notebooks for sherpa/sherpa#1246

bg.reg

annulus(3147.6844,4519.83,73.123398,146.2468)
-circle(3036.4982,4611.592,18.016347)
-circle(3199.2587,4600.8753,11.875275)
-box(3148.8565,4415.5092,228.06552,49.230003,0)

gistfile1.txt

The same notebook run with the main branch (`old`) and #1246 (`new`)
showing the issue that is fixed.

Note that re-creating the original filter from

    notice2d('src.reg')
    notice2d('bg.reg')
    orig = get_filter()

with

    notice2d()
    notice2d(orig)

fails with main (it filters out everything) but works with #1246

Also of note is that it appears that the same data has been selected
in both cases (as the `calc_stat_info` returns the same number of
points and the same statistic value), it's just that the filter
specification is wrong.

src.reg

circle(3145.9471,4520.7031,15.800277)

test-new.ipynb

test-old.ipynb

Unfortunately the get_data() calls, which generate matplotlib image plots, are not shown here, but they show the expected behavior of the filters.