Douglas Blank

Oh yeah, the low-level code requires a DB upgrade. I put a draft of that code here: https://github.com/gramps-project/gramps/blob/e1cab51a2ab4466cdc09cf856e3365242bb643af/scripts/unpickle.py

And every place that saves the pickle, we'll have to update that too. (Then some day we can get rid of the pickled blobs completely).

Sounds good!

Well, Business Logic appears in code everywhere! In views, gramplets of all kinds, and filters/rules. And is replicated in multiple places. I don’t think that is very organized. The second...

> How much quicker is using SQLite to extract JSON fields over returning the entire JSON for an object and using python to extract the fields? I think as long...

But letting the SQL layer pull out the parts will be as fast as possible I think.

Here is an interesting version: 1. The `BusinessLogic` is stand-alone and contains all of the functions necessary to support the methods defined here. It uses jsonpath_ng to parse the json_path's...

Oh, this works really well. I'm going to test what the speed difference is between the generic + jsonpath_ng vs dbapi + JSON_EXTRACT. Place your bets.

I can probably refactor the generic `jsonpath_ng` version so that parser doesn't have to parse the same query over and over. But the time differences are pretty stunningly different so...

Caching the parser for each json_path in the generic method gives: * Generic: 0.14 seconds * SQL: 0.05 seconds