Thanks for the context, that helps! I got tripped up slightly by the way the documentation is arranged which
suggested it ought to be possible to run, for example, mix
ecto.gen.migration xyz, but since those tasks are now in the ecto_sql
repo they were no longer present.
The situation is roughly the same as far as I know. It seems like being able to arrange migration scripts (even if limited to index management) in a similar way to that of ecto_sql could still be useful, and might make a good vehicle for other Mongo features like schema. The mix tasks for dumping / loading are similarly useful too (perhaps using mongodump / mongorestore).
This wouldn't preclude running a Task on startup if needed. I have also wondered about supporting some sort of "declare indexes" functionality that would allow a user to define the indexes they want and have the adapter take care of managing the creation / removal of them under the hood to match.
At the risk of creating huge amounts of work for myself... has an `ecto_nosql` project been considered?
Any other guidance / tips / recommendations on the best way of proceeding with this and in general with helping to get ecto_mongo back on track are much appreciated.