[BUG] qasm3_to_qir: "Variable c3 is not an array", despite c3 being an array.

[Sola85]

Environment

• qBraid-QIR version: 0.2.3
• Python version: 3.11.5
• Operating system: Linux c5ce507e9ef3 5.15.153.1-microsoft-standard-WSL2

What happened?

The following Qiskit circuit is converted to qasm3 using qiskit.qasm3.

qubits = QuantumRegister(1)
clbits = ClassicalRegister(1)
circuit = QuantumCircuit(qubits, clbits)
(q0,) = qubits
(c0,) = clbits
 
circuit.h(q0)
circuit.measure(q0, c0)
with circuit.switch(c0) as case:
    with case(0):
        circuit.x(q0)
    with case(1):
        circuit.z(q0)
circuit.measure(q0, c0)

This yields the following QASM3:

OPENQASM 3.0;
include "stdgates.inc";
bit[1] c3;
int switch_dummy;
qubit[1] q3;
h q3[0];
c3[0] = measure q3[0];
switch_dummy = c3[0];
switch (switch_dummy) {
  case 0 {
    x q3[0];
  }
  case 1 {
    z q3[0];
  }
}
c3[0] = measure q3[0];

I get the following error message: qbraid_qir.qasm3.exceptions.Qasm3ConversionError: Indexing error. Variable c3 is not an array and Error at line 8, column 18 in QASM file.

I.e. the indexing c3[0] seems to fail.

Suggestions (Optional)

No response

[TheGupta2012]

Hi @Sola85 , thanks for raising an issue!

Since, internally, we are handling c3[0] = measure q3[0]; as classical assignment statements we see this kind of error message.c3 is technically not an array (because it is a bit type scalar variable with 1 bit length), the classical assignment handler raises an error. Whenever we have indexing on the lhs, our semantic analyser expects that the variable is of type array.

Also, we are not correctly handling the syntax c3[0] = measure q3[0]; in our pyqasm.unroll function. It makes sense to convert this statement to a QuantumMeasurementStatement instead of a ClassicalAssignment before appending to the unrolled AST.

[ryanhill1]

@TheGupta2012 Based on your evaluation, would it make sense to transfer this issue to the pyqasm repo?

[TheGupta2012]

Yea that makes more sense @ryanhill1.
@Sola85 closing this as I've opened the issue in pyqasm

[ryanhill1]

@TheGupta2012 Transferred and re-opened original issue, closed your duplicate

[Sola85]

Awesome, thanks! Will try it out as soon as I find the time!