def cfpq_hellings(g: LabelGraph, cfg: GrammarCNF):
num_vert = g.num_vert
start_sym = cfg.start_symbol
result = LabelGraph()
result.num_vert = num_vert
m = deque()
for variable in cfg.variables:
result.graph_dict[variable] = Matrix.sparse(BOOL, num_vert, num_vert)
if cfg.generate_epsilon():
for v in range(num_vert):
result.graph_dict[start_sym][v, v] = True
for label in g.graph_dict:
term = Terminal(label)
result.graph_dict[term] = g.graph_dict[label].dup()
for v_from, v_to in g.get_edges(label):
for production in cfg.productions:
if (len(production.body) == 1 and production.body[0] == term):
head = production.head
result.graph_dict[head][v_from, v_to] = True
for label in result.graph_dict:
for i, j in result.get_edges(label):
m.append((label, i, j))
# 3rd step: cfpq on modified matrix
while m:
var, v, u = m.popleft()
for var_left in result.graph_dict:
for v_new, v_ in result.get_edges(var_left):
if (v_ == v):
for production in cfg.pair_productions:
if (production.body[1] == var
and production.body[0] == var_left):
if (v_new, u) not in result.get_edges(
production.head):
result.graph_dict[production.head][v_new,
u] = True
m.append((production.head, v_new, u))
for var_right in result.graph_dict:
for u_, u_new in result.get_edges(var_right):
if (u_ == u):
for production in cfg.pair_productions:
if (production.body[1] == var_right
and production.body[0] == var):
if (v, u_new) not in result.get_edges(
production.head):
result.graph_dict[production.head][
v, u_new] = True
m.append((production.head, v, u_new))
return result.graph_dict[start_sym]
def test_cfpq_grammar_2():
for i in range(NUM_GRAPHS):
g_2 = GrammarCNF.from_text(TEST_GRAMMARS[1])
for i in range(NUM_GRAPHS):
graph = LabelGraph().from_txt(
os.path.join(DATA_DIR, f'graph_{i}.txt'))
result = cfpq_matrix_mult(graph, g_2)
expected = set()
with open(os.path.join(DATA_DIR, f'expected_{1}_{i}.txt'),
'r') as f:
for line in f:
v, to = line.split(' ')
expected.add((int(v), int(to)))
edges = set(LabelGraph.get_reachable(result))
assert edges == expected
def test_cfpq_brackets():
for i in range(NUM_GRAPHS):
brackets_cnf = GrammarCNF.from_text(TEST_GRAMMARS[0])
for i in range(NUM_GRAPHS):
graph = LabelGraph().from_txt(
os.path.join(DATA_DIR, f'graph_{i}.txt'))
result = cfpq_hellings(graph, brackets_cnf)
expected = set()
with open(os.path.join(DATA_DIR, f'expected_{0}_{i}.txt'),
'r') as f:
for line in f:
v, to = line.split(' ')
expected.add((int(v), int(to)))
edges = set(LabelGraph.get_reachable(result))
assert edges == expected
def cfpq_matrix_mult(g: LabelGraph, cfg: GrammarCNF):
num_vert = g.num_vert
if (num_vert == 0):
return Matrix.sparse(BOOL, num_vert, num_vert)
result = LabelGraph()
start_sym = cfg.start_symbol
result.num_vert = num_vert
for variable in cfg.variables:
result.graph_dict[variable] = Matrix.sparse(BOOL, num_vert, num_vert)
for label in g.graph_dict:
term = Terminal(label)
result.graph_dict[term] = g.graph_dict[label].dup()
for v_from, v_to in g.get_edges(label):
for production in cfg.productions:
if (len(production.body) == 1 and production.body[0] == term):
head = production.head
result.graph_dict[head][v_from, v_to] = True
if cfg.generate_epsilon():
for v in g.vertices:
result.graph_dict[start_sym][v, v] = True
matrix_changing = True
with semiring.LOR_LAND_BOOL:
while matrix_changing:
matrix_changing = False
for production in cfg.pair_productions:
head = production.head
body = production.body
prev_nvals = result.graph_dict[head].nvals
tmp = result.graph_dict[body[0]] @ result.graph_dict[body[1]]
result.graph_dict[head] = result.graph_dict[head] + tmp
if (prev_nvals != result.graph_dict[head].nvals):
matrix_changing = True
return result.graph_dict[start_sym]
def test_cfpq_empty_graph():
brackets_cnf = GrammarCNF.from_text(TEST_GRAMMARS[0])
result = cfpq_matrix_mult(LabelGraph(), brackets_cnf)
expected = set()
edges = set(LabelGraph.get_reachable(result))
assert edges == expected
import os
from src.antlr_utils import parse
from src.grammar_cnf import GrammarCNF
import pytest
@pytest.mark.parametrize("grammar", [GrammarCNF.from_txt("dbql_grammar.txt")])
@pytest.mark.parametrize(
"test_input, expected",
[
('''
connect "azat/home/db" ;
select edges
from query term("s")*|term("b")+.term("c")?;
''', True),
('''
select edges from name "sparsegraph" ;
''', True),
('''
connect "azat/home/db" ;
''', True),
('''
connect "azat/home/db" ;
select edges from name "sparsegraph_256.txt" ;
''', True),
('''
connect "azat/home/db" ;
select edges
from startAndFinal(set(1, 2, 3), set (4, 5, 6))
of name "sparsegraph" ;
''', True),
def test_1_word_accepts():
cnf = GrammarCNF.from_text(TEST_GRAMMARS[0])
word_accepts = 'aabb'
assert cnf.contains(word_accepts) is True
def test_2_epsilon_not_accepts():
cnf = GrammarCNF.from_text(TEST_GRAMMARS[1])
assert cnf.contains('') is False
def test_3_accepts():
cnf = GrammarCNF.from_text('S -> \( S \) | eps')
assert cnf.contains('()')
assert cnf.contains('(())')
def test_2_accepts():
cnf = GrammarCNF.from_text(TEST_GRAMMARS[1])
assert cnf.contains('abab')
assert cnf.contains('ab')
def test_1_epsilon_accepts():
cnf = GrammarCNF.from_text(TEST_GRAMMARS[0])
word_epsilon = ''
assert cnf.contains(word_epsilon) is True