def build_LR1_automaton(self, G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0, lookaheads=(G.EOF,))
start = frozenset([start_item])
closure = self.closure_lr1(start, firsts)
automaton = State(frozenset(closure), True)
pending = [ start ]
visited = { start: automaton }
while pending:
current = pending.pop()
current_state = visited[current]
for symbol in G.terminals + G.nonTerminals:
closure_current = self.closure_lr1(current, firsts)
goto = self.goto_lr1(closure_current, symbol, just_kernel=True)
if len(goto) == 0:
continue
try:
next_state = visited[goto]
except KeyError:
next_closure = self.closure_lr1(goto, firsts)
visited[goto] = next_state = State(frozenset(next_closure), True)
pending.append(goto)
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton
def build_LR0_automaton(G: Grammar):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
automaton = State(start_item, True)
pending = [start_item]
visited = {start_item: automaton}
while pending:
current_item = pending.pop()
if current_item.IsReduceItem:
continue
current_state = visited[current_item]
next_symbol = current_item.NextSymbol
next_item = current_item.NextItem()
#- $(X --> a.cB) ---c---> (X --> ac.B) con c in V_T
#- $(X --> a.YB) ---Y---> (X --> aY.B) con Y in V_N
next_state = get_state(visited, pending, next_item)
current_state.add_transition(next_symbol.Name, next_state)
if next_symbol in G.nonTerminals:
sym_productions = G.symbDict[next_symbol.Name].productions
#- $(X --> a.YB) ---e---> (Y --> .Z) con Y in V_N
for pr in [Item(x, 0) for x in sym_productions]:
trans_state = get_state(visited, pending, pr)
current_state.add_epsilon_transition(trans_state)
return automaton.to_deterministic(lr0_formatter)
def build_lr0_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
start = frozenset([start_item])
automaton = State(closure_lr0(start), True)
pending = [start]
visited = {start: automaton}
while pending:
current = pending.pop()
current_state = visited[current]
current_closure = current_state.state
for symbol in G.terminals + G.nonTerminals:
kernel = goto_lr0(current_closure, symbol)
if kernel == frozenset():
continue
try:
next_state = visited[kernel]
except KeyError:
next_state = visited[kernel] = State(closure_lr0(kernel), True)
pending.append(kernel)
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton
def _build_automaton(self):
start = State('start')
automatons = self.regexs
for i in automatons:
start.add_epsilon_transition(i)
final = start.to_deterministic()
return final
def build_LALR_automaton(G):
assert len(G.startSymbol.productions) == 1, "Grammar must be augmented"
lr1_automaton = build_LR1_automaton(G)
same_kernel = {}
for node in lr1_automaton:
just_center = frozenset([item.Center() for item in node.state])
try:
same_kernel[just_center].append(node)
except KeyError:
same_kernel[just_center] = [node]
start = frozenset([item.Center() for item in lr1_automaton.state
]) # como cabecera solo quedan los items sin lookahead
automaton = State(
lr1_automaton.state, True
) # en visited se guarda el estado que corresponde a la fusion de estaods ocn el mismo nucleo
pending = [start]
visited = {start: automaton}
while pending:
current = pending.pop()
current_state = visited[current] # se van a actualizar
# todos los estados con los que el estado actual tiene alguna transicion
lr1_state = same_kernel[current][0]
# chequear que cada estado del cjto analizado tenga esa transicion
for symbol in G.terminals + G.nonTerminals:
if lr1_state.has_transition(symbol.Name):
state = lr1_state.transitions[symbol.Name][0]
center_items = frozenset(
[item.Center() for item in state.state])
try:
next_state = visited[center_items]
except KeyError:
kernel_set = same_kernel[center_items]
items_with_lookahead = {}
for node in kernel_set:
for item in node.state:
try:
current_item = items_with_lookahead[
item.Center()]
except KeyError:
current_item = items_with_lookahead[
item.Center()] = set()
current_item.update(item.lookaheads)
completed_items = [
Item(item.production, item.pos, lookaheads)
for item, lookaheads in items_with_lookahead.items()
]
next_state = State(frozenset(completed_items), True)
visited[center_items] = next_state
pending.append(center_items)
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton
def DerivationTree(derivation, G):
lent = len(derivation)
nonTerminalstack = []
root = State(G.startSymbol.Name)
nonTerminalstack.append(root)
while lent > 0:
lent -= 1
next_production = derivation[lent]
print("next_production", next_production)
currentNode = nonTerminalstack.pop()
# assert currentNode.state == next_production.Left.Name, "Wrong derivation"
if next_production.IsEpsilon:
currentNode.add_transition(" ", State("epsilon", True))
for symbol in next_production.Right:
if symbol.IsTerminal:
currentNode.add_transition(" ", State(symbol.Name, True))
else:
nonTerminalstack.append(State(symbol.Name))
currentNode.add_transition(
" ", nonTerminalstack[len(nonTerminalstack) - 1])
return root
def build_larl1_automaton(G, firsts=None):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
if not firsts:
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0, lookaheads=ContainerSet(G.EOF))
start = frozenset([start_item.Center()])
closure = closure_lr1([start_item], firsts)
automaton = State(frozenset(closure), True)
pending = [start]
visited = {start: automaton}
while pending:
current = pending.pop()
current_state = visited[current]
current_closure = current_state.state
for symbol in G.terminals + G.nonTerminals:
goto = goto_lr1(current_closure, symbol, just_kernel=True)
closure = closure_lr1(goto, firsts)
center = frozenset(item.Center() for item in goto)
if center == frozenset():
continue
try:
next_state = visited[center]
centers = {item.Center(): item for item in next_state.state}
centers = {
item.Center(): (centers[item.Center()], item)
for item in closure
}
updated_items = set()
for c, (itemA, itemB) in centers.items():
item = Item(c.production, c.pos,
itemA.lookaheads | itemB.lookaheads)
updated_items.add(item)
updated_items = frozenset(updated_items)
if next_state.state != updated_items:
pending.append(center)
next_state.state = updated_items
except KeyError:
visited[center] = next_state = State(frozenset(closure), True)
pending.append(center)
if current_state[symbol.Name] is None:
current_state.add_transition(symbol.Name, next_state)
else:
assert current_state.get(
symbol.Name) is next_state, 'Bad build!!!'
automaton.set_formatter(multiline_formatter)
return automaton
def _build_automaton(self):
start = State('start')
regexs = self.regexs
for regex in regexs:
start.add_epsilon_transition(regex)
return start.to_deterministic()
def _build_parsing_table(self):
G = self.G.AugmentedGrammar(True)
automaton = build_LR1_automaton(G)
centers = {}
for i, node in enumerate(automaton):
node.idx = i
try:
same_center = centers[node_centers(node)]
centers[node_centers(node)] = State(
compress(node.state.union(same_center.state)), True)
except KeyError:
centers[node_centers(node)] = node
centers[node_centers(node)].idx = i
self.automaton = automaton
self.centers = centers
Goto = {}
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
for node in centers.values():
for symbol in G.terminals + G.nonTerminals:
temp = goto_lr1(node.state, symbol, firsts)
if not temp:
continue
Goto[(node.idx,
symbol)] = centers[node_centers(State(temp, True))]
for node in centers.values():
idx = node.idx
for item in node.state:
# Your code here!!!
# - Fill `self.Action` and `self.Goto` according to `item`)
# - Feel free to use `self._register(...)`)
if item.IsReduceItem:
production = item.production
if production.Left == G.startSymbol:
self._register(self.action, (idx, G.EOF), ("OK", None))
else:
for lookahead in item.lookaheads:
self._register(self.action, (idx, lookahead),
("REDUCE", production))
else:
next_symbol = item.NextSymbol
if next_symbol.IsTerminal:
self._register(
self.action, (idx, next_symbol),
("SHIFT", Goto[(node.idx, next_symbol)].idx))
else:
self._register(self.goto, (idx, next_symbol),
Goto[(node.idx, next_symbol)].idx)
def build_LR0_automaton(G):
assert len(G.startSymbol.productions) == 1, "Grammar must be augmented"
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
automaton = State(start_item, True)
pending = [start_item]
visited = {start_item: automaton}
while pending:
current_item = pending.pop()
if current_item.IsReduceItem:
continue
# (Decide which transitions to add)
# agregar las epsilon transiciones
# a estados donde el item posee producciones a partir del simbolo actual en la posicion 0
# y agregar la transicion a partir del simbolo siguiente
next_item = current_item.NextItem()
try:
next_state = visited[next_item]
except KeyError:
next_state = State(next_item, True)
visited[next_item] = next_state
pending.append(next_item)
if current_item.NextSymbol.IsNonTerminal:
epsilon_productions = current_item.NextSymbol.productions
else:
epsilon_productions = None
current_state = visited[current_item]
# (Adding the decided transitions)
current_state.add_transition(current_item.NextSymbol.Name, next_state)
if epsilon_productions:
for eproduction in epsilon_productions:
epItem = Item(eproduction, 0)
try:
epState = visited[epItem]
except KeyError:
epState = State(epItem, True)
visited[epItem] = epState
pending.append(epItem)
current_state.add_epsilon_transition(epState)
return automaton
def build_LALR1_automaton(G):
automaton = build_LR1_automaton(G)
stKernel = {}
for node in automaton:
kernel = frozenset([item.Center() for item in node.state])
try:
stKernel[kernel].append(node)
except KeyError:
stKernel[kernel] = [node]
initial = frozenset([item.Center() for item in automaton.state])
automaton = State(automaton.state, True)
visited = {initial: automaton}
pending = [initial]
while pending:
current = pending.pop()
current_state = visited[current]
lr1_state = stKernel[current][0]
for symbol in G.terminals + G.nonTerminals:
if symbol.Name in lr1_state.transitions:
dest_core = frozenset([
item.Center()
for item in lr1_state.transitions[symbol.Name][0].state
])
try:
next_state = visited[dest_core]
except KeyError:
union_core = {center: set() for center in dest_core}
for node in stKernel[dest_core]:
for item in node.state:
union_core[item.Center()].update(item.lookaheads)
union_core = frozenset([
Item(center.production,
center.pos,
lookaheads=lookaheads)
for center, lookaheads in union_core.items()
])
next_state = State(union_core, True)
visited[dest_core] = next_state
pending.append(dest_core)
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton
def get_state(visited, pending, item):
if not item in visited.keys():
new_state = State(item, True)
pending.append(item)
visited[item] = new_state
return new_state
return visited[item]
def build_lalr_automaton(G):
def centers(items: [Item]):
return frozenset(item.Center() for item in items)
def lookaheads(items: [Item]):
return {item.Center(): item.lookaheads for item in items}
def subset(items1, items2):
return all(items1[i] <= items2[i] for i in items1)
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0, (G.EOF, ))
start = State(frozenset(closure_lr1([start_item], firsts)), True)
pending = [start_item]
visisted_centers = {centers(start.state): start}
visited = {start_item: start}
while pending:
current_state = visited[pending.pop()]
for symbol in G.terminals + G.nonTerminals:
next_item = frozenset(goto_lr1(current_state.state, symbol,
firsts))
if next_item:
try:
next_state = visisted_centers[centers(next_item)]
if not subset(lookaheads(next_item),
lookaheads(next_state.state)):
next_state.state = compress(
list(next_state.state) + list(next_item))
pending.append(frozenset(next_state.state))
visited[frozenset(next_state.state)] = next_state
except KeyError:
next_state = State(next_item, True)
pending += [next_item]
visisted_centers[centers(next_item)] = next_state
visited[next_item] = next_state
current_state.add_transition(symbol.Name, next_state)
return start
def build_LR0_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
automaton = State(start_item, True)
pending = [start_item]
visited = {start_item: automaton}
while pending:
current_item = pending.pop()
if current_item.IsReduceItem:
continue
# Your code here!!! (Decide which transitions to add)
current_state = visited[current_item]
next_symbol = current_item.NextSymbol
next_item = current_item.NextItem()
try:
next_state = visited[next_item]
except KeyError:
next_state = State(next_item, True)
visited[next_item] = next_state
current_state.add_transition(next_symbol.Name, next_state)
# print('Added trans: ', current_state, f'==={next_symbol.Name}===>', next_state)
pending.append(next_item)
if next_symbol.IsNonTerminal: # If symbol is non terminal, add e-productions to Y->.alpha
for prod in next_symbol.productions:
new_item = Item(prod, 0)
try:
new_state = visited[new_item]
item_was_found = True
except KeyError:
new_state = State(new_item, True)
visited[new_item] = new_state
item_was_found = False
current_state.add_epsilon_transition(new_state)
# print('Added e-trans:', current_state, '==>', new_state)
if not item_was_found:
pending.append(new_item)
return automaton
def _build_regexs(self, table):
regexs = []
for n, (token_type, text_regex) in enumerate(table):
automaton = State.from_nfa(self.regex(text_regex))
for state in automaton:
if state.final: state.tag = (n, token_type)
regexs.append(automaton)
return regexs
def build_LR0_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
automaton = State(start_item, True)
pending = [start_item]
visited = {start_item: automaton}
while pending:
current_item = pending.pop()
if current_item.IsReduceItem:
continue
if current_item.NextSymbol.IsTerminal:
if not current_item.NextItem in visited:
visited[current_item.NextItem] = State(current_item.NextItem,
True)
pending.append(current_item.NextItem)
else:
for production in current_item.NextSymbol.productions:
item = Item(production, 0)
if not item in visited:
visited[item] = State(item, True)
pending.append(item)
if not current_item.NextItem in visited:
visited[current_item.NextItem] = State(current_item.NextItem,
True)
pending.append(current_item.NextItem)
current_state = visited[current_item]
# Your code here!!! (Add the decided transitions)
current_state.add_transition(current_item.NextSymbol.Name,
visited[current_item.NextItem])
if current_item.NextSymbol.IsNonTerminal:
for production in current_item.NextSymbol.productions:
current_state.add_epsilon_transition(visited[Item(
production, 0)])
# automaton.set_formatter(multiline_formatter_lr0)
return automaton
def build_automaton(G: Grammar):
"""
Build the finite automaton for
a regular grammar
"""
states = {
nonTerminal: State(nonTerminal.Name)
for nonTerminal in G.nonTerminals
}
final_state = State('F\'', True)
start_in_right = False
epsilon_production = False
for nonTerminal in G.nonTerminals:
for production in nonTerminal.productions:
right = production.Right
# Start Symbol produces epsilon
if isinstance(right, Epsilon) and nonTerminal == G.startSymbol:
epsilon_production = True
continue
start_in_right |= G.startSymbol in right
n = len(right)
# X --> w
if n == 1 and isinstance(right[0], Terminal):
states[nonTerminal].add_transition(right[0].Name,
final_state)
continue
# X --> w Y
if n == 2 and isinstance(right[0], Terminal) and isinstance(
right[1], NonTerminal):
states[nonTerminal].add_transition(right[0].Name,
states[right[1]])
continue
return states[G.startSymbol], False
states[G.startSymbol].final = epsilon_production
return states[G.startSymbol], not (start_in_right
and epsilon_production)
def build_LR0_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
automaton = State(start_item, True)
pending = [start_item]
visited = {start_item: automaton}
while pending:
current_item = pending.pop()
if current_item.IsReduceItem:
continue
# Your code here!!! (Decide which transitions to add)
new_item = current_item.NextItem()
transition_symbol = current_item.NextSymbol.Name
next_items = [(new_item, transition_symbol)]
if current_item.NextSymbol in G.nonTerminals:
for production in current_item.NextSymbol.productions:
new_item = Item(production, 0)
next_items.append((new_item, G.Epsilon))
current_state = visited[current_item]
# Your code here!!! (Add the decided transitions)
for item, symbol in next_items:
try:
state = visited[item]
except KeyError:
state = State(item, True)
visited[item] = state
pending.append(item)
if symbol == G.Epsilon:
current_state.add_epsilon_transition(state)
else:
current_state.add_transition(symbol, state)
return automaton
def build_LR1_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0, lookaheads=(G.EOF, ))
start = frozenset([start_item])
closure = closure_lr1(start, firsts)
automaton = State(frozenset(closure), True)
pending = [start]
visited = {start: automaton}
while pending:
current = pending.pop()
current_state = visited[current]
for symbol in G.terminals + G.nonTerminals:
# Your code here!!! (Get/Build `next_state`)
dest = goto_lr1(current_state.state, symbol, just_kernel=True)
if not dest:
continue
if dest in visited:
next_state = visited[dest]
else:
items = goto_lr1(current_state.state,
symbol,
firsts,
just_kernel=False)
next_state = State(frozenset(items), True)
visited[dest] = next_state
pending.append(dest)
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton
def Compact_Automata(self, automata):
new_states = {}
for state in automata:
new_states[state] = state
states_to_compress = []
for state1 in automata:
if not new_states[state1] == state1:
continue
states_to_compress = [state1]
for state2 in automata:
if state1 == state2 or not new_states[state2] == state2:
continue
node1 = state1.state
node2 = state2.state
are_equals = False
if len(node1) == len(node2):
for item1 in node1:
are_equals = False
for item2 in node2:
if item1.Center() == item2.Center():
are_equals = True
if not are_equals:
break
if are_equals:
states_to_compress.append(state2)
compress_set = ContainerSet()
for state in states_to_compress:
node = state.state
compress_set.update(ContainerSet(*node))
new_node = self.compress(compress_set)
new_state = State(frozenset(new_node), True)
for state in states_to_compress:
new_states[state] = new_state
new_automata = new_states[automata]
for state in automata:
for key in state.transitions:
for to_state in state.transitions[key]:
try:
assert new_states[to_state] in new_states[state].transitions[key]
except:
new_states[state].add_transition(key, new_states[to_state])
return new_automata
def _build_regexs(table):
regexs = []
for n, (token_type, regex) in enumerate(table):
NFA = Regex.build_automaton(regex)
automaton, states = State.from_nfa(NFA, get_states=True)
for state in automaton:
if state.final:
state.tag = [(n, token_type)]
regexs.append(automaton)
return regexs
def build_LR1_automaton(G):
assert len(G.startSymbol.productions) == 1, "Grammar must be augmented"
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0, lookaheads=(G.EOF, ))
start = frozenset([start_item]) # como cabecera solo queda el kernel
closure = closure_lr1(start, firsts)
automaton = State(frozenset(closure),
True) # en visited si se guarda el estado completo
pending = [start]
visited = {start: automaton}
while pending:
current = pending.pop()
current_state = visited[current]
closure = closure_lr1(current, firsts)
for symbol in G.terminals + G.nonTerminals:
# (Get/Build `next_state`)
# closure = closure_lr1(current,firsts)
goto = goto_lr1(closure, symbol, firsts, True)
if not goto:
continue
try:
next_state = visited[goto]
except KeyError:
next_state = visited[goto] = State(
frozenset(closure_lr1(goto, firsts)), True)
pending.append(goto)
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton
def _build_regexs(self, table):
regexs = []
for n, (token_type, regex) in enumerate(table):
dfa = Regex(regex)
automaton_list = State.from_nfa(
dfa.automaton, 'texto random pra que haga lo que quiero')
for i in automaton_list[1]:
if (i.final):
i.tag = (n, token_type)
regexs.append(automaton_list[0])
return regexs
def build_LR1_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0, lookaheads=(G.EOF, ))
start = frozenset([start_item])
closure = closure_lr1(start, firsts)
automaton = State(frozenset(closure), True)
pending = [start]
visited = {start: automaton}
while pending:
current = pending.pop()
current_state = visited[current]
for symbol in G.terminals + G.nonTerminals:
# Your code here!!! (Get/Build `next_state`)
next_state_key = goto_lr1(current_state.state,
symbol,
just_kernel=True)
# next_state_key = frozenset([i.NextItem() for i in current_state.state if i.NextSymbol == symbol])
if not next_state_key:
continue
try:
next_state = visited[next_state_key]
except KeyError:
next_state_items = goto_lr1(current_state.state, symbol,
firsts)
next_state = State(frozenset(next_state_items), True)
pending.append(next_state_key)
visited[next_state_key] = next_state
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton
def build_LR0_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
automaton = State(start_item, True)
pending = [ start_item ]
visited = { start_item: automaton }
while pending:
current_item = pending.pop()
if current_item.IsReduceItem:
continue
state = visited[current_item]
next_item = current_item.NextItem()
try:
new_state = visited[next_item]
except KeyError:
new_state = visited[next_item] = State(next_item, True)
pending.append(next_item)
next_symbol = current_item.NextSymbol
state.add_transition(next_symbol.Name, new_state)
if next_symbol.IsNonTerminal:
for prod in next_symbol.productions:
item = Item(prod, 0)
try:
visited[item]
except KeyError:
visited[item] = State(item, True)
pending.append(item)
state.add_epsilon_transition(visited[item])
current_state = visited[current_item]
return automaton
def build_LR0_automaton(G):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0)
automaton = State(start_item, True)
pending = [start_item]
visited = {start_item: automaton}
while pending:
current_item = pending.pop()
if current_item.IsReduceItem:
continue
next_item = current_item.NextItem()
symbol = current_item.NextSymbol
try:
visited[current_item].add_transition(symbol.Name,
visited[next_item])
except:
visited[next_item] = State(next_item, True)
visited[current_item].add_transition(symbol.Name,
visited[next_item])
pending.append(next_item)
if symbol.IsNonTerminal:
prod = symbol.productions
for p in prod:
item = Item(p, 0)
try:
visited[current_item].add_epsilon_transition(visited[item])
except KeyError:
visited[item] = State(item, True)
visited[current_item].add_epsilon_transition(visited[item])
pending.append(item)
return automaton
def _build_regexs(self, table):
regexs = {}
for n, (token_type, regex) in enumerate(table):
# Your code here!!!
# - Remember to tag the final states with the token_type and priority.
# - <State>.tag might be useful for that purpose ;-)
auto = regex_automaton(regex)
auto, states = State.from_nfa(auto, True)
for st in states:
if st.final:
st.tag = (n, token_type)
regexs[token_type] = auto
return regexs
def state_transpose(initial_state: State):
"""
returns the state equivalent to initial_state on the transpose graph and a dictionary mapping idx to new states transposed
"""
state_dict = {
} # state_dict[key] = (original_state,copy_of_original_state_with_transposed_transitions)
idx = 0
for x in initial_state:
new_state = State(x.state, x.final, x.formatter, x.shape)
if not hasattr(x, 'idx'):
x.idx = idx
idx += 1
new_state.idx = x.idx
state_dict[x.idx] = (x, new_state)
for x in initial_state:
for key, states in x.transitions.items():
for state in states:
state_dict[state.idx][1].add_transition(
key, state_dict[x.idx][1])
return state_dict[initial_state.idx][1], state_dict
def build_tree(self, current_node, index, parse_list, left_to_right):
new_nodes = []
if parse_list[index[0]].Right:
for x in parse_list[index[0]].Right:
new_nodes.append(State(x))
else:
new_nodes.append(State(parse_list[index[0]].Right))
if not left_to_right: new_nodes.reverse()
for x in new_nodes:
current_node.add_transition('', x)
if x.state.IsNonTerminal:
index[0] += 1
self.build_tree(x, index, parse_list, left_to_right)
if not left_to_right:
try:
current_node.transitions[''].reverse()
except:
pass
def build_lr1_automaton(G, firsts=None):
assert len(G.startSymbol.productions) == 1, 'Grammar must be augmented'
if not firsts:
firsts = compute_firsts(G)
firsts[G.EOF] = ContainerSet(G.EOF)
start_production = G.startSymbol.productions[0]
start_item = Item(start_production, 0, lookaheads=(G.EOF, ))
start = frozenset([start_item])
closure = closure_lr1(start, firsts)
automaton = State(frozenset(closure), True)
pending = [start]
visited = {start: automaton}
while pending:
current = pending.pop()
current_state = visited[current]
current_closure = current_state.state
for symbol in G.terminals + G.nonTerminals:
kernel = goto_lr1(current_closure, symbol, just_kernel=True)
if kernel == frozenset():
continue
try:
next_state = visited[kernel]
except KeyError:
goto = closure_lr1(kernel, firsts)
visited[kernel] = next_state = State(frozenset(goto), True)
pending.append(kernel)
current_state.add_transition(symbol.Name, next_state)
automaton.set_formatter(multiline_formatter)
return automaton