I am exploring fast ways to perform SAT solving in Z3 (Python). To do so, I am trying to imitate the results of Chapter 5.1 of https://theory.stanford.edu/~nikolaj/programmingz3.html#sec-blocking-evaluations.
The code I am using is the following:
def block_modelT(s, terms): #I used the name 'block_modelT' so as not to use 'block_model'
m = s.model()
s.add(Or([t != m.eval(t) for t in terms]))
def all_smt(s, terms):
while sat == s.check():
print(s.model())
block_modelT(s, terms)
Thus, I execute it:
pip install z3-solver
from z3 import *
x, y, z = Bools('x y z')
vars = [x,y,z]
phi_0 = x #ignore that I call them phi: it is because I am used to do it like that for predicates (i.e. literals) of other theories: e.g., if x is Int, then phi_0 = (x>0)
phi_1 = y
phi_2 = z
phi = And(phi_0, Or(phi_1, phi_2))
all_smt(s, vars)
And get the following error:
1 def block_modelT(s, terms):
2 m = s.model()
----> 3 s.add(Or([t != m.eval(t) for t in terms]))
TypeError: 'builtin_function_or_method' object is not iterable
Any help?
EDIT:
The problem has been solved so I finally try the next piece of code:
def all_smt(s, initial_terms):
def block_term(s, m, t):
s.add(t != m.eval(t))
def fix_term(s, m, t):
s.add(t == m.eval(t))
def all_smt_rec(terms):
if sat == s.check():
m = s.model()
yield m
for i in range(len(terms)):
s.push()
block_term(s, m, terms[i])
for j in range(i):
fix_term(s, m, terms[j])
for m in all_smt_rec(terms[i:]):
yield m
s.pop()
for m in all_smt_rec(list(initial_terms)):
yield m
I execute it: all_smt(s, varss) #has the same name
. And get a generator <generator object all_smt at 0x7...
instead of a valid assignment (i.e. a model). How can obtain the correct answer [z = False, y = True, x = True], [z = True, x = True]
? I tried both print
and return
.
Your program doesn't define an s
; and you haven't added the formula to it. The following works for me:
from z3 import *
def block_modelT(s, terms): #I used the name 'block_modelT' so as not to use 'block_model'
m = s.model()
s.add(Or([t != m.eval(t) for t in terms]))
def all_smt(s, terms):
while sat == s.check():
print(s.model())
block_modelT(s, terms)
x, y, z = Bools('x y z')
vars = [x,y,z]
phi_0 = x #ignore that I call them phi: it is because I am used to do it like that for predicates (i.e. literals) of other theories: e.g., if x is Int, then phi_0 = (x>0)
phi_1 = y
phi_2 = z
phi = And(phi_0, Or(phi_1, phi_2))
s = Solver()
s.add(phi)
all_smt(s, vars)
This prints:
[z = False, y = True, x = True]
[z = True, x = True]