Feature/Implement Qubit Register Consolidation

[vinayswamik]

Summary of changes

Implement device qubit(__PYQASM_QUBITS__) support in QasmVisitor and QasmModule

• Added a new parameter device_qubits to QasmVisitor and QasmModule to manage qubit consolidation.
• Introduced _get_pyqasm_device_qubit_index method for qubit index mapping and _qubit_register_consolidation to ensure compatibility with device constraints.
• Updated measurement, barrier, and gate operations to utilize the new device qubit logic.
• Added unit tests to validate the functionality of qubit consolidation and error handling for exceeding device qubit limits.

Closes #187

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 97.58065% with 3 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/pyqasm/transformer.py	96.92%	2 Missing ⚠️
src/pyqasm/visitor.py	97.82%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[TheGupta2012]

Hi @vinayswamik , the changes look good!

A point to consider for the design -

Consolidation of qubits inside the qasm operations should be done as a "finalize" step. Presently, you are updating the qubit identifiers directly during the unroll / validate phase and subsequently adding the __PYQASM_QUBITS__ declaration. While it works for the test cases, this transformation logic should be separated in a new step.

A cleaner approach would be to refactor the transformation of statements (which you have done inside each of reset, barrier, etc.) as part of the _qubit_register_consolidation method. It should read the unrolled statements and transform the qubits for each of them, possibly utilizing the Qasm3Transformer module. I only see self._get_pyqasm_device_qubit_index being utilized in your changes and can be decoupled from the visitor code.

[vinayswamik]

Hi @vinayswamik , the changes look good!

A point to consider for the design -

Consolidation of qubits inside the qasm operations should be done as a "finalize" step. Presently, you are updating the qubit identifiers directly during the unroll / validate phase and subsequently adding the __PYQASM_QUBITS__ declaration. While it works for the test cases, this transformation logic should be separated in a new step.

A cleaner approach would be to refactor the transformation of statements (which you have done inside each of reset, barrier, etc.) as part of the _qubit_register_consolidation method. It should read the unrolled statements and transform the qubits for each of them, possibly utilizing the Qasm3Transformer module. I only see self._get_pyqasm_device_qubit_index being utilized in your changes and can be decoupled from the visitor code.

Yeah,I thought about doing all the changes in finalize. But, the reason why i did changes inside the visit operations is, stmts like branching statements are still exists as branching statements even after unrolling.
If we want to do changes inside them, we might need to go again into _visit_branching_statement and get all possible statements that require changes. which costs additional code inside _visit_branching_statementand outside as well.

My suggestion would be move all the current code in (reset, barrier, measurement, gate) to a new function in Qasm3Transformer. then invoke that only when consolidate_qubits == True in visit operations.

And also we can do consolidation with/without providing device_qubits. Here we consider device_qubits = total qubits in the code.

[TheGupta2012]

Yeah,I thought about doing all the changes in finalize. But, the reason why i did changes inside the visit operations is, stmts like branching statements are still exists as branching statements even after unrolling. If we want to do changes inside them, we might need to go again into _visit_branching_statement and get all possible statements that require changes. which costs additional code inside _visit_branching_statementand outside as well.

My suggestion would be move all the current code in (reset, barrier, measurement, gate) to a new function in Qasm3Transformer. then invoke that only when consolidate_qubits == True in visit operations.

And also we can do consolidation with/without providing device_qubits. Here we consider device_qubits = total qubits in the code.

Yea, that's a good point, the branches might exist in the code even after unroll. As you said, let's refactor the current code to the transformer and use the flag consolidate_qubits to trigger it.

[TheGupta2012]

Besides the minor comment fix, changes lgtm @vinayswamik ! Once the CI fix is merged, can you rebase as well? Should be good to merge post that

[TheGupta2012]

Lgtm

[TheGupta2012]

Lgtm