Deprecate Two/ThreeQubitGate
#4207
Conversation
daxfohl
requested review from
a team,
cduck,
vtomole and
wcourtney
as code owners
daxfohl
changed the title
ThreeQubitGateTwo/ThreeQubitGate
daxfohl
changed the title
Two/ThreeQubitGateTwo/ThreeQubitGate
|
Wondering now if we should make Gate generic based on shape of supported qubits. So Gate[Tuple[Qubit, Qubit]] would be a two-qubit gate. That lets us be more explicitly typesafe without requiring multiple inheritance. There are some subtleties like ControlledGates, but maybe we could work it out. |
|
I personally prefer the current structure of checking gate attributes (eg: via Let's merge this PR and discuss the proposal on a separate issue. |
tanujkhattar
added
the
automerge
CirqBot
added
the
front_of_queue_automerge
CirqBot
removed
automerge
|
Hey @daxfohl , does |
|
It shouldn't. Where are you seeing it? |
|
Oh that's right. Sorry about that. Was confused about something. |
|
It happens. I just spent half a day trying to figure out why a test framework was rejecting my tests and finally had someone point out that my code wasn't building. :) |
Closes quantumlib#2164 Copies `Two/ThreeQubitGate` to cirq.testing due to its convenience there, and deprecates the originals. BREAKING CHANGE: @balopat @95-martin-orion @tanujkhattar During the deprecation period, `isinstance(TwoQubitGate)` has been shimmed such that it will now return True iff _num_qubits_() returns 2, regardless of whether it's actually part of the class hierarchy (and same for ThreeQubitGate). It's not expected that this will cause real-world breakages, but possible. This can go before or after quantumlib#4236, as they are disjoint problems.
Closes quantumlib#2164 Copies `Two/ThreeQubitGate` to cirq.testing due to its convenience there, and deprecates the originals. BREAKING CHANGE: @balopat @95-martin-orion @tanujkhattar During the deprecation period, `isinstance(TwoQubitGate)` has been shimmed such that it will now return True iff _num_qubits_() returns 2, regardless of whether it's actually part of the class hierarchy (and same for ThreeQubitGate). It's not expected that this will cause real-world breakages, but possible. This can go before or after quantumlib#4236, as they are disjoint problems.
8 participants
Closes #2164
Copies
Two/ThreeQubitGateto cirq.testing due to its convenience there, and deprecates the originals.BREAKING CHANGE: @balopat @95-martin-orion @tanujkhattar During the deprecation period,
isinstance(TwoQubitGate)has been shimmed such that it will now return True iff num_qubits() returns 2, regardless of whether it's actually part of the class hierarchy (and same for ThreeQubitGate). It's not expected that this will cause real-world breakages, but possible.This can go before or after #4236, as they are disjoint problems.