Skip to content

Stabilize Euclidean MTT distances at extreme scales #4968

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
FlorianPfaff wants to merge 2 commits into from
+16 −1
Closed

Stabilize Euclidean MTT distances at extreme scales #4968

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e24c1ab
Stabilize Euclidean MTT pairwise norms
FlorianPfaff Jul 31, 2026
d3955f3
Test extreme finite Euclidean MTT distances
FlorianPfaff Jul 31, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion src/pyrecest/evaluation/get_distance_function.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -255,7 +255,7 @@ def _euclidean_mtt_distance(x1, x2, *, cutoff_distance: float) -> float:
return float(cutoff_distance * abs(first.shape[0] - second.shape[0]))

deltas = first[:, None, :] - second[None, :, :]
costs = numpy.linalg.norm(deltas, axis=2)
costs = numpy.hypot.reduce(deltas, axis=2)
costs = numpy.minimum(costs, float(cutoff_distance))
row_indices, column_indices = linear_sum_assignment(costs)
matched_cost = float(costs[row_indices, column_indices].sum())
Expand Down
15 changes: 15 additions & 0 deletions tests/evaluation/test_euclidean_mtt_distance_orientation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,3 +41,18 @@ def test_empty_target_dimension_disambiguates_small_dim_first_target_set():

np.testing.assert_allclose(distance(no_targets, dim_first_targets), 21.0)
np.testing.assert_allclose(distance(dim_first_targets, no_targets), 21.0)


def test_euclidean_mtt_distance_preserves_extreme_finite_norm_below_cutoff():
distance = get_distance_function(
"euclidean_mtt",
{"cutoff_distance": 1.5e308},
)
first = np.array([[1.0e308, 1.0e308]])
second = np.array([[0.0, 0.0]])
expected = np.hypot(1.0e308, 1.0e308)

with np.errstate(over="raise", invalid="raise"):
result = distance(first, second)

np.testing.assert_allclose(result, expected, rtol=1.0e-15)
Loading