Add Trotter state-error baseline

[saktar-lanl]

Summary

Adds a baseline/ directory holding reference Trotterization-error measurements for the QHAT molecular Hamiltonian library, so a new Trotter or time-evolution technique can be measured against a fixed, documented set of conditions.

For each Hamiltonian the metric is the state error

err = || psi_trotter(T) - psi_exact(T) ||_2

starting from the Hartree-Fock determinant, with both states evolved under the same generator so no global phase enters.

What's included

• compute_trotter_state_error.py — computes the baseline; resumable, one CSV per (basis, mapping, species)
• three plot scripts covering active space, molecule and evolution-time views
• state_errors/ — 32 CSVs: 9 diatomics, hgbs-5 and sto-6g, JW and BK
• plots/ — sample figures
• README.md — conventions, how to run, and how to evaluate a new technique against it

Grid: T = 2^6, 2^8, 2^9 by delta_t = 1, 1/2, 1/4 by Trotter order 1, 2, 4, giving 27 rows per (Hamiltonian, mapping). Each CSV carries a provenance header recording the conditions that produced its rows.

Change to existing code

hamiltonian_generator/build_config_L_sweep.py gains monatomic systems H through Ne, generated at the origin with no bond-length sweep and written to library/atoms/<atom>/<basis>/.

[saktar-lanl]

Updated with the single-atom results, so the baseline now covers all 19 systems: 9 molecules and 10 atoms, in both basis sets and both fermion-to-qubit mappings.

The README has been updated to match.

[BrendanKKrueger]

@saktar-lanl: Did you actually want to add all those data files and plots to the repository? It's unclear to me if details of a specific study belong in the repository versus in the shared drive, so I wanted to ask your thoughts on the matter.