[bwang257]

search_rocket unpacks get_inertias() in the wrong order:
get_inertias returns (longitudinal, rotational) and rocket.py:24 assigns it to inertia_z, inertia_i, so rocket.inertia comes out as (rotational, rotational, longitudinal). In OpenRocket, "longitudinal" is the pitch/yaw inertia, not the roll inertia. AbstractRKSimulationStepper.java (328-330) in RocketPy divides the "longitudinal" inertia into pitch and yaw moments.

This PR swaps the unpack order (single line change) and adds a test that asserts that roll inertia < pitch inertia, skipping examples with no inertia data. This test fails initially and passes with the fix.

Also fixed the order in the example parameters.json files and tests.

[juliomachad0]

Hi @bwang257 !, the PR looks great and the physics explanation makes total sense!
Could you please just run make format (or sort the import math above import pytest in tests/acceptance/test_ork_extractor.py) so the linter check passes in CI?
Once that's done, I'll merge it. Thanks!

[bwang257]

Hi @juliomachad0!, thanks! I ran make format and the linter check should pass in CI now. Could you approve the workflow run so lint can re-run?

Once again, thanks for taking a look at this PR!