little help with timeseries and spans

I need help finding the right path to accomplish some custom "visual masking"
and corresponding array-mask construction:

For much of what I need, scikits.timeseries initially sounded useful but
either I misunderstand how to use it, or it just can't do most of what I want,
which is:

1: create/apply a mask that invalidates/masks the data that is outside
normal business hours; i.e. mask out the weekends and anything between
18:00 and 06:00 the next day.

2: The second task is to create vertical spans in the plot to show the mask

So, for #2 what I need is essentially two collections of axvspan patches:

Axvspan collection 'a' begins at 17h00 each Friday, and ends at 09h00 each Monday.
Collection 'b' begins at 17h00 each weekday, and ends at 09h00 the following morning, but
is masked out by collection 'a'.

Axes.fill_betweenx() looks like it's *not* what I need....

Most of this app operates at finer granularity than what pyplot provides, so I'm dealing directly
with the individual axes and artists; the solution needs to work in that domain.

So I'd be grateful for ideas....