def add_new_rule(grammar: Grammar):
'''Пополнить грамматику правилом A → aN , где A ∈ V N , a ∈ V T и
N - новый нетерминал, для каждого правила вида A → a , если в грамматике нет
соответствующего ему правила A → aB , где B ∈ V N .'''
rules_with_terms = {
rule
for rule in grammar.rules if rule.right_part != EMPTY_SEQUENCE
and set(rule.right_part) <= grammar.terminals
}
rules_without_other_nonterms = {
rule
for rule in grammar.rules
if len(rule.right_part) == 2 and rule.left_part in rule.right_part
}
rules = rules_with_terms - {
Rule(rule.left_part,
get_term_nonterm(rule, grammar.terminals)[0])
for rule in rules_without_other_nonterms
}
if rules:
new_nonterm = ('N'
if 'N' not in grammar.nonterminals else random.choice(
list(set(ascii_uppercase) - grammar.nonterminals)))
grammar.nonterminals.add(new_nonterm)
new_rules = {
Rule(rule.left_part, rule.right_part + new_nonterm)
for rule in rules
}
grammar.rules.update(new_rules)
def get_rules(sets, acceleration_values, consumption_values):
new_rules = []
# 1. if consumption is low => efficiency high
antecedent = Antecedent(sets['consumption']['low'], consumption_values)
consequence = Consequence(sets['efficiency']['high'])
new_rules.append(Rule(antecedent, consequence))
# 2. if consumption is high => efficiency low
antecedent = Antecedent(sets['consumption']['high'], consumption_values)
consequence = Consequence(sets['efficiency']['low'])
new_rules.append(Rule(antecedent, consequence))
# 3. if acceleration is negative and consumption low => efficiency high
antecedent_acceleration = Antecedent(sets['acceleration']['negative'],
acceleration_values)
antecedent_consumption = Antecedent(sets['consumption']['low'],
consumption_values)
consequence = Consequence(sets['efficiency']['high'])
new_rules.append(
Rule(And(antecedent_acceleration, antecedent_consumption),
consequence))
# 4. if acceleration is positive or consumption high => efficiency low
antecedent_acceleration = Antecedent(sets['acceleration']['positive'],
acceleration_values)
antecedent_consumption = Antecedent(sets['consumption']['high'],
consumption_values)
consequence = Consequence(sets['efficiency']['low'])
new_rules.append(
Rule(Or(antecedent_acceleration, antecedent_consumption), consequence))
return new_rules
def on_packet_in(self, packet, switch, inport):
print("\n>> ", inport, "switch", switch.id, packet)
# the magic oracle function knows what to do with the packet...
use_port = get_port_by_ip(switch, packet.ip_dst)
if use_port >= 0:
r = Rule()
r.MATCH('IP_DST', packet.ip_dst)
r.ACTION('OUTPUT', use_port)
send_rule(r, switch)
# important to not loose any packets!
send_packet(packet, switch, use_port)
def toCNF(self):
"""
each rule must be of format
A → BC (A,B,C ∈ N)
A → a (A ∈ N, a ∈ Σ)
S → ε if S is not on the right side of any rule
"""
import CFG
G = self.toOwn()
N = G.N.copy()
P = CFG.P()
i = 0
while i < len(N):
A = N[i]
for rule in G.P.get(A, P[A]):
A = N[i]
if len(rule) >= 2:
# change rule A -> abcd... to A -> a<bcd...>
B, *C = rule
B1 = Rule(B if B.isupper() else f"{B}̄")
C1 = Rule((f"<{''.join(C)}>" if len(C) > 1 else
C[0] if C[0].isupper() else f"{C[0]}̄"))
P[A].add(Rule(f"{B1}{C1}"))
A = Rule(f"{C1[0]}")
# continue generating <bcd...> -> b<cd...>
while len(A[0]) >= 2 and A[0] != "<>":
rule = Rule(f"{C1[0][1]}<{C1[0][2:-1]}>")
B, *C = rule
B1 = Rule(B if B.isupper() else f"{B}̄")
C1 = Rule((f"{''.join(C[0])}" if len(C[0]) > 1 else
C[0] if C[0].isupper() else f"{C[0]}̄"))
P[A].add(Rule(f"{B1}{C1}"))
A = Rule(f"{C1[0]}")
else:
P[A].add(Rule(rule))
i += 1
return CFG(N, self.Σ, P, self.S)
def resolve(A, B):
for _ in range(len(P[A])):
rule = list(P[A].pop(0))
if rule[0].islower():
for i, c in enumerate(rule[1:]):
if c.islower() and len(c) == 1:
rule[i + 1] = f"{c}̄"
rule = "".join(rule)
P[A].add(Rule(rule))
elif rule[0] == B:
rules = Set()
for rule1 in P[B]:
rules.add(Rule(rule1 + "".join(rule[1:])))
P[A] |= rules
else:
P[A].add(Rule(rule))
def remove_left_recursion(self):
def calc_potentials():
potentials = {}
for A in N:
potentials.setdefault(A, Set())
for rule in P[A]:
if rule[0] in N:
potentials[A].add(rule[0])
return potentials
N = self.N.copy()
P = self.P.copy()
for i, A in enumerate(N):
for B in N[:i + 1]:
if not B in calc_potentials()[A]:
continue
if A == B:
α = [rule for rule in P[A] if rule.startswith(B)]
β = [rule for rule in P[A] if not rule.startswith(B)]
N = Set(f"{A}'").union(N)
P[f"{A}'"] |= Set(Rule(rule[1:]) for rule in α) | Set(
Rule(rule[1:] + f"{A}'") for rule in α)
P[A] = Set(Rule(rule) for rule in β) | Set(
Rule(rule + f"{A}'") for rule in β)
else:
α = [rule for rule in P[A] if rule.startswith(B)]
β = [rule for rule in P[A] if not rule.startswith(B)]
P[A] = Set(Rule(rule) for rule in β) | Set(
Rule(ruleB + rule[1:]) for rule in α for ruleB in P[B])
return CFG(N, self.Σ.copy(), P, self.S)
def Earley(self):
for i in range(0, len(self.word) + 1):
self.Scan(i)
old_length = len(self.D[i])
self.Predict(i)
self.Complete(i)
new_length = len(self.D[i])
while new_length != old_length:
old_length = new_length
self.Predict(i)
self.Complete(i)
new_length = len(self.D[i])
if Situation(Rule('S1', 'S'), 0, 1) in self.D[len(self.word)]:
return 'YES'
else:
return 'NO'
def make_finite_state_machine(grammar: Grammar):
'''Преобразовать регулярную грамматику в конечный автомат'''
add_new_rule(grammar)
# Начальный символ грамматики S принять за начальное состояние КА H.
start_state = grammar.start_symbol
# Из нетерминалов образовать множество состояний автомата Q = V N ∪ {N },
states = grammar.nonterminals
# а из терминалов – множество символов входного алфавита T = V T .
input_symbols = grammar.terminals
transitions = transform_rules_to_transitions(grammar)
# Если в грамматике имеется правило S → ε , где S - начальный символ грамматики,
# то поместить S во множество заключительных состояний.
final_states = {grammar.start_symbol} if Rule(
grammar.start_symbol, EMPTY_SEQUENCE) in grammar.rules else set()
fsm = FSM(states, input_symbols, transitions, start_state, final_states)
# Если получен НКА, то преобразовать его в ДКА
if not is_dfsm(fsm):
fsm = transform_nfsm_to_dfsm(fsm)
return fsm
def __init__(self, grammar, word):
self.word = word
self.grammar = grammar
self.D = [set() for i in range(len(self.word) + 1)]
self.D[0].add(Situation(Rule('S1', 'S'), 0, 0))
def on_connect(self, switch):
print "switch connected", switch.id
if switch.id == 1:
r = Rule()
r.MATCH('IP_DST', '20.0.0.0/16')
r.ACTION('OUTPUT', 5)
send_rule(r, switch)
if switch.id == 2:
r = Rule()
r.MATCH('IP_DST', '20.0.1.0/24')
r.ACTION('OUTPUT', 2)
send_rule(r, switch)
r = Rule()
r.MATCH('IP_DST', '20.0.2.0/24')
r.ACTION('OUTPUT', 3)
send_rule(r, switch)
r = Rule()
r.MATCH('IP_DST', '20.0.3.0/24')
r.ACTION('OUTPUT', 4)
send_rule(r, switch)
r = Rule()
r.MATCH('IP_DST', '20.0.4.0/24')
r.ACTION('OUTPUT', 5)
send_rule(r, switch)
def get_mongo(mongo_address="mongodb://localhost:27017/", mongo_db="python", mongo_collection="tokens", logfile=None):
try:
client = pymongo.MongoClient(mongo_address)
db = client[mongo_db]
collection = db[mongo_collection]
current = []
if mongo_collection == "tokens":
for x in collection.find():
left = None
right = None
if 'leftborder' in x and x['leftborder']:
if 'allow_symbols' in x['leftborder']:
left = Border(x['leftborder']['allow_symbols'])
if 'their_dict' in x['leftborder']:
left.their_dict = x['leftborder']['their_dict']
if 'rightborder' in x and x['rightborder']:
if 'allow_symbols' in x['rightborder']:
right = Border(x['rightborder']['allow_symbols'])
if 'their_dict' in x['rightborder']:
right.their_dict = x['rightborder']['their_dict']
current.append(Token(x['prior'], x['reg'], x['name'], left, right))
elif mongo_collection == "rules":
for x in collection.find():
role = []
ignore_symbols = {}
visa_verse = []
visa_verse_tokens = []
use_parent_final_words = False
inner = False
could_be_end = None
check_could_be_end = None
if 'role' in x:
role = x['role']
if 'ignore_symbols' in x:
ignore_symbols = x['ignore_symbols']
if 'inner' in x:
inner = x['inner']
if 'visa_verse' in x:
visa_verse = x['visa_verse']
if 'visa_verse_tokens' in x:
visa_verse_tokens = x['visa_verse_tokens']
if 'use_parent_final_words' in x:
use_parent_final_words = x['use_parent_final_words']
if 'check_could_be_end' in x:
check_could_be_end = x['check_could_be_end']
if 'could_be_end' in x:
could_be_end = x['could_be_end']
current.append(
Rule(x['keyword'], x['allow_tokens'], x['final_words'], role, ignore_symbols, inner, visa_verse,
visa_verse_tokens, use_parent_final_words, could_be_end, check_could_be_end))
elif mongo_collection == "bugs":
for x in collection.find():
id = 0
object = ""
line = 0
column = 0
description = ""
priority = 1
important = 1
type = ""
date = ""
author = ""
program = ""
version = "1.0.0"
state = "Open"
manage_by = ""
possible_decision = ""
if 'id' in x:
id = x['id']
if 'object' in x:
object = x['object']
if 'line' in x:
line = x['line']
if 'column' in x:
column = x['column']
if 'description' in x:
description = x['description']
if 'priority' in x:
priority = x['priority']
if 'important' in x:
important = x['important']
if 'type' in x:
type = x['type']
if 'date' in x:
date = x['date']
if 'author' in x:
author = x['author']
if 'program' in x:
program = x['program']
if 'version' in x:
version = x['version']
if 'state' in x:
state = x['state']
if 'manage_by' in x:
manage_by = x['manage_by']
if 'possible_decision' in x:
possible_decision = x['possible_decision']
current.append(
Bug(id, object, line, column, description, priority, important, type, date, author, program,
version, state, manage_by, possible_decision))
except Exception as ex:
print('The error has occured')
print(ex)
finally:
if client:
client.close()
if current:
return current
def test():
print(Rule("SaSaS").chars)
print(Rule("aX'b").chars)
print(Rule("ab̄c").chars)
print(Rule("ab̂c").chars)
print(Rule("a<bc<d>f>e").chars)
def scores(s, best_rules, f=None):
#scores = {}
bestscore = 0
bestrule = []
a = 0
top_rules = {}
for atag in s.keys():
if len(atag) > 4:
stat = s[atag]
vtags = atag.split()
if len(vtags) < 2:
continue
for y in vtags:
try:
freq = s[y]
except KeyError:
s[y] = ({0: {}}, {0: {}}, {0: {}}, 0)
continue
for i, cont_type in enumerate(s[y][:3]):
for distance in cont_type:
for context in cont_type[distance]:
fr = -sys.maxint
for z in vtags:
if y != z:
# relative frequency
try:
incontext_z = s[z][i][distance][context]
except:
incontext_z = 0.0
try:
freq_z = s[z][3]
relf = float(s[y][3]) / float(freq_z) * float(incontext_z)
except:
freq_z = 0.0
relf = 0.0
if relf >= fr:
fr = relf
try:
w = incontext_z
except:
w = 0
#print vtags
x = s[y][i][distance][context] - float(fr)
curr_rule = Rule(*[atag, y, (distance, context), i])
curr_rule.id = feature_type(context)
if x > bestscore and s[y][i][distance][context] != w:
try:
a = stat[i][distance][context]
except KeyError:
continue
bestscore = x
bestrule = curr_rule
top_rules = {}
top_rules[bestrule] = a
elif x == bestscore and s[y][i][distance][context] != w:
try:
a = stat[i][distance][context]
except KeyError:
continue
bestrule = curr_rule
top_rules[bestrule] = a
return top_rules, bestscore, a
try:
a = stat[i][distance][context]
except KeyError:
continue
bestrule = curr_rule
top_rules[bestrule] = a
return top_rules, bestscore, a
def random_choice(corpus):
for s in corpus:
for token in s:
try:
if token.has_ambig():
token.disambiguate(random.choice(token.getPOStags().split('_')))
except:
pass
if __name__ == '__main__':
inc = read_corpus(sys.stdin)
#s = context_stats(inc, join_context=True)
#print s
#rs = scores(s, [])
#pprint(rs[1:3])
'''for x in rs[1].keys():
print x.display()
print rs[1][x]
print rs[2]'''
r = Rule('ADVB NOUN', 'NOUN', (1, 'PNCT'), 0)
from utils import Rule
from algorithm import Earley
if __name__ == '__main__':
n = int(input())
grammar = []
for i in range(n):
s = input().split('->')
grammar.append(Rule(s[0],s[1]))
word = input()
algo = Earley(grammar, word)
print(algo.Earley())