Shared Tavis–Cummings / cavity–emitter model factory

[simsaidan]

Summary

Add a reusable factory that builds the open cavity–emitter (Tavis–Cummings–like) problem used in README Example 2 / the Sims et al. workflow, so users can swap parameters without copying operator soup.

Motivation

The README and examples/simple_algo_benchmark.py hard-code ω_c, ω_e, κ, γ, g, Fock truncation, Lindblad ops, local vs full embeddings, and measurement helpers. That duplication makes it hard to:

• change coupling / decay / truncation for a paper figure
• keep Example 2 and docs in sync
• write short benchmarks or tests against a known physical model

Proposed API

from shadowsim.models import tavis_cummings

model = tavis_cummings(
    omega_c=245000,
    omega_e=245000,
    kappa=...,          # or decay rates as used today
    gamma=...,
    g=100,
    cavity_levels=4,    # truncation → two-qubit Fock encoding or explicit dim
    # optional: initial_fock=2, etc.
)

# Suggested fields (names flexible):
model.full_hamiltonians      # list[Hamiltonian] for QuTiP path
model.local_hamiltonians     # list[LocalHamiltonian] for Split-J path
model.c_ops_full             # list[Operator]
model.c_ops_local            # list[LocalOperator]
model.psi0                   # State
model.e_ops                  # observables for QuTiP
model.measurement_groups     # for Split-J, if applicable
model.reducers               # cavity_population, population_one, etc.