def parse_table(name, table, subsets):
lines = table.split('\n')
if len(lines) < 4:
raise ValueError,\
"Rule %s has too few lines to be an FSA table." % name
pairs1 = lines[1].strip().split()
pairs2 = lines[2].strip().split()
if len(pairs1) != len(pairs2):
raise ValueError,\
"Rule %s has pair definitions that don't line up." % name
pairs = [KimmoPair(p1, p2) for p1, p2 in zip(pairs1, pairs2)]
finals = []
fsa = FSA()
for line in lines[3:]:
line = line.strip()
if not line: continue
groups = re.match(r'(\w+)(\.|:)\s*(.*)', line)
if groups is None:
raise ValueError,\
"Can't parse this line of the state table for rule %s:\n%s"\
% (name, line)
state, char, morestates = groups.groups()
if fsa.start() == 0: fsa.set_start(state)
if char == ':': finals.append(state)
fsa.add_state(state)
morestates = morestates.split()
if len(morestates) != len(pairs):
raise ValueError,\
"Rule %s has a row of the wrong length:\n%s\ngot %d items, should be %d"\
% (name, line, len(morestates), len(pairs))
for pair, nextstate in zip(pairs, morestates):
fsa.insert_safe(state, pair, nextstate)
fsa.set_final(finals)
return KimmoFSARule(name, fsa, subsets)
def from_dfa_dict(name, states, subsets):
fsa = FSA()
pairs = set([KimmoPair.make('@')])
for (statename, trans) in states.items():
for label in trans:
if label != 'others':
pairs.add(KimmoPair.make(label))
for (statename, trans) in states.items():
parts = statename.split()
source = parts[-1]
if not parts[0].startswith('rej'):
fsa.add_final(source)
if fsa.start() == 0 and source in ['begin', 'Begin', '1', 1]:
fsa.set_start(source)
if source in ['start', 'Start']:
fsa.set_start(source)
used_pairs = set()
for label in trans:
if label != 'others':
used_pairs.add(KimmoPair.make(label))
for label, target in trans.items():
if label.lower() == 'others':
fsa.insert_safe(source, KimmoPair.make('@'), target)
for pair in pairs.difference(used_pairs):
fsa.insert_safe(source, pair, target)
else:
fsa.insert_safe(source, KimmoPair.make(label), target)
return KimmoFSARule(name, fsa, subsets)
def parse_table(name, table, subsets):
lines = table.split('\n')
if len(lines) < 4:
raise ValueError,\
"Rule %s has too few lines to be an FSA table." % name
pairs1 = lines[1].strip().split()
pairs2 = lines[2].strip().split()
if len(pairs1) != len(pairs2):
raise ValueError,\
"Rule %s has pair definitions that don't line up." % name
pairs = [KimmoPair(p1, p2) for p1, p2 in zip(pairs1, pairs2)]
finals = []
fsa = FSA()
for line in lines[3:]:
line = line.strip()
if not line: continue
groups = re.match(r'(\w+)(\.|:)\s*(.*)', line)
if groups is None:
raise ValueError,\
"Can't parse this line of the state table for rule %s:\n%s"\
% (name, line)
state, char, morestates = groups.groups()
if fsa.start() == 0: fsa.set_start(state)
if char == ':': finals.append(state)
fsa.add_state(state)
morestates = morestates.split()
if len(morestates) != len(pairs):
raise ValueError,\
"Rule %s has a row of the wrong length:\n%s\ngot %d items, should be %d"\
% (name, line, len(morestates), len(pairs))
for pair, nextstate in zip(pairs, morestates):
fsa.insert_safe(state, pair, nextstate)
fsa.set_final(finals)
return KimmoFSARule(name, fsa, subsets)
def from_dfa_dict(name, states, subsets):
fsa = FSA()
pairs = set([KimmoPair.make('@')])
for (statename, trans) in states.items():
for label in trans:
if label != 'others':
pairs.add(KimmoPair.make(label))
for (statename, trans) in states.items():
parts = statename.split()
source = parts[-1]
if not parts[0].startswith('rej'):
fsa.add_final(source)
if fsa.start() == 0 and source in ['begin', 'Begin', '1', 1]:
fsa.set_start(source)
if source in ['start', 'Start']:
fsa.set_start(source)
used_pairs = set()
for label in trans:
if label != 'others':
used_pairs.add(KimmoPair.make(label))
for label, target in trans.items():
if label.lower() == 'others':
fsa.insert_safe(source, KimmoPair.make('@'), target)
for pair in pairs.difference(used_pairs):
fsa.insert_safe(source, pair, target)
else:
fsa.insert_safe(source, KimmoPair.make(label), target)
return KimmoFSARule(name, fsa, subsets)
def _parse_context(self, tokens, i, reverse):
(j, tree) = self._parse_list(tokens, i)
if j == i: return (i, None)
sigma = set()
self._collect_alphabet(tree, sigma)
fsa = FSA(sigma)
final_state = self._build_fsa(fsa, fsa.start(), tree, reverse)
fsa.set_final([final_state])
#fsa.pp()
dfa = fsa.dfa()
#dfa.pp()
dfa.prune()
#dfa.pp()
return (j, dfa)
def _parse_context(self, tokens, i, reverse):
(j, tree) = self._parse_list(tokens, i)
if j == i: return (i, None)
sigma = set()
self._collect_alphabet(tree, sigma)
fsa = FSA(sigma)
final_state = self._build_fsa(fsa, fsa.start(), tree, reverse)
fsa.set_final([final_state])
#fsa.pp()
dfa = fsa.dfa()
#dfa.pp()
dfa.prune()
#dfa.pp()
return (j, dfa)
