Skip to content

Modify SquaredMahalanobis __call__ to accept a 2D state_vector in state2 #1272

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

Open
ajland wants to merge 6 commits into
base: main
Choose a base branch
from
Open

Modify SquaredMahalanobis __call__ to accept a 2D state_vector in state2 #1272

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
de7a0ea
Modify SquaredMahalanobis __call__ to accept a 2D state_vector in state2
ajland Mar 30, 2026
a546e81
Merge branch 'dstl:main' into vectorize_mahalanobis
ajland Mar 31, 2026
b32b6f2
Alter einsum to compute result more directly. Clarify docstring.
Mar 31, 2026
4fb84b4
Fix mahalanobis test to be less strict
Mar 31, 2026
5a60f3f
Add to tests to check the new one-to-many mahalanobis distance functi…
Apr 1, 2026
066dcd7
Fix flake8 errors
Apr 1, 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
23 changes: 15 additions & 8 deletions stonesoup/measures/state.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -188,32 +188,39 @@ def __call__(self, state1, state2):

Parameters
----------
state1 : :class:`~.State`
state2 : :class:`~.State`
state1 : :class:`~.State` whose `state_vector` has shape (n, 1)
state2 : :class:`~.State` whose `state_vector` has shape (n, m)

Returns
-------
float
Union(float, np.ndarray)
Squared Mahalanobis distance between a pair of input :class:`~.State`
objects
objects. Returns a 1D array if state2 has >1 length along axis 1.

"""
state_vector1 = getattr(state1, 'mean', state1.state_vector)
state_vector2 = getattr(state2, 'mean', state2.state_vector)

if len(state_vector1) != len(state_vector2):
raise ValueError(
f"Shape mismatch between state1 and state2 along axis 0, \
{len(state_vector1)} != {len(state_vector2)}."
)

if self.mapping is not None:
u = state_vector1[self.mapping, 0]
v = state_vector2[self.mapping2, 0]
v = state_vector2[self.mapping2, :]
# extract the mapped covariance data
vi = self._inv_cov(state1, tuple(self.mapping))
else:
u = state_vector1[:, 0]
v = state_vector2[:, 0]
v = state_vector2[:, :]
vi = self._inv_cov(state1)

delta = u - v
delta = -(v.T-u)

return np.dot(np.dot(delta, vi), delta)
# Return the diagonal elements of A@B@A.T
return np.einsum('ij,jk,ik->i', delta, vi, delta).squeeze()[()]

@staticmethod
def _inv_cov(state, mapping=None):
Expand Down
55 changes: 46 additions & 9 deletions stonesoup/measures/tests/test_state.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,9 +21,11 @@
stateB_u = State(u, timestamp=t)

v = StateVector([[11.], [10.], [100.], [2.]])
vm = StateVectors(np.random.random((4, 10)))
vi = CovarianceMatrix(np.diag([20., 3., 7., 10.]))

state_v = GaussianState(v, vi, timestamp=t)
state_vm = State(vm)
stateB_v = State(v, timestamp=t)


Expand All @@ -48,9 +50,14 @@ def test_euclideanweighted():

def test_mahalanobis():
measure = measures.Mahalanobis()
assert measure(state_u, state_v) == distance.mahalanobis(u[:, 0],
v[:, 0],
np.linalg.inv(ui))
assert measure(state_u, state_v) == pytest.approx(distance.mahalanobis(u[:, 0],
v[:, 0],
np.linalg.inv(ui)))
result_nm = measure(state_u, state_vm)
for i, vec in enumerate(vm):
assert result_nm[i] == pytest.approx(distance.mahalanobis(u[:, 0],
vec[:, 0],
np.linalg.inv(ui)))


def test_hellinger():
Expand Down Expand Up @@ -141,13 +148,23 @@ def test_hellinger_partial_mapping(mapping_type):
def test_mahalanobis_full_mapping(mapping_type):
mapping = mapping_type(np.arange(len(u)))
measure = measures.Mahalanobis(mapping=mapping)
assert measure(state_u, state_v) == distance.mahalanobis(u[:, 0],
v[:, 0],
np.linalg.inv(ui))
assert measure(state_u, state_v) == pytest.approx(distance.mahalanobis(u[:, 0],
v[:, 0],
np.linalg.inv(ui)))
result_nm = measure(state_u, state_vm)
for i, vec in enumerate(vm):
assert result_nm[i] == pytest.approx(distance.mahalanobis(u[:, 0],
vec[:, 0],
np.linalg.inv(ui)))
measure = measures.Mahalanobis(mapping=mapping, mapping2=mapping)
assert measure(state_u, state_v) == distance.mahalanobis(u[:, 0],
v[:, 0],
np.linalg.inv(ui))
assert measure(state_u, state_v) == pytest.approx(distance.mahalanobis(u[:, 0],
v[:, 0],
np.linalg.inv(ui)))
result_nm = measure(state_u, state_vm)
for i, vec in enumerate(vm):
assert result_nm[i] == pytest.approx(distance.mahalanobis(u[:, 0],
vec[:, 0],
np.linalg.inv(ui)))


def test_mahalanobis_partial_mapping(mapping_type):
Expand All @@ -157,23 +174,43 @@ def test_mahalanobis_partial_mapping(mapping_type):
assert measure(state_u, state_v) == \
distance.mahalanobis([10, 1],
[11, 10], np.linalg.inv(reduced_ui))
result_nm = measure(state_u, state_vm)
for i, vec in enumerate(vm):
assert result_nm[i] == pytest.approx(distance.mahalanobis(u[mapping, 0],
vec[mapping, 0],
np.linalg.inv(reduced_ui)))
mapping = np.array([0, 3])
reduced_ui = CovarianceMatrix(np.diag([100, 10]))
measure = measures.Mahalanobis(mapping=mapping)
assert measure(state_u, state_v) == \
distance.mahalanobis([10, 1],
[11, 2], np.linalg.inv(reduced_ui))
result_nm = measure(state_u, state_vm)
for i, vec in enumerate(vm):
assert result_nm[i] == pytest.approx(distance.mahalanobis(u[mapping, 0],
vec[mapping, 0],
np.linalg.inv(reduced_ui)))

mapping = mapping_type([0, 1])
measure = measures.Mahalanobis(mapping=mapping, mapping2=mapping)
assert measure(state_u, state_v) == \
distance.mahalanobis([10, 1],
[11, 10], np.linalg.inv(reduced_ui))
result_nm = measure(state_u, state_vm)
for i, vec in enumerate(vm):
assert result_nm[i] == pytest.approx(distance.mahalanobis(u[mapping, 0],
vec[mapping, 0],
np.linalg.inv(reduced_ui)))
mapping = np.array([0, 3])
measure = measures.Mahalanobis(mapping=mapping, mapping2=mapping)
assert measure(state_u, state_v) == \
distance.mahalanobis([10, 1],
[11, 2], np.linalg.inv(reduced_ui))
result_nm = measure(state_u, state_vm)
for i, vec in enumerate(vm):
assert result_nm[i] == pytest.approx(distance.mahalanobis(u[mapping, 0],
vec[mapping, 0],
np.linalg.inv(reduced_ui)))

mapping = mapping_type([0, 1])
mapping2 = np.array([0, 3])
Expand Down