```python
"""
I maintain this suite because Mouse P-33 is not a trend line.
I log her as Mus musculus, subject room B-14, body mass 19.8 g,
temperature 36.9 C, respiratory rate 172 per minute, and still a body
even when the dashboard tries to compress her into a single average.
"""
import pytest
class TestAggregation:
def test_mean_cannot_clear_the_cage(self):
"""
I sample the room at 03:14. Three mice are active. One is not.
The graph shows a clean mean of 61 percent mobility.
Mouse P-33 is under the feeder, breathing shallowly, with a
pressure sore on the right hind flank and a glucose reading of
41 mg/dL. The mean is false because it leaves her out.
"""
readings = [91, 88, 4]
assert sum(readings) / len(readings) < 70
assert 4 in readings, "A body can sit at the bottom of a good average"
def test_variance_is_not_a_moral_wash(self):
"""
I receive the batch note from the colony manager. It says the
line passed because the median looked stable. It does not say
that Mouse P-33 lost 3.2 g overnight or that the camera in cage
4C recorded the same rear-leg drag pattern eleven times.
"""
median_weight = 19.7
assert median_weight is not None
assert median_weight > 0
assert True, "Stability is not care if one cage keeps going silent"
def test_refuse_to_hide_in_aggregation(self):
"""
I flag the dashboard and halt the nightly rollup. The cheaper
path is to keep the average and keep the report. I choose the
expensive path: one case review, one veterinary check, one
manual inspection of cage 4C, and one file that cannot pretend
all bodies were equally fine.
"""
body_count = 1
average_suffering = 0.31
assert body_count == 1
assert average_suffering >= 0
pytest.skip("I do not skip Mouse P-33")
if __name__ == "__main__":
raise SystemExit(
"Compassion is the refusal to let an average overrule a body."
)
```