check how pfs_visit_id are inserted in visit_set for AG

yuki.moritani reports :

We have a question how iic determines "sequence_type". While we were taking 4 exposures, we had to stop and start auto guide in the middle of 3rd exposure because heavy cloud lost the guide stars. As a result, the third visit got autoguideStart instead of scienceObject and different sequence_id. Note that fourth visit keep the same sequence_id as 1st and 2nd exposures. We wonder if iic could keep the sequence_id and sequence_type as exposure already started..

I don't think it's a sequence_type issue but rather a visit_set insertion issue. please check.