Adjust tolerance in compute_lambda_thc_test.py test (#1208)

After updating dependency versions, one of the tests in
`compute_lambda_thc_test.py` fails. This seems to be due to some
numerical drift introduced by these version updates:

   * `pyscf` updated from 2.9.0 to 2.12.1
   * `numpy` updated from 1.26.4 to 2.2.6

The test involves a small electronic structure calculation followed by a
BFGS optimization (perform_bfgs_opt=True with bfgs_maxiter=10 in
kpoint_thc_via_isdf). Even small changes in the floating-point
operations from newer numpy/scipy versions or updates to pyscf's
integral evaluations can cause these sorts of minor deviations in the
final output after optimization iterations.

Author: mhucka

src/openfermion/resource_estimates/pbc/thc/compute_lambda_thc_test.py

@@ -78,4 +78,4 @@ def test_kpoint_thc_lambda():
     )
     helper = KPTHCDoubleTranslation(kpt_thc.chi, kpt_thc.zeta, mf)
     lambda_data = compute_lambda(hcore_mo, helper)
-    assert np.isclose(lambda_data.lambda_total, 93.84613761765415)
+    assert np.isclose(lambda_data.lambda_total, 93.8478, atol=2e-3)