Confidence bounds for regression tests

[jnsbck]

Regression is now evaluated with a single sample t-test, i.e. the baseline times = conf bounds. This is much nicer imo and giving a confidence region for tests to pass should also make it more robust to potential runtime fluctuations on the github runners.

Also made a few smaller follow up fixes for #475.

[jnsbck]

/update_regression_baselines

[github-actions]

Regression Baseline Update

✅ Process completed

test_runtime(num_cells=1, artificial=False, connect=False, connection_prob=0.0, voltage_solver=jaxley.stone)
🔴 build_time: (+245.14% vs 0.547s).
🔴 compile_time: (+7.76% vs 18.720s).
🔴 run_time: (+1.69% vs 2.817s).

test_runtime(num_cells=1, artificial=False, connect=False, connection_prob=0.0, voltage_solver=jax.sparse)
🔴 build_time: (+13.18% vs 0.317s).
🔴 compile_time: (+16.14% vs 3.107s).
🔴 run_time: (+22.99% vs 2.442s).

test_runtime(num_cells=10, artificial=False, connect=True, connection_prob=0.1, voltage_solver=jaxley.stone)
🔴 build_time: (+151.64% vs 1.628s).
🔴 compile_time: (+3.82% vs 29.477s).
🟠 run_time: (-0.61% vs 19.020s).

test_runtime(num_cells=10, artificial=False, connect=True, connection_prob=0.1, voltage_solver=jax.sparse)
🔴 build_time: (+52.30% vs 1.253s).
🔴 compile_time: (+3.75% vs 26.463s).
🔴 run_time: (+0.29% vs 25.086s).

test_runtime(num_cells=1000, artificial=True, connect=True, connection_prob=0.001, voltage_solver=jaxley.stone)
🟠 build_time: (-1.99% vs 110.895s).
🔴 compile_time: (+5.56% vs 45.626s).
🔴 run_time: (+0.74% vs 41.248s).

test_runtime(num_cells=1000, artificial=True, connect=True, connection_prob=0.001, voltage_solver=jax.sparse)
🔴 build_time: (+1.15% vs 108.675s).
🔴 compile_time: (+9.06% vs 60.998s).
🔴 run_time: (+4.69% vs 59.059s).

✨ Changes have been pushed directly to this PR