def setup_param_group(self, group):
frame = tk.Frame(self.param_tabs)
self.param_tabs.add(frame, text=group.title)
i = 0
for p1, p2 in pairs(group.params):
self.setup_param(frame, p1, i, 0)
if p2:
self.setup_param(frame, p2, i, 1)
i += 1
def get_sequence(self):
# NEW APPROACH
if len(self.assembly) >= 2:
for contig1,contig2 in util.pairs(self.assembly):
uniq1 = contig1.get_unique_sequence()
uniq2 = contig2.get_unique_sequence()
overlap = contig1.get_overlap_sequence(contig2)
if uniq1 not in self.sequence:
self.sequence.append(uniq1)
self.sequence.append(overlap)
self.sequence.append(uniq2)
# cleanup steps
self.sequence = "".join(self.sequence)
self.length = len(self.sequence)
else:
self.sequence = self.assembly[0].sequence
def recover_token(oracle):
token = oracle.create_token()
ans = bytearray()
for c1, c2 in util.pairs(util.blocks(token, 16)):
p2 = bytearray(16)
for i in reversed(range(16)):
padding = pad(bytes(i), 16)
for guess in range(256):
p2[i] = guess
iv = xor(c1, p2, padding)
if i == 15: # Ugly hack; prevent creating new valid padding.
iv[:14] = bytearray(14)
if oracle.is_padding_ok(bytes(iv + c2)):
break
ans += p2
return bytes(unpad(ans, 16))
def mappings(x, y, getTransformations=True, repeats=False):
tempy = y
if len(y) < len(x):
tempy.extend([None] * (len(x) - len(y)))
mappings = util.pairs(x, tempy)
if len(x) < len(y):
mappings = util.extended_pairs(x, y)
# Mappings defined as follows: Given x=[1], y=[1,2], we want
queue = PriorityQueue()
for mapping in mappings:
ftf = FigureTransformation()
for objectMap in mapping:
name = objectMap[0].getName()
otf = ObjectTransformation(objectMap,
ftf,
transform=getTransformations,
try_repeats=repeats)
ftf.add(name, otf)
yield ftf
def heuristic(ks):
pairs = util.pairs(util.blocks(ct, ks))
return sum(hamming(x, y) for x, y in pairs if len(x) == len(y))
spy = price['SPY'].copy()
price = price.drop('SPY', 1)
delta = pd.DataFrame(index=price.index, columns=price.columns)
for c in price.columns:
delta[c] = price[c].pct_change()
days = delta.shape[0]
symbols = delta.shape[1]
i = 0
corr_tab = np.empty(shape=[days, symbols**2])
for c1 in delta.columns:
for c2 in delta.columns:
corr_tab[:,i] = delta[c1].rolling(window=w).corr(delta[c2])
i += 1
avg_corr = pd.DataFrame(index=delta.index, columns=['Correlation'])
avg_corr['Correlation'] = 100 * (corr_tab.sum(axis=1) - symbols) / 2 / u.pairs(symbols)
avg_corr['SPY'] = spy
avg_corr['Date'] = pd.to_datetime(avg_corr.index)
avg_corr['Smooth SPY'] = avg_corr['SPY'].rolling(window=s).mean()
avg_corr['Smooth Corr'] = avg_corr['Correlation'].rolling(window=s).mean()
avg_corr = avg_corr.iloc[w+s:,:] #truncate the wasted w+s days
avg_corr.loc[avg_corr['Correlation'] <= lo, 'Low'] = True
box = avg_corr.loc[:,['Date', 'Low']]
box = box[box['Low'] == True]
top = avg_corr['SPY'].max()
bottom = avg_corr['SPY'].min()
box['Top'] = top
box['Bottom'] = bottom
box['Right'] = box['Date'] + pd.DateOffset(2)