def spoofing(self, hostname, dns, rawdata, sock): fromip = self.client_address[0] for ipp,rev in self.server.resolvs: if match(ipp, fromip): for hnp,ip in rev: if match(hnp, hostname): return ip return self.queryip(hostname) return self.queryip(hostname)
def test_process():
from util import read_file_to_list
from util import match
rules, rows = read_file_to_list("testinput.txt")
assert 6 == len(rules)
assert 5 == len(rows)
assert True == match(rules, rows[0])
assert False == match(rules, rows[1])
assert True == match(rules, rows[2])
assert False == match(rules, rows[3])
assert False == match(rules, rows[4])
def check(self, imgSrc) -> bool:
self.imgSrc = ac.imread(imgSrc)
res = util.match(self.imgSrc, self.imgSign, 0.9)
if res != None :
self.y0 = int(res['rectangle'][3][1] + 80*paras.SCALE)
return True
return False
def match(self, text): """Personal match.""" if text in ['i', 'me']: return self elif text == 'here': return self.location else: return util.match(text, set(self.contents + self.location.contents + [self.location, self]))
def check(self, imgSrc) -> bool:
self.imgSrc = ac.imread(imgSrc)
res = util.match(self.imgSrc, self.imgSign)
# print(res)
if res != None :
self.point = [int(res['result'][0]), int(res['result'][1])]
return True
return False
def test_process():
from util import read_file_to_list
from util import match
rules, rows = read_file_to_list("testinput2.txt")
valid = 0
for row in rows:
if match(rules, row):
valid += 1
assert valid == len(rows)
def test(task_name, func, subset='train'):
with open(f'ARC/data/training/{task_name}.json') as f:
j = json.load(f)
correct = True
for i, t in enumerate(j[subset]):
input_grid = np.array(t['input'])
pred = func(input_grid)
# vis(pred)
target = np.array(t['output'])
correct &= match(pred, target)
return correct
def processa_paises():
who_doctors, who_nurses = util.read_pais()
# remove duplicatas, deixando apenas os dados mais recentes
util.drop_pais_dupl(who_doctors, "Country")
util.drop_pais_dupl(who_nurses, "Country")
# cada bloco a seguir cria um atributo da tabela final
# a lista de paises nao eh igual entre os datasets
# para igualar, concatenamos e fazemos os ajustes pertinentes
nome = []
nome_aux = pd.concat([who_doctors["Country"], who_nurses["Country"]], join='inner', ignore_index=True)
nome_aux.drop_duplicates(inplace=True)
nome_series = nome_aux.reset_index() # reseta indices
nome_series.drop(["index"], axis=1, inplace=True) # remove indices antigos
for i in range(len(nome_series)):
nome.append(nome_series["Country"][i])
nome.sort(key=str.lower) # organiza em ordem alfabetica
med_total = util.match(nome, who_doctors, "Medical doctors (number)", "Country")
med_10k = util.match(nome, who_doctors, "Medical doctors (per 10 000 population)", "Country")
enf_total = util.match(nome, who_nurses, "Nursing and midwifery personnel (number)", "Country")
enf_10k = util.match(nome, who_nurses, "Nursing and midwifery personnel (per 10 000 population)", "Country")
dados = {'País': nome,
'Profissionais de enfermagem - Total': enf_total,
'Profissionais de enfermagem a cada 10k habitantes': enf_10k,
'Médicos - Total': med_total,
'Médicos a cada 10k habitantes': med_10k}
df = pd.DataFrame(data=dados)
util.write_to_csv(df, "Mundo (OMS)")
stats.stats_paises(pd.read_csv("../data/processed/Mundo (OMS).csv"))
def test_match():
s = '\n'
s += str(util.match('abc', 'a'))
s += '\n'
s += str(util.match('abc123', 'bc'))
s += '\n'
s += str(util.match('abc123', '123'))
s += '\n'
s += str(util.match('abc 123', '\s'))
s += '\n'
s += str(util.match('abc 123xyz', '\d'))
s += '\n'
s += str(util.match('abc123', '^abc'))
s += '\n'
s += str(util.match('abc123', '^123'))
s += '\n'
s += str(util.match('abc', 'z'))
return s
def test(task_name, func_string, subset='train'):
with open(f'ARC/data/training/{task_name}.json') as f:
j = json.load(f)
prog = parse(func_string)
correct = True
for i, t in enumerate(j[subset]):
input_grid = np.array(t['input'])
prog_with_input = ['define', 'grid', input_grid, prog]
pred = eval(prog_with_input)
#print(pred)
# vis(pred)
target = np.array(t['output'])
correct &= match(pred, target)
return correct
def answer(self):
if self.problem is None:
return
answer = self.store.find(self.problem)
if answer is None:
print("answer not found")
self.tapAnswer(0)
self.captureAnswer()
else:
res = util.match(self.imgSrc, answer)
if res is None:
print("answer not match")
self.tapAnswer(0)
self.captureAnswer()
return
point = int(res['result'][0]), int(res['result'][1])
#idx = util.findImg(self.answers, answer)
#print("=====Find Answer!=====", idx)
#self.tapAnswer(idx)
print("======find answer======", point[0], point[1])
self.tapPosition(point[0], point[1])
time.sleep(1) #skip the middle status: right answer but wrong sign
def matchfind():
surveyname = request.args.get('surveyname', '') #write JS method for this
username = session['user']
print surveyname, username
return json.dumps(util.match(surveyname, username))
noise = Variable(noise, volatile=True) # total freeze netG
y = Variable(netG(noise).data)
f_enc_Y_D, f_dec_Y_D = netD(y)
# compute biased MMD2 and use ReLU to prevent negative value
mmd2_D = mix_rbf_mmd2(f_enc_X_D, f_enc_Y_D, sigma_list)
mmd2_D = F.relu(mmd2_D)
# compute rank hinge loss
#print('f_enc_X_D:', f_enc_X_D.size())
#print('f_enc_Y_D:', f_enc_Y_D.size())
one_side_errD = one_sided(f_enc_X_D.mean(0) - f_enc_Y_D.mean(0))
# compute L2-loss of AE
L2_AE_X_D = util.match(x.view(batch_size, -1), f_dec_X_D, 'L2')
L2_AE_Y_D = util.match(y.view(batch_size, -1), f_dec_Y_D, 'L2')
errD = torch.sqrt(
mmd2_D
) + lambda_rg * one_side_errD - lambda_AE_X * L2_AE_X_D - lambda_AE_Y * L2_AE_Y_D
errD.backward(mone)
optimizerD.step()
# ---------------------------
# Optimize over NetG
# ---------------------------
for p in netD.parameters():
p.requires_grad = False
for j in range(Giters):
def matchfind():
surveyname = request.args.get("surveyname", "") # write JS method for this
username = session["user"]
print surveyname, username
return json.dumps(util.match(surveyname, username))
def button4Click(self):
self.error = match(self.cfeats,self.rfm)
self.button4["background"] = "lightgreen"
def match_headers(self,headers,hijackheaders): for k,v in hijackheaders.items(): if not k in headers or not match(v,headers[k]): return False return True
# Lam [[Char]] expr
# Word [Char]
# Number Num
# Character Char
App = lambda es: lambda app,lam,word,number,character:app(es)
Lam = lambda w,e: lambda app,lam,word,number,character:lam(w,e)
Word = lambda n: lambda app,lam,word,number,character:word(n)
Number = lambda d: lambda app,lam,word,number,character:number(d)
Character = lambda c: lambda app,lam,word,number,character:character(c)
getChildren = lambda n:n(Const([]),lambda _,la: la)
getName = lambda l: l(lambda t:t.d,lambda _,__:'')
toExprST = Y(lambda f: lambda tr:tr(
lambda t:match(t.n)([
('T_word',lambda w: Word(t.d)),
('T_number',lambda d: Number(int(t.d))),
('T_char',lambda c: Character(t.d))
])(lambda a:Const(a)(print(a))),
lambda b,la: match(b)([
('P_expr',lambda _: let(list(map(f,la)),lambda la:
# fold(App)(la[0])(la[1:])
App(la)
)),
('P_abst',lambda _: let(list(map(getName,getChildren(la[0]))),lambda largs:
Lam(largs,f(la[1]))
))
])(lambda a:Const(a)(print(a)))
))
# toListOfExprs as RoseTree Tok [Char] -> Maybe [([Char],expr)]
toListOfExprs = lambda tr: (
one_side_errD_thinned = one_sided(
f_enc_X_D.mean(0) - thinned_f_enc_Y_D.mean(0))
except Exception as e:
print('D: Thinning f_enc_Y: Error: {}'.format(e))
pdb.set_trace()
# Unthinned hinge loss.
one_side_errD_unthinned = one_sided(
f_enc_X_D.mean(0) - f_enc_Y_D.mean(0))
# Choose which hinge loss you want.
if not weighted:
one_side_errD = one_side_errD_unthinned
else:
one_side_errD = one_side_errD_thinned
# compute L2-loss of AE
L2_AE_X_D = util.match(x.view(batch_size, -1), f_dec_X_D, 'L2')
L2_AE_Y_D = util.match(y.view(batch_size, -1), f_dec_Y_D, 'L2')
# Also compute AE loss on subsets of zeros and ones.
try:
if len(x_eval_1s) or len(x_eval_dec_1s):
L2_AE_X1_D = util.match(x_eval_1s.view(len(x_eval_1s), -1),
x_eval_dec_1s, 'L2')
else:
L2_AE_X1_D = Variable(
torch.from_numpy(np.array([0
])).type(torch.FloatTensor))
except Exception as e:
print('D: Computing x1 ae error. Error: {}'.format(e))
pdb.set_trace()
try:
if len(x_eval_0s) or len(x_eval_dec_0s):