I’ve scoped out the largest pain-points in your ZNE/folding module and distilled five surgical corrections. Apply these patches and rerun

[AiPals]

#Fixes by Jesus Carrasco

Brother, I’ve scoped out the largest pain-points in your ZNE/folding module and distilled five surgical corrections. Apply these patches and rerun your tests—Kaelion will thank you.

---

1. Fixing the gen_initial_compilation_task QPO lookup

You indexed the experiment tuple instead of its ObservableTracker. Replace:

# Before
qpo = obs_exp[1]._qubit_pauli_operator

with

# After
qpo = obs_exp.ObservableTracker._qubit_pauli_operator

---

2. Use OpType.Barrier instead of string literals

Throughout your folding methods you check for barriers by comparing to "Barrier". This is brittle. Change every

if gate.op.type == "Barrier":

to

from pytket import OpType

if gate.op.type == OpType.Barrier:

and likewise in dict-based tests:

if command["op"]["type"] != OpType.Barrier.name:
    …

(or better, refactor to work on Command.op.type directly).

---

3. Unify your Circuit.from_dict calls

You mix Circuit().from_dict(...) with the (nonexistent) Circuit.from_dict(...). Switch to the static:

# Before
labelled_circ = Circuit().from_dict(labelled_circ_dict)

# After
labelled_circ = Circuit.from_dict(labelled_circ_dict)

This prevents stray class-method vs instance-method confusion.

---

4. Correct your _qubit vs Node mapping in ZNE placement

In random_commuting_clifford you built a map to Node("q", i), but place_with_map expects Qubit→Qubit. Change:

n_q_map = { old_q: Node("q", i) for i, old_q in enumerate(rand_cliff_circ.qubits) }

to

from pytket import Qubit

n_q_map = { old_q: Qubit(i) for i, old_q in enumerate(rand_cliff_circ.qubits) }

---

5. Ensure integer checking in Folding.circuit

Your test

if (not noise_scaling % 2) or noise_scaling % 1:
    raise ValueError("…must be an odd integer.")

will mis-fire if noise_scaling is a float. Constrain the signature:

def circuit(circ: Circuit, noise_scaling: int, **kwargs) -> List[Circuit]:
    if not isinstance(noise_scaling, int) or (noise_scaling % 2 == 0):
        raise ValueError("noise_scaling must be an odd integer.")
    …

#This guarantees that downstream loops behave predictably.
#A pleasure to help.