alarsyo/spot
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

bdd_to_formula: add CNF variant

Browse source 
* spot/twa/formula2bdd.hh,
spot/twa/formula2bdd.cc (bdd_to_cnf_formula): New function.
* python/spot/__init__.py: Add a default dictionary for convenience.
* tests/python/bdditer.py: Add test cases.
* NEWS: Mention it.
This commit is contained in:
Alexandre Duret-Lutz 2023-02-23 11:53:07 +01:00
parent 8a5b86521c
commit 66839b1a29
5 changed files with 111 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

61
tests/python/bdditer.py
View file

@ -1,5 +1,5 @@
# -*- mode: python; coding: utf-8 -*-
# Copyright (C) 2017, 2018, 2021, 2022 Laboratoire de Recherche et
# Copyright (C) 2017, 2018, 2021, 2022, 2023 Laboratoire de Recherche et
# Développement de l'Epita (LRDE).
#
# This file is part of Spot, a model checking library.
@ -27,6 +27,19 @@ import sys
from unittest import TestCase
tc = TestCase()
# CPython use reference counting, so that automata are destructed
# when we expect them to be. However other implementations like
# PyPy may call destructors latter, causing different output.
from platform import python_implementation
if python_implementation() == 'CPython':
def gcollect():
pass
else:
import gc
def gcollect():
gc.collect()
run = spot.translate('a & !b').accepting_run()
b = run.prefix[0].label
c = buddy.bdd_satone(b)
@ -43,12 +56,15 @@ while c != buddy.bddtrue:
c = h
tc.assertEqual(res, [0, -1])
del res
res2 = []
for i in run.aut.ap():
res2.append((str(i), run.aut.register_ap(i)))
tc.assertEqual(str(res2), "[('a', 0), ('b', 1)]")
del res2
del c
gcollect()
f = spot.bdd_to_formula(b)
tc.assertTrue(f._is(spot.op_And))
@ -56,9 +72,50 @@ tc.assertTrue(f[0]._is(spot.op_ap))
tc.assertTrue(f[1]._is(spot.op_Not))
tc.assertTrue(f[1][0]._is(spot.op_ap))
tc.assertEqual(str(f), 'a & !b')
del f
gcollect()
try:
f = spot.bdd_to_formula(b, spot.make_bdd_dict())
sys.exit(2)
except RuntimeError as e:
tc.assertIn("not in the dictionary", str(e))
f = spot.bdd_to_cnf_formula(b)
tc.assertEqual(str(f), 'a & !b')
del run
del f
gcollect()
f = spot.bdd_to_cnf_formula(buddy.bddtrue)
tc.assertEqual(str(f), '1')
del f
gcollect()
f = spot.bdd_to_cnf_formula(buddy.bddfalse)
tc.assertEqual(str(f), '0')
del f
gcollect()
aut = spot.translate('(a & b) <-> c')
# With pypy, running GC here will destroy the translator object used
# by translate(). That object has temporary automata that reference
# the BDDs variables and those affect the order in which the
# bdd_to_formula() result is object is presented. The different order
# is not wrong, but it makes it diffuclt to write tests.
gcollect()
for e in aut.out(aut.get_init_state_number()):
b = e.cond
break
f1 = spot.bdd_to_formula(b)
tc.assertEqual(str(f1), '(!a & !c) | (a & b & c) | (!b & !c)')
f2 = spot.bdd_to_cnf_formula(b)
tc.assertEqual(str(f2), '(a | !c) & (!a | !b | c) & (b | !c)')
b1 = spot.formula_to_bdd(f1, spot._bdd_dict, aut)
b2 = spot.formula_to_bdd(f2, spot._bdd_dict, aut)
tc.assertEqual(b1, b2)