def output_bqpjson(outfile, as_qubo, problem):
"Output weights and strengths in bqpjson format, either Ising or QUBO."
# Prepare the "easy" fields.
bqp = {}
bqp["version"] = "1.0.0"
bqp["id"] = random.randint(2**20, 2**60)
bqp["scale"] = 1.0
bqp["offset"] = 0.0
if as_qubo:
bqp["variable_domain"] = "boolean"
else:
bqp["variable_domain"] = "spin"
# Prepare the list of all variables.
var_ids = set(problem.weights.keys())
for q1, q2 in problem.strengths.keys():
var_ids.add(q1)
var_ids.add(q2)
bqp["variable_ids"] = sorted(var_ids)
# Prepare the linear terms.
lin_terms = []
for q, wt in sorted(problem.weights.items()):
lin_terms.append({
"id": q,
"coeff": wt})
bqp["linear_terms"] = lin_terms
# Prepare the quadratic terms.
quad_terms = []
strengths = qmasm.canonicalize_strengths(problem.strengths)
for (q1, q2), wt in sorted(strengths.items()):
quad_terms.append({
"id_tail": q1,
"id_head": q2,
"coeff": wt})
bqp["quadratic_terms"] = quad_terms
# Prepare some metadata.
metadata = {}
if as_qubo:
metadata["description"] = "QUBO problem compiled by QMASM (https://github.com/lanl/qmasm)"
else:
metadata["description"] = "Ising problem compiled by QMASM (https://github.com/lanl/qmasm)"
metadata["command_line"] = qmasm.get_command_line()
metadata["generated"] = datetime.datetime.utcnow().isoformat()
if hasattr(problem, "embedding"):
# Physical problem
def attempt_assign(key, func):
"Try assigning a key, but don't complain if we can't."
try:
metadata[key] = func()
except KeyError:
pass
attempt_assign("dw_url", lambda: os.environ["DW_INTERNAL__HTTPLINK"])
attempt_assign("dw_solver_name", lambda: qmasm.solver_name)
props = qmasm.solver.properties
attempt_assign("dw_chip_id", lambda: props["chip_id"])
L, M, N = qmasm.chimera_topology(qmasm.solver)
metadata["chimera_cell_size"] = L*2
metadata["chimera_degree"] = max(M, N)
metadata["equivalent_ids"] = sorted(problem.chains)
metadata["variable_names"] = {s: problem.embedding[n]
for s, n in qmasm.sym_map.symbol_number_items()}
else:
metadata["variable_names"] = {s: [n]
for s, n in qmasm.sym_map.symbol_number_items()}
bqp["metadata"] = metadata
# Output the problem in JSON format.
outfile.write(json.dumps(bqp, indent=2, sort_keys=True) + "\n")
def output_bqpjson(outfile, as_qubo, problem):
"Output weights and strengths in bqpjson format, either Ising or QUBO."
# Prepare the "easy" fields.
bqp = {}
bqp["version"] = "1.0.0"
bqp["id"] = random.randint(2**20, 2**60)
bqp["scale"] = 1.0
bqp["offset"] = 0.0
if as_qubo:
bqp["variable_domain"] = "boolean"
else:
bqp["variable_domain"] = "spin"
# Prepare the list of all variables.
var_ids = set(problem.weights.keys())
for q1, q2 in problem.strengths.keys():
var_ids.add(q1)
var_ids.add(q2)
bqp["variable_ids"] = sorted(var_ids)
# Prepare the linear terms.
lin_terms = []
for q, wt in sorted(problem.weights.items()):
lin_terms.append({
"id": q,
"coeff": wt})
bqp["linear_terms"] = lin_terms
# Prepare the quadratic terms.
quad_terms = []
strengths = qmasm.canonicalize_strengths(problem.strengths)
for (q1, q2), wt in sorted(strengths.items()):
quad_terms.append({
"id_tail": q1,
"id_head": q2,
"coeff": wt})
bqp["quadratic_terms"] = quad_terms
# Prepare some metadata.
metadata = {}
if as_qubo:
metadata["description"] = "QUBO problem compiled by QMASM (https://github.com/lanl/qmasm)"
else:
metadata["description"] = "Ising problem compiled by QMASM (https://github.com/lanl/qmasm)"
metadata["command_line"] = qmasm.get_command_line()
metadata["generated"] = datetime.datetime.utcnow().isoformat()
if hasattr(problem, "embedding"):
# Physical problem
def attempt_assign(key, func):
"Try assigning a key, but don't complain if we can't."
try:
metadata[key] = func()
except KeyError:
pass
attempt_assign("dw_url", lambda: os.environ["DW_INTERNAL__HTTPLINK"])
attempt_assign("dw_solver_name", lambda: qmasm.solver_name)
props = qmasm.solver.properties
attempt_assign("dw_chip_id", lambda: props["chip_id"])
L, M, N = qmasm.chimera_topology(qmasm.solver)
metadata["chimera_cell_size"] = L*2
metadata["chimera_degree"] = max(M, N)
metadata["equivalent_ids"] = sorted(problem.chains)
metadata["variable_names"] = {s: problem.embedding[n]
for s, n in qmasm.sym_map.symbol_number_items()}
else:
metadata["variable_names"] = {s: [n]
for s, n in qmasm.sym_map.symbol_number_items()}
bqp["metadata"] = metadata
# Output the problem in JSON format.
outfile.write(json.dumps(bqp, indent=2, sort_keys=True) + "\n")