Clarification on applying individual transformations #238
Description
My goal is to apply a single transformation to a graph that satisfies the 'graph-like' precondition.
First approach
I tried to apply the following transformations mentioned in #199 (to_gh and spider_simp):
fname = os.path.join('..','circuits','Fast', 'mod5_4_before')
circ = zx.Circuit.load(fname).to_basic_gates()
x = circ.to_graph() # We first have to convert the circuit into a ZX-graph
initial_graph = copy.deepcopy(x)
# turn all red spiders into green spiders
zx.to_gh(initial_graph)
# simplify: remove excess HAD's, fuse along non-HAD edges, remove parallel edges and self-loops
zx.simplify.spider_simp(initial_graph, quiet=True)
# checking transformed graph
print(f'zx.simplify.is_graph_like:{zx.simplify.is_graph_like(initial_graph)}')
Output:
zx.simplify.is_graph_like:False
But the graph was not 'graph-like'.
Second approach
My second approach was using specific functions to convert to 'graph-like'.
I converted the 'mod5_4_before' circuit to a graph, and later to a 'graph-like' graph using the 'to_graph_like' function.
I have to say that to_graph_like function appears in a code block with a note:
"#The functions below haven't been updated in a while. Use at your own risk.", although the function appears to implement the requirements necessary for a graph to satisfy the 'graph-like' condition according to 'Hybrid quantum-classical circuit simplification
with the ZX-calculus'. (page 6).
After this, I applied the transformation 'zx.rules.remove_ids' to the graph of type 'graph-like', and checked the type of graph using the 'is_graph_like' function, this function returned False -> so, it was not a graph of type 'graph-like'.
fname = os.path.join('.','circuits','Fast', 'mod5_4_before')
circ = zx.Circuit.load(fname).to_basic_gates()
x = circ.to_graph() # We first have to convert the circuit into a ZX-graph
graph = copy.deepcopy(x)
print('--------------------------------------')
graph_aux = copy.deepcopy(graph)
zx.simplify.to_graph_like(graph_aux)
print(f'zx.simplify.is_graph_like(graph_aux):{zx.simplify.is_graph_like(graph_aux)}')
print('--------------------------------------')
zx.draw(graph_aux, labels=True)
print('--------------------------------------')
matcher_params = zx.rules.match_ids_parallel(graph_aux)
print(f'matcher_params[0]:{matcher_params[0]}')
print('--------------------------------------')
zx.rules.apply_rule(graph_aux, zx.rules.remove_ids,[matcher_params[0]])
zx.draw(graph_aux, labels=True)
print(f'zx.simplify.is_graph_like:{zx.simplify.is_graph_like(graph_aux)}')
Output:
As I said at the beginning of this comment, my goal is to apply a transform to a circuit.
In this case, I tested the 'zx.rules.remove_ids' transformation, and the result does not seem to be satisfactory.
I am very confused, please:
- Could someone explain to me this result?
And - Could someone tell me how can I apply any transformation individually without getting errors?