👋 Having some issue with AutoETS in spark as part of the default MSTL's trend_forecaster. I think it's numba related. (error). 1. Can I use another forecaster? I tried AutoARIMA, but it complains autoARIMA doesn't have predict_interval method implemented

'AutoARIMA' object has no attribute 'prediction_intervals'

. I actually don't mind which trend forecaster is used at this stage, i just want to get this running first. The documentation about which forecaster can be used is lackingd, that has

prediction_intervals

implemented? 2. I'm currently on

numba 0.51.2

, is this a problem? I see that on git repo, we need at least

numba=0.55

, is this true? I can try to get myself upgraded.