fix: handle 1D interpolation result in resample for single-state ODEs - #1

tests/test_trajectory_analysis.py (2)

388-463: Consider DRYing repeated single-state ODE definitions.

The same ODE/Trajectory scaffolding is repeated across several tests (Line 390–Line 413, Line 420–Line 428, Line 452–Line 460). A small helper would reduce noise and make intent clearer.

Example refactor direction

 class TestResampleSingleState:
+    `@staticmethod`
+    def _make_single_state_traj(with_output=False):
+        class ODE(co.ODESystem):
+            a = parameter()
+            x = state()
+            if with_output:
+                dynamic_output.velocity = -a * x
+            dot[x] = -a * x
+
+        class Traj(ODE.TrajectoryAnalysis):
+            tf = 10
+            initial[x] = 1
+            if with_output:
+                vel = dynamic_output.velocity
+
+        return Traj

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/test_trajectory_analysis.py` around lines 388 - 463, Multiple tests
duplicate the same single-state ODE and TrajectoryAnalysis scaffolding (classes
named ODE and Traj) — create a small helper to return a ready-to-use trajectory
instance and use it in each test instead of repeating the class definitions.
Implement a factory function (e.g., make_single_state_traj(a=0.5, tf=10,
initial=1, dynamic_output=False)) that builds the ODE/Traj classes (including
optional dynamic_output and exposing vel when needed) and returns an
instantiated sim; then update test_resample_single_state*,
test_resample_single_state_values, test_resample_single_state_no_output,
test_resample_single_state_with_dynamic_output, and
test_resample_single_state_small_dt to call make_single_state_traj(...) and call
sim.resample(...) as before. Ensure the helper preserves parameter names and
initial state behavior so existing assertions against resampled.t and
resampled.x continue to work.

---

433-449: Assert dynamic output values, not only successful execution.

At Line 447–Line 448, the test proves the call doesn’t crash, but it won’t catch incorrect dynamic_output values/shaping. Add a value assertion on resampled.vel so this path is actually validated.

Proposed test strengthening

     sim = Traj(a=0.5)
     resampled = sim.resample(1.0, include_events=False, include_output=True)
     assert resampled.t.size > 0
+    expected_vel = -0.5 * np.exp(-0.5 * resampled.t)
+    np.testing.assert_allclose(resampled.vel, expected_vel, rtol=1e-4)

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/test_trajectory_analysis.py` around lines 433 - 449, The test currently
only ensures resample runs; add an assertion that dynamic_output values are
correct by comparing resampled.vel against the analytical velocity -a * exp(-a *
t) (or elementwise -a * x where x = exp(-a * t)) from the Traj/Timestep results;
use the same a passed to Traj(a=0.5) and compare elementwise across resampled.t
(or at the t=1.0 sample) with an approximate-equality check (e.g.,
allclose/approx) to validate both value and shaping for dynamic_output.velocity
in test_resample_single_state_with_dynamic_output.

tests/test_trajectory_analysis.py (1)

388-483: Consider extracting a shared helper/fixture for ODE/Traj setup.

The repeated class definitions are very similar across tests; a small fixture would reduce duplication and future maintenance overhead.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/test_trajectory_analysis.py` around lines 388 - 483, Several tests
repeat near-identical ODE and Traj classes (e.g., ODE and Traj in
test_resample_single_state, test_resample_single_state_values,
test_resample_single_state_with_dynamic_output, etc.); extract a small helper or
pytest fixture like make_sim(...) that builds and returns the Traj simulator
instance (or the Traj class) configured by flags: states (['x'] or ['x','v']),
parameter a, tf, initial values, and an optional dynamic_output flag to attach
dynamic_output.velocity; replace direct class definitions in each test with
calls to make_sim(a=0.5, tf=10, initial={'x':1}, dynamic_output=True/False) so
tests reuse the factory and reduce duplication while keeping names ODE and Traj
references inside the helper for clarity.