LocalLimit

As the oscillator is made local, every TLO quantity reduces to the LLO path

9/9 passed in 1865ms

Test	What it does	Result
test_limit_calib	tlo_calib read at the power actually sampled returns ratio exactly 1 and a monitor reading of exactly $1+v_{el}$.	✅ pass
test_limit_continuous	The key-rate deficit falls linearly with the shot-noise ratio's departure from 1, to under 1e-11 bit/symbol at a part in ${10}^{12}$.	✅ pass
test_limit_epr	tlo_cov at T = 1, xi = 0, ratio 1 equals GaussianState.epr's covariance to 1e-14.	✅ pass
test_limit_estimate	At ratio 1 tlo_estimate returns T and xi unchanged, both conventions coincident and both signatures exactly zero, over nine channels.	✅ pass
test_limit_physical	tlo_cov satisfies V + i*Omega/2 >= 0 over nine channels and ratios inside the estimator's physical range.	✅ pass
test_limit_pipeline	tlo_estimate at ratio 1 agrees with run_symbols' oracle xi, four-seed mean, within ARB_TOL.	✅ pass
test_limit_range	tlo_range at ratio 1 returns the quoted half-range unchanged.	✅ pass
test_limit_rate	cv_rate on the ratio-1 estimate equals cv_rate on the true channel exactly, in both detectors and both trust models.	✅ pass
test_limit_shot	tlo_shot at a transparent LO arm returns the characterisation unit and $v_{el}$ to the last bit.	✅ pass