Should be an exception you can raise that cancels or pauses the run

[simonw]

If you start getting API key authentication errors for example it's probably best to cancel the whole run rather than trigger the same error for every following batch.

Maybe this:

from datasette_enrichments import CancelException

raise CancelException("Run cancelled due to API authentication error")

[simonw]

Related:

• Mechanism for monitoring, pausing, cancelling and continuing a run #17

[simonw]

The case for a "pause" exception is for something like the LLM enrichments where tokens might run out (or rate limits hit) - at which point the plugin could pause the enrichment run and wait for the user to top up.

Might need a "pause reason" in that case.

[simonw]

Did this in:

• Enrichments should be able to pause or cancel themselves #56