By default, circuits constructed by the pyAQASM library have a somewhat compressed structure.
Circuits consists in a main body, and each gate might have a subcircuit implementation that recursively describes the flow of the circuit.
Lets consider, for instance, a QFT adder:
# we define a simple function that prints the instructions in the main body of a circuit:
def print_main(circuit):
from qat.core.util import extract_syntax
for op in circuit.ops:
print(extract_syntax(circuit.gateDic[op.gate], circuit.gateDic), op.qbits)
from qat.lang.AQASM import *
from qat.lang.AQASM.qftarith import add
prog = Program()
qbits = prog.qalloc(4)
prog.apply(add(2, 2), qbits)
circuit = prog.to_circ()
print_main(circuit)
This is not very explicit. In fact, if one iterates over the circuit, we can see that the circuit contains many more instructions:
for op in circuit.iterate_simple():
print(op)
What happened here is that the iterator iterate_simple emulated an execution stack when iterating over the circuit and unfolded the full circuit.
Equivalently, one could simply ask the to_circ method to unfold all the circuit:
circuit_full = prog.to_circ(inline=True)
print_main(circuit)
CircuitInliner is a plugin that performs this inlining process in the stack. It should be used in case some other plugins requires an inlined circuit to run properly.
from qat.plugins import CircuitInliner
from qat.core import Batch
plugin = CircuitInliner()
print("==== Before 'compile'")
print_main(circuit)
batch = Batch(jobs=[circuit.to_job()])
batch = plugin.compile(batch, None)
print("==== After 'compile'")
print_main(circuit)
As you can see, the circuit was modified in place. This behavior can be changed via the 'inplace' parameter:
circuit = prog.to_circ()
plugin = CircuitInliner(inplace=False)
print("==== Before 'compile'")
print_main(circuit)
batch = Batch(jobs=[circuit.to_job()])
batch = plugin.compile(batch, None)
print("==== After 'compile'")
print_main(circuit)
print("==== output:")
print_main(batch.jobs[0].circuit)