The qaTree.json file produced by qaCut.groovy isn't the last place this file is modified; see flowchart documentation, where the qaTree.json file is subsequently used in various "Manual QA" scripts. If we modify qaCut.groovy, we may want to modify everything downstream too.

Alternatively, we could just use your JSON <=> SQLite converters. Then we could just continue to use the JSON file internally, then when ready to upload the final version to the QADB, convert the JSON file to a SQLite file. When we need to update that SQLite file, bring it back here, convert to JSON, edit, then revert to SQLite. This would be the easier approach for now, and if we end up benefiting significantly from the change to SQLite, we could then consider refactoring everything here to just use SQLite instead of JSON.

We will also have to convert chargeTree.json, but maybe it's time to combine it with qaTree.json (which could be done in qaCut.groovy)

Sorry I forgot about chargeTree.json. I agree it is probably easiest to first use JSON internally and convert to SQLite for the QADB. If we are going to keep JSON as the internal format though, it might be easier to just keep qaTree.json and chargeTree.json separate for now since that works well already, and then we can combine them into a single SQL database with a converter script after all the manual QA. In the future, if we switch to using SQL internally then we could just combine them into a single database upstream of the manual QA.

Sounds good. So the converter could take both qaTree.json and (optionally) chargeTree.json as input and would output a single DB (and vice versa for the opposite converter). I say optionally because we are now thinking of removing chargeTree.json altogether, but that's a separate discussion; for now let's keep supporting it.

I removed the SQL database creation routine in qaCut.groovy so it should be identical to main now. I also added SQL <-> JSON conversion scripts for chargeTree.json. Converting chargeTree.json to SQL for a dataset will add all that data as a separate table called ${dataset}_chargeTree the entries of which can be cross referenced with the qaTree*.json table ${dataset} via the file and run numbers.

Holding off on this for now, since ideas for the future of QADB are under discussion.