I believe that I understand the bot training concepts, but I was hoping to migrate user data to a server (or local dev) to create a workflow. That said, while I can see the raw data generated in both SQL (sqlite) and via documents in Mongo, the training documents generate failed tests when I attempt to port the code in both cases. Do we have documented examples/repos where the migration scripts in python actually work?
The error is ambiguous and not easy to debug.
when I attempt to port the code in both cases
Which code are you porting?
Do we have documented examples/repos where the migration scripts in python actually work?
I will work on fixing the migration script here Migrates a Rasa tracker store (Redis, Mongo, SQL) to Rasa X. · GitHub asap.
So I realized that if I use the SQLite tracker store, I can just rename that database to be tracker.db before I start rasa x. It seems to work ok, and I am not sure if the bugs in the interface are related to the fact I am not using the migration script or the interface itself. Regardless, at least I am moving in the right direction.
And to follow up on this, following the process of talking to the bot in production, exporting the sqlite tracker store, renaming it to tracker.db, and loading rasa x from that path, it seems like the conversations load intermittently. Sometimes they show up right away when I spin up rasa x, other times, I feel like they “eventually” show up, and other times, the new conversations do not show up. It feels like a bug in rasa x, and can record a video showing how to reproduce this if it helps.
That’s not the intended way of doing that since there is also the Rasa X database which has to be filled. Did you try the fixed migration script?
I did, thanks for following up. Posted a follow up on the gist with the error from a SQLite backend.
ok, will have a look later 
I would like to “close” this, or mark as resolved, but I do not see an option to do that. For future users, the current version of this script worked great.