Python wrappers / utility functions

[bjader]

Ideally we would have our examples have minimal calls to the Julia namespace. For a Python user, each call to jl.foo() requires knowing the function name, arguments and behaviour up-front, since an IDE won't be able to link this to corresponding Julia function. Furthermore, boilerplate code such as:

qc = transpile(qc, backend=backend, basis_gates=list(SUPPORTED_GATES))

circuit, edges = circuit_description(qc)

qmap = graph_to_grid(graph_from_edges(edges))

# Set tensor network truncation parameters
chi = 4
cutoff = 1e-12

psi_bpc, errors = jl.tn_from_circuit(circuit, qmap, edges, chi, cutoff)

Could all be replaced by a Python function that executes all the above with helpful argument names and documentation:

def tn_from_circuit(circuit: QuantumCircuit, chi, cutoff))
    qc = transpile(qc, backend=backend, basis_gates=list(SUPPORTED_GATES))
    circuit, edges = circuit_description(qc)
    qmap = graph_to_grid(graph_from_edges(edges))
    return jl.tn_from_circuit(circuit, qmap, edges, chi, cutoff)

A similar change could be made for having to call to:

obs_jl = jl.translate_observable(observable_description(obs), qmap)
z_eval = np.real(jl.expect(psi_bpc, obs_jl))

for any expectation value