ENH - Propagate bad-angle flags from L1B Histogram to L2 in GLOWS

[maxinelasp]

Parent issue: #1276

Per Section 12.3.4 of the algorithm document, bad-angle flags from L1B must propagate to L2 using OR logic: if a given flag is True in any L1B histogram for the observation day, it is True in L2.

Currently DailyLightcurve sets histogram_flag_array = np.zeros(...) (placeholder) and there is a # todo: bad angle filter comment in glows_l2.py:82.

Implementation:

• Collect the histogram_flag_array from all good-time L1B blocks
• OR them together per bin to produce the L2 histogram_flag_array

Depends on #2721 (bad-angle flags must be non-zero in L1B first).