def manage_sb_compilation(phi, psi, m_atom, a, E):
if OPTIMIZED:
monitoring_atom = ds.Literal(m_atom, False)
added_precond = (None, False)
if can_make_true(a, phi):
R_phi = regression_aux(phi, a)
rho = ds.Or([R_phi.negate(), monitoring_atom]).simplified()
added_precond = (rho, True)
if can_make_true(a, psi):
R_psi = regression_aux(psi, a)
if R_psi != psi:
add_cond_eff(E, create_cond_eff(R_psi, monitoring_atom))
return added_precond
else:
monitoring_atom = ds.Literal(m_atom, False)
R_phi = regression_aux(phi, a)
if R_phi == phi:
added_precond = (None, False)
else:
rho = ds.Or([R_phi.negate(), monitoring_atom]).simplified()
added_precond = (rho, True)
R_psi = regression_aux(psi, a)
if R_psi != psi:
add_cond_eff(E, create_cond_eff(R_psi, monitoring_atom))
return added_precond
def manage_sometime_compilation(phi, m_atom, a, E):
if OPTIMIZED:
if can_make_true(a, phi):
monitoring_atom = ds.Literal(m_atom, False)
R = regression_aux(phi, a)
add_cond_eff(E, create_cond_eff(R, monitoring_atom))
else:
monitoring_atom = ds.Literal(m_atom, False)
R = regression_aux(phi, a)
if R != phi:
add_cond_eff(E, create_cond_eff(R, monitoring_atom))
def convert_formula(formula):
if isinstance(formula, pddl.Atom):
return ds.Literal(atom_str(formula), False)
elif isinstance(formula, pddl.NegatedAtom):
return ds.Literal(atom_str(formula), True)
elif isinstance(formula, pddl.Disjunction):
return ds.Or([convert_formula(part) for part in formula.parts])
elif isinstance(formula, pddl.Conjunction):
return ds.And([convert_formula(part) for part in formula.parts])
elif isinstance(formula, pddl.Truth):
return ds.TRUE()
elif isinstance(formula, pddl.Falsity):
return ds.FALSE()
def manage_sa_compilation(phi, psi, m_atom, a, E):
if OPTIMIZED:
R1 = regression_aux(phi, a)
R2 = regression_aux(psi, a)
monitoring_atom = ds.Literal(m_atom, False)
if can_make_true(a, phi) or can_falsify(a, psi):
cond = ds.And([R1, R2.negate()]).simplified()
cond_eff = create_cond_eff(cond, monitoring_atom.negate())
add_cond_eff(E, cond_eff)
if can_make_true(a, psi):
add_cond_eff(E, create_cond_eff(R2, monitoring_atom))
else:
R1 = regression_aux(phi, a)
R2 = regression_aux(psi, a)
monitoring_atom = ds.Literal(m_atom, False)
if phi != R1 or psi != R2:
cond = ds.And([R1, R2.negate()]).simplified()
cond_eff = create_cond_eff(cond, monitoring_atom.negate())
add_cond_eff(E, cond_eff)
if psi != R2:
add_cond_eff(E, create_cond_eff(R2, monitoring_atom))
def manage_amo_compilation(phi, m_atom, a, E):
if OPTIMIZED:
if not can_make_true(a, phi):
return None, False
monitoring_atom = ds.Literal(m_atom, False)
R = regression_aux(phi, a)
if R == phi:
return None, False
else:
rho = ds.Or([R.negate(), monitoring_atom.negate(),
phi]).simplified()
add_cond_eff(E, create_cond_eff(R, monitoring_atom))
return rho, True
else:
monitoring_atom = ds.Literal(m_atom, False)
R = regression_aux(phi, a)
if R == phi:
return None, False
else:
rho = ds.Or([R.negate(), monitoring_atom.negate(),
phi]).simplified()
add_cond_eff(E, create_cond_eff(R, monitoring_atom))
return rho, True
def tcore(F, A, I, G, C):
relevancy_dict = build_relevancy_dict(C)
A_prime = []
G_prime = []
I_prime, F_prime = get_monitoring_atoms(C, I)
compiled_action = 0
for c in LTC(C):
monitoring_atom = ds.Literal(c.monitoring_atom, False)
G_prime.append(monitoring_atom)
G_prime = ds.And(G_prime)
for a in A:
E = []
relevant_constraints = get_relevant_constraints(a, relevancy_dict)
if len(relevant_constraints) > 0:
compiled_action += 1
for c in relevant_constraints:
if c.kind == ds.ALWAYS:
precondition, to_add = manage_always_compilation(c.gd1, a)
elif c.kind == ds.ATMOSTONCE:
precondition, to_add = manage_amo_compilation(
c.gd1, c.monitoring_atom, a, E)
elif c.kind == ds.SOMETIMEBEFORE:
precondition, to_add = manage_sb_compilation(
c.gd1, c.gd2, c.monitoring_atom, a, E)
if c.kind == ds.ALWAYS or c.kind == ds.ATMOSTONCE or c.kind == ds.SOMETIMEBEFORE:
if to_add:
a.precondition.append(precondition)
if c.kind == ds.SOMETIME:
manage_sometime_compilation(c.gd1, c.monitoring_atom, a, E)
elif c.kind == ds.SOMETIMEAFTER:
manage_sa_compilation(c.gd1, c.gd2, c.monitoring_atom, a, E)
for eff in E:
a.effects.append(eff)
if ds.FALSE() not in a.precondition:
A_prime.append(a)
G_new = ds.And([G, G_prime]).simplified()
#print("Compiled actions: {}".format(compiled_action))
return F + F_prime, A_prime, I + I_prime, G_new
def get_fresh_monitoring_atom(name, number):
return ds.Literal('{}{}{}'.format(name, SEPARATOR, number), False)