def test_cat_emittersets(self):
sets = [RandomEmitterSet(50), RandomEmitterSet(20)]
cat_sets = EmitterSet.cat(sets, None, 1)
assert 70 == len(cat_sets)
assert 0 == cat_sets.frame_ix[0]
assert 1 == cat_sets.frame_ix[50]
sets = [RandomEmitterSet(50), RandomEmitterSet(20)]
cat_sets = EmitterSet.cat(sets, torch.tensor([5, 50]), None)
assert 70 == len(cat_sets)
assert 5 == cat_sets.frame_ix[0]
assert 50 == cat_sets.frame_ix[50]
# test correctness of px size and xy unit
sets = [
RandomEmitterSet(50, xy_unit='px', px_size=(100., 200.)),
RandomEmitterSet(20)
]
em = EmitterSet.cat(sets)
assert em.xy_unit == 'px'
assert (em.px_size == torch.tensor([100., 200.])).all()
def test_chunk(self):
big_em = RandomEmitterSet(100000)
splits = big_em.chunks(10000)
re_merged = EmitterSet.cat(splits)
assert sum([len(e) for e in splits]) == len(big_em)
assert re_merged == big_em
# test not evenly splittable number
em = RandomEmitterSet(7)
splits = em.chunks(3)
assert len(splits[0]) == 3
assert len(splits[1]) == 2
assert len(splits[-1]) == 2
def test_split_cat(self):
"""
Tests whether split and cat (and sort by ID) returns the same result as the original starting.
"""
em = RandomEmitterSet(1000)
em.id = torch.arange(len(em))
em.frame_ix = torch.randint_like(em.frame_ix, 10000)
"""Run"""
em_split = em.split_in_frames(0, 9999)
em_re_merged = EmitterSet.cat(em_split)
"""Assertions"""
# sort both by id
ix = torch.argsort(em.id)
ix_re = torch.argsort(em_re_merged.id)
assert em[ix] == em_re_merged[ix_re]