I would like to run a UCI engine with chess.engine and get both multipv results, as they become available, and the value of bestmove.

It seems that engine.play(...) has no way of returning multipvs before bestmove has been sent, while engine.analysis(...) simply throws away the value of bestmove.

Most of the times I can recover the bestmove from the first pv, but in the rare cases where the engine responds too fast to give any info lines, this doesn't work.

Am I missing something?