if you want to use csdid , you probably can try csdid2. Its much faster, because it avoid the bottle neck of using Stata on the surface, doing everything in Mata.
it will not help you with continuous treatment tho.
I m working on that for jwdid tho

Thank you!.
And perhaps the example of panel vs Repeated crossection may be of use:
friosavila.github.io/app_metrics/...
This has the code attached.
Hope its helpful