def _create_cfg_from_regex(cls, head, regex, track_variables):
dfa = regex.to_epsilon_nfa().to_deterministic().minimize()
transitions = dfa._transition_function._transitions
state_to_var = {}
productions, terminals, variables = set(), set(), set()
for state in dfa.states:
state_to_var[state] = cfg.Variable(f'{state}:{cls.__var_state_counter}')
cls.__var_state_counter += 1
variables.update(state_to_var.values())
for start_state in dfa.start_states:
productions.add(cfg.Production(head, [state_to_var[start_state]]))
for state_from in transitions:
for edge_symb in transitions[state_from]:
state_to = transitions[state_from][edge_symb]
current_prod_head = state_to_var[state_from]
current_prod_body = []
if any(letter.isupper() for letter in edge_symb.value) and not track_variables:
var = cfg.Variable(edge_symb.value)
variables.add(var)
current_prod_body.append(var)
elif edge_symb.value != 'eps':
term = cfg.Terminal(edge_symb.value)
terminals.add(term)
current_prod_body.append(term)
current_prod_body.append(state_to_var[state_to])
productions.add(cfg.Production(current_prod_head, current_prod_body))
if state_to in dfa.final_states:
productions.add(cfg.Production(state_to_var[state_to], []))
if not productions:
return cfg.CFG(variables, terminals, head, {cfg.Production(head, [])})
return cfg.CFG(variables, terminals, head, productions)
def to_normal_form(self):
new = super().to_normal_form()
if self.is_eps_reachable:
new.productions.add(
cfg.Production(self.start_symbol, [])
)
return custom_CFG.from_CFG(new)
def _initialize_production_from_start_in_to_cfg(self, start):
productions = []
for state in self._states:
productions.append(
cfg.Production(start, [
self._cfg_variable_converter.to_cfg_combined_variable(
self._start_state, self._start_stack_symbol, state)
]))
return productions
def _process_transition_and_state_to_cfg_safe(self, productions, state,
transition):
head = self._get_head_from_state_and_transition(state, transition)
bodies = self._get_all_bodies_from_state_and_transition(
state, transition)
if transition[INPUT][INPUT_SYMBOL] != Epsilon():
_prepend_input_symbol_to_the_bodies(bodies, transition)
for body in bodies:
productions.append(cfg.Production(head, body, filtering=False))
def _read_line(cls, line, productions, terminals, variables):
prod_s = line.split(' ', 1)
head_s = prod_s[0]
if len(prod_s) > 1:
body_s = prod_s[1]
else:
body_s = ''
head = cfg.Variable(head_s.strip())
variables.add(head)
body = []
for body_component in body_s.split():
if body_component[0] in string.ascii_uppercase:
body_var = cfg.Variable(body_component)
variables.add(body_var)
body.append(body_var)
else:
body_ter = cfg.Terminal(body_component)
terminals.add(body_ter)
body.append(body_ter)
productions.add(cfg.Production(head, body))