pandas helper: where join

[kalenkovich]

For things like being close enough time-wise, overlapping intervals, etc. But also just for whenever you want to write an SQL join with the "where" clause.

• cross-join or inner-join at the group level,
• apply the where clause (probably .loc[where_fun] where where_fun is an argument),
• right/left-join if the overall join is a left join