Harjot Singh

Notes so far - Search indexes use a LOT of storage, if you pregenerate the possible results for every query (about 86MB just for SGGS, with only line IDs attached)....

Perhaps we only need to solve one issue, aside from dropping the ORM. The smart sorting: ``` select * from `line_content` inner join `lines` on `lines`.`id` = `line_content`.`line_id` inner join...

with generated serach index: ``` select * from `search_only` inner join `line_content` on `line_content`.`line_id` = `search_only`.`line_id` inner join `lines` on `lines`.`id` = `line_content`.`line_id` inner join `shabads` on `shabads`.`id` = `lines`.`shabad_id`...

Indeed! Which is why Partial Sync support is planned for the next release!

No, because we'd still have to rename the table

Need to modify the `Translations` model to parse that field as JSON before it's returned. See objection docs.

When building the source JSON files, yes. The issue is when quoting the DB, which does not call JSON.parse on output

Great call! This is exactly what I've been working on 😄

An option to overcome this is to chunk downloads/sync at the app level, paired with the DB API.

Linked to #1881