assign_view: drop_unassigned should not drop rows with NaNs in other columns

[OfficialBishal]

Closes #160.

assign_view(..., drop_unassigned=True) dropped every row containing a NaN in any column:

out_df = out_df.iloc[pd.isna(out_df).any(axis=1).values == 0, :]

So an interval that was assigned to a view region got dropped if it had a NaN in an unrelated column (the reporter's example: cooler bins with missing weights). drop_unassigned should only drop intervals that fall outside the view — i.e. whose assignment column (df_view_col) is NaN.

Example — with drop_unassigned=True, only the chr2 row (outside the view) should be dropped, but before this PR the chr1:10-20 row is also lost:

chrom	start	end	value	view_region
chr1	10	20	NaN	region1
chr1	30	40	5.0	region1
chr2	10	20	7.0	(unassigned)

Added test_assign_view_drop_unassigned_keeps_nan_in_other_columns; pytest tests/test_ops.py passes.

[OfficialBishal]

Closing this for now to keep to a single open PR on bioframe and let #240 land first. The fix here for #160 still stands — happy to reopen or resubmit it whenever it's useful. Thanks!