Much of atomate2 is set up so that result documents (i.e., emmet schema task docs) are written to a central database - whether that’s MongoDB, a JSON file, etc.
However, is there a way to duplicate the results of calculations in the run directory itself? e.g., I run 3 structure relaxations, and each relaxation has the task document written to its corresponding run directory. I use the word “duplicate” because I imagine jobflow still needs the central database to manage workflows, but having some results in the run directory itself would simplify conventional ways of analyzing runs.