename the original one in run_dir (with a suffix like _input, _user or _origin)
• point the process that consume the setting after to the output one

The analysis settings file in run_dir/analysis_settings.json seems to not be used by the rest of the computation.

Moved so that the user / input analysis settings file can be found in run_dir/input/analysis_settings.json for reference.

The run_dir/output/analysis_settings.json is the one with the merged settings.