Update TketBackend to derive from BackendV2
#389
Conversation
cqc-alec
force-pushed
the
tketbackend
branch
from
dfcbcb8 to
1dff824
Compare
| gate_name = _gate_str_2_optype_rev[optype] | ||
| if gate_name in standard_gate_mapping: | ||
| gate_obj = standard_gate_mapping[gate_name] | ||
| if gate_obj.num_qubits in [1, 2] or inspect.isclass(gate_obj): |
There was a problem hiding this comment.
See https://github.com/Qiskit/qiskit/blob/2ef371ae0d159a6dfd643805f3e5e5fdec37ab88/qiskit/transpiler/target.py#L1066 -- have to restrict the gateset to match what is accepted by the Target method.
There was a problem hiding this comment.
Do I understand correctly that the gate_obj variable is not always a class? Presumably thats why the inspect check is needed.
There was a problem hiding this comment.
Yes, some gates such as CCX were rejected when I omitted this check. I am not sure when gate_obj is a class, but simply copied that from the code linked above.
| ] | ||
| basis_gates = [] | ||
| for optype in pred.gate_set: | ||
| if optype in _gate_str_2_optype_rev.keys(): |
There was a problem hiding this comment.
Perhaps worth adding a comment explaining what these "if" statements are for?
| from pytket.architecture import FullyConnected | ||
|
|
||
|
|
||
| def _extract_basis_gates(backend: Backend) -> list[str]: | ||
| standard_gate_mapping = get_standard_gate_name_mapping() | ||
| for pred in backend.required_predicates: | ||
| if type(pred) == GateSetPredicate: |
There was a problem hiding this comment.
Rather pedantic but is there a reason not to use "is" or "isinstance" here vs "=="?
Closes #350 .
This change would break one of the examples in the manual, and perhaps also one or two places in the example notebooks (I haven't checked). This shouldn't affect the docs build since it's pinned to a specific version of pytket-qiskit, but we should probably revise those sections. It seems that the relevant parts of
qiskit_algorithms(a community-maintained project) haven't yet been updated to use the new backend class.This PR removes the last of the deprecation warnings from the qiskit tests that were coming from our code.