alternative_splicing_events/altsplice_test crash with TypeError on real data (numpy.bool_ index)

[YalanBi]

Discovered while re-running the tutorial notebooks against a freshly reconstructed real ENCODE chr8 transcriptome (not a fixture artifact -- reproduces on any gene with an exon-skipping-type alternative splicing event).

Error

TypeError: tuple indices must be integers or slices, not numpy.bool

Raised from SegmentGraph.find_splice_bubbles (splice_graph.py), at:

inB_sets[i + 1][unspliced and node_id == i].add(transcript_id)

unspliced = self[i].end == node_B.start -- when exon coordinates are numpy int64 (as they are for real alignment-derived data, not just plain Python int fixtures), this comparison yields numpy.bool_ instead of Python bool. Python only accepts bool/int as a sequence index; numpy.bool_ is not a subclass of int and is rejected.

Impact

This breaks Transcriptome.alternative_splicing_events(), Transcriptome.altsplice_test() (differential splicing), and likely Transcriptome.coordination_test(), for any real transcriptome with numpy-typed exon coordinates -- i.e. essentially all real-world usage.

Fix

Coerce to Python bool at the indexing site.