def check_hessenberg_tie_breaking(g, n_rows, n_cols, values, result, compare):
if result.get('error_msg'):
check(result, n_rows, n_cols, values)
return
#
rowp, colp = result['rowp'], result['colp']
rperm = [-1] * len(rowp)
for i, r in enumerate(rowp):
rperm[r] = i
# Last occupied columns rowwise, empty rows allowed
n_rows = len(rperm)
last_elem = [
max(colp[c - n_rows] for c in g[r]) if g[r] else -1 for r in rperm
]
log('Last elems:', last_elem)
row_weights = get_row_weights(g, n_rows)
perm_weights = [row_weights[r] for r in rperm]
# Checking if we really break ties if the last elements equal
for i, ((c1, c2), (w1, w2)) in enumerate(
zip(pairwise(last_elem), pairwise(perm_weights))):
log('i:', i)
log('Last elems:', c1, c2)
log('Weights:', w1, w2)
#assert (c1, w1) <= (c2, w2)
assert compare(c1, w1, c2, w2)
def compute_name_value_category(self):
has_single_color = True
for c1, c2 in pairwise(self.cards):
if c1.color != c2.color:
has_single_color = False
is_suite = True
for c1, c2 in pairwise(self.cards):
if c1.number < c2.number - 1 or \
c1.number == c2.number:
is_suite = False
is_set = True
for c1, c2 in pairwise(self.cards):
if c1.number != c2.number:
is_set = False
if has_single_color:
# Color suite. Strongest combination.
# The smallest card number gives its strength
# relative to other color suites
if is_suite:
self.name = "Color suite"
self.category = 4
self.value = self.cards[0].number
# Color.
# The sum of the card numbers gives its
# strength relative to other colors
else:
self.name = "Color"
self.category = 2
self.value = self.cards[0].number + \
self.cards[1].number + \
self.cards[2].number
elif is_suite:
# Suite. Strongest combination.
# The smallest card number gives its strength
# relative to other suites
if is_suite:
self.name = "Suite"
self.category = 1
self.value = self.cards[0].number
elif is_set:
# Set. Second strongest combination.
# Any card number gives its strength
# relative to other sets
self.name = "Set"
self.category = 3
self.value = self.cards[0].number
else:
# Sum: weakest combination.
# The sum of the card numbers gives its
# strength relative to other sums.
self.name = "Sum"
self.category = 0
self.value = self.cards[0].number + \
self.cards[1].number + \
self.cards[2].number
def L_admissible(Gamma):
# L-admissibility
iotaGamma = [(-1)**i * g for i, g in enumerate(Gamma)]
la = ((all(g > 0 for g in iotaGamma) or all(g < 0 for g in iotaGamma))
and (all(f >= g for f, g in pairwise(iotaGamma))
or all(f <= g for f, g in pairwise(iotaGamma))))
if not la:
print("Vorticities {} are not L-admissible.".format(Gamma))
return la
def check_nondecreasing_row_weights(bip, rowp, colp, row_weights):
# Last occupied columns rowwise, empty rows allowed
c_index = { name : i for i, name in enumerate(n for n in colp) }
last_elem = [max(c_index[c] for c in bip[r]) if bip[r] else -1 for r in rowp]
log('Last elems:', last_elem)
perm_weights = [row_weights[r] for r in rowp]
# Checking if we really break ties if the last elements equal
for i, ((c1,c2), (w1,w2)) in enumerate(zip(pairwise(last_elem), pairwise(perm_weights))):
log('i:', i)
log('Last elems:', c1, c2)
log('Weights:', w1, w2)
assert (c1, w1) <= (c2, w2)
def __main__():
with open(sys.argv[1]) as f:
cfg = parse_dot(f)
all_nodes = sorted(cfg.nodes(), key=natural_sort_key)
for n, next_n in pairwise(all_nodes):
print("%s:" % n)
print(" nop()")
succ = cfg.succ(n)
if len(succ) == 0:
print(" return")
elif len(succ) == 1:
if succ[0] != next_n:
print(" goto %s" % succ[0])
else:
old_len = len(succ)
succ = [x for x in succ if x != next_n]
else_node = None
if len(succ) == old_len:
# next_n wasn't in succ
else_node = succ[-1]
succ = succ[:-1]
s = ", ".join(["($cond) goto %s" % x for x in succ])
if else_node:
s += ", else goto %s" % else_node
print(" if " + s)
def dump_c(cfg, stream=sys.stdout):
labels = find_used_labels(cfg)
func_start = True
for (addr, info), nxt in pairwise(cfg.iter_rev_postorder()):
bblock = info["val"]
if func_start:
label = cfg.parser.label_from_addr(bblock.addr)
if label[0].isdigit():
label = "fun_" + label
if ("estimated_params" in cfg.props):
print("// Estimated params: %s" %
sorted(list(cfg.props["estimated_params"])),
file=stream)
if cfg.props["trailing_jumps"]:
print(
"// Trailing jumps not removed, not rendering CFG edges as jumps",
file=stream)
print("void %s()\n{" % label, file=stream)
func_start = False
if addr in labels:
print("\nl%s:" % addr, file=stream)
bblock.dump(stream, indent=1, printer=print_inst)
if not cfg.props["trailing_jumps"]:
for succ in cfg.succ(addr):
cond = cfg.edge(addr, succ).get("cond")
if not cond and nxt and succ == nxt[0]:
continue
stream.write(" ")
if cond:
stream.write("if %s " % cond)
print("goto l%s;" % succ, file=stream)
print("}", file=stream)
def hierarchyReduction(hierarchy, dim = 2, classic=False):
if(hierarchy.representatives == None):
return
docs = map(lambda x: x.center, hierarchy.representatives)
if (hierarchy.center != None):
docs.append(hierarchy.center)
if (hierarchy.simMat == None):
hierarchy.simMat = utils.pairwise(docs, lambda x,y: x.similarity(y))
points = []
if(classic):
points = classicMDS(hierarchy.simMat,dim)
else:
points = reduction(hierarchy.simMat,dim)
if (hierarchy.center != None):
hierarchy.centerPos = points[-1]
hierarchy.mdsPos = points[:-1]
else:
hierarchy.mdsPos = points[:]
map(lambda x: hierarchyReduction(x,dim), hierarchy.representatives)
def reduce(self, classic=False, dim=2):
if (self.representatives == None):
return
docs = map(lambda x: x.center, self.representatives)
if (self.center != None):
docs.append(self.center)
if (self.simMat == None):
self.simMat = utils.pairwise(docs, lambda x, y: x.similarity(y))
points = []
if (classic):
points = mds.classicMDS(self.simMat, dim)
else:
points = mds.reduction(self.simMat, dim)
if (self.center != None):
self.centerPos = points[-1]
self.mdsPos = points[:-1]
else:
self.mdsPos = points[:]
map(lambda x: x.reduce(classic, dim), self.representatives)
def main(args):
if(len(args) != 1):
print "Usage: mds.py C clustering.pkl"
print " C is the cluster in clustering.pkl to display"
sys.exit(0)
#C = int(args[1])
#path = args[2]
print "Loading"
#clustering = utils.load_obj(path)
#docs = clustering[C].members
docs = doc.get_docs_nested(driver.get_data_dir("small"))
print "Calculating Pairwise Similarities"
similarities = utils.pairwise(docs, lambda x,y: x.similarity(y))
#print "INITIAL SIMILARITIES:"
#utils.print_mat(similarities)
#similarities = [[0,93,82,133],[93,0,52,60],[82,52,0,111],[133,60,111,0]]
print "Starting MDS"
#pos = reduction(similarities)
pos = classicMDS(similarities)
print "MDS:"
utils.print_mat(pos)
def scrape(origin, destination, start_date, end_date):
# Creates a file with the cheapest outgoing and return flights for a given
# date range.
items = int((end_date - start_date).days) - 1
date_list = daterange(start_date, end_date)
data = {}
for index, (depart_date, return_date) in enumerate(pairwise(date_list)):
try:
depart_list, return_list = search_flight(destination, depart_date,
return_date, params)
except:
log.error(f"Some error occured, retrying item {index+1}/{items}")
data = open_file(f'{destination}.json')
data[str(depart_date)]['outgoing'] = min_depart
data[str(return_date)]['return'] = min_return
with open(f'{destination}.json', 'w') as f:
json.dump(data,
f,
sort_keys=True,
indent=8,
separators=(',', ': '))
log.info(f"Search {index+1}/{items} complete: {depart_date} to "
f"{return_date}. ({min_depart}MYR, {min_return}MYR)")
def get_updated_link_shape(link, probes, get_distance=get_distance):
[ref_node_shape, *_] = link.shapeInfo.split('|')
[ref_node_lat, ref_node_lon, slope] = ref_node_shape.split('/')
ref_latlon = (float(ref_node_lat), float(ref_node_lon))
get_distance_to_ref_node = lambda p: get_distance(
(p.matchedLatitude, p.matchedLongitude), ref_latlon)
unique_probes = dedup(probes, get_distance_to_ref_node)
sorted_by_distance = sorted(unique_probes, key=get_distance_to_ref_node)
slopes = [
compute_slope(p1, p2) for (p1, p2) in pairwise(sorted_by_distance)
]
slopes = [s for s in slopes if s]
average_slope = mean(slopes) if len(slopes) > 0 else 0.0
result = []
for node in link.shapeInfo.split('|'):
[lat, lon, slope] = node.split('/')
result += [
'{}/{}/{}'.format(lat, lon, slope if slope else average_slope)
]
return '|'.join(result)
def max_acceleration_pass(track: Track, car: Car):
"""
Calculates the maximum acceleration (and corrects maximum velocity if necessary) in
one backwards pass.
"""
def calc(this: Point, nxt: Point):
# Find maximum acceleration.
this.max_acceleration = (nxt.max_velocity**2 -
this.max_velocity**2) / 2
if this.max_acceleration > car.acceleration:
# Can't speed up enough, so correct next max velocity.
this.max_acceleration = car.acceleration
nxt.max_velocity = calc_velocity(this.max_velocity,
this.max_acceleration)
if nxt.next:
calc(this.next, nxt.next)
if this.max_acceleration < -1 * car.braking:
# Can't slow down enough, so correct previous max velocity.
this.max_acceleration = -1 * car.braking
this.max_velocity = calc_velocity(nxt.max_velocity, car.braking)
nxt.max_velocity = calc_velocity(this.max_velocity,
this.max_acceleration)
track.points[-1].max_acceleration = 0
for second, first in pairwise(track.points[::-1]):
calc(first, second)
def plot_one(data_x, data_y, filename=None, show=False):
print("Producing information plane image")
cmap = plt.get_cmap('gnuplot')
for id, e in enumerate(pairwise(zip(data_x, data_y))):
(x1, y1), (x2, y2) = e
plt.plot((x1, x2), (y1, y2),
linewidth=0.2,
alpha=0.9,
color=cmap(min(.9, 0.15 + id * 0.1)))
point_size = 300
plt.scatter(x1,
y1,
s=point_size,
alpha=0.9,
color=cmap(min(.9, 0.15 + id * 0.1)))
plt.scatter(x2,
y2,
s=point_size,
alpha=0.9,
color=cmap(min(.9, 0.15 + id * 0.1)))
plt.xlabel('I(X,T)')
plt.ylabel('I(Y,T)')
if filename is not None:
print("Saving image to file : ", filename)
start = time.time()
plt.savefig(filename, dpi=1000)
end = time.time()
print("Time taken to save to file {:.3f}s".format((end - start)))
if show:
plt.show()
plt.cla()
def parse_contributions(text):
def is_invalid(s):
tests = [
s['first_name'][0].isupper(), s['last_name'][0].isupper(),
len(s['first_name']) >= 2,
len(s['last_name']) >= 2
]
if not all(tests):
log.debug("Discarting invalid contribution {}".format(s))
# itertools.filterfalse() returns list of items where the predicate returns
# false, hence the inverse logic
return not all(tests)
end = re.search(r'$', text)
speakers = itertools.filterfalse(lambda x: is_invalid(x.groupdict()),
Regex.speaker_reg_.finditer(text))
# TODO: remove/fix
aw_data = ''
with open('../data/deputies.json') as f:
aw_data = f.read()
aw_data = json.loads(aw_data)
contributions = []
for m, m1 in pairwise(itertools.chain(speakers, [end])):
contrib = {
'speaker': match_abgeordnetenwatch(m.groupdict(), aw_data),
'start_idx': m.start(),
'end_idx': m1.start(),
'speech': text[m.end():m1.start()]
}
contributions.append(contrib)
return contributions
def __compute_ride_fares(self, ride: Ride):
slot_fares = self.rides_config.slot_fares
remaining_duration = ride.duration
ride_fares = []
start_time = ride.startTime.time()
for current_slot_fare, next_slot_fare in cycle(pairwise(slot_fares + [slot_fares[0]])):
if start_time < next_slot_fare.start or next_slot_fare.start == time(0, 0, 0):
slot_duration = RidesService.__compute_slot_duration(current_slot_fare, next_slot_fare)
ride_end_time = (ride.startTime + ride.duration).time()
is_ride_end_in_slot = current_slot_fare.start <= ride_end_time and remaining_duration <= slot_duration
if is_ride_end_in_slot:
ride_fares.append((remaining_duration, current_slot_fare.fare))
remaining_duration = timedelta(0)
else:
slot_fare_end = datetime.combine(date.min, current_slot_fare.start) + slot_duration
ride_in_slot_duration = slot_fare_end - datetime.combine(date.min, start_time)
remaining_duration = remaining_duration - ride_in_slot_duration
ride_fares.append((ride_in_slot_duration, current_slot_fare.fare))
start_time = slot_fare_end.time()
if remaining_duration <= timedelta(0):
break
return ride_fares
def break_up_audio_file_by_timestamps(audio_file, timestamps):
audio_file_name, audio_file_ext = os.path.splitext(audio_file)
audio_segment = AudioSegment.from_file(audio_file)
intervals = pairwise(itertools.chain((0,), timestamps, (len(audio_segment),)))
slices = cut_audio_segment(audio_segment, intervals)
for idx, current in enumerate(slices, 1):
current.export(f'{audio_file_name}_{idx}{audio_file_ext}', format=audio_file_ext[1:])
def get_cost_score(self) -> float:
# demand ~ capacity
# time ~ due_time
self.initial_port()
for source, dest in utils.pairwise([0] + self.route + [0]):
if self.check_capacity(dest):
# current vehicle has the capacity to go from source to dest
if not self.check_time_and_go(source, dest):
# current vehicle hasn't enough time to go to dest -> new vehicle
# current vehicle should go back from source to deport
self.move_vehicle(source, 0)
# current_load = 0
# new vehicle starts from deport heading dest
self.add_vehicle()
self.move_vehicle(0, dest)
else:
# current vehicle hasn't the capacity to go to dest
# current vehicle should go back from source to deport
self.move_vehicle(source, 0)
# head from deport to dest
distance = self.get_distance(
0, dest) # just for speeding up (caching)
if not self.check_time_and_go(0, dest, distance):
# too late to go from deport to dest on current vehicle -> new vehicle
self.add_vehicle()
self.move_vehicle(0, dest, distance)
total_travel_cost = Chromosome.get_travel_cost(self.total_travel_dist)
total_vehicles_and_deport_working_hours_cost = self.get_vehicle_count_preference_cost(
vehicles_count=self.vehicles_count,
deport_working_hours=self.max_elapsed_time)
return total_travel_cost + total_vehicles_and_deport_working_hours_cost
def translated_operation(a, rng, operator):
"""
Return an array of corresponding results from operations between
diametrically opposed translated slices.
Parameters
----------
a : array-like
Input array.
rng : int
Half of the Chebyshev distance between the two inputs
in each pairs.
operator : function
Binary operator used to compute results
Return
------
out : MaskedArray
Array of results where additional dimension correspondent
to each pair of translated slice.
"""
# get centered pixels for 2x2 and 3x3 based on rng
a__ = a[:,1:-1,1:-1] if rng != 0 \
else (a[:,:-1,:-1]+a[:,:-1,1:]+a[:,1:,1:]+a[:,1:,:-1]) /4
# compute angle difference
out = ma.masked_array(
[operator(a__, a[ts]) for ts in TRANSLATING_SLICES_[rng]])
# Rearrange axes:
for dim1, dim2 in pairwise(range(out.ndim)):
out = out.swapaxes(dim1, dim2)
return out, a__
def get_variational_parametrized_qc(qubits, params):
register = get_quantum_register(len(qubits))
qc = get_new_qc(register)
i = 0
parameters_are_exhausted = False
while not parameters_are_exhausted:
for qubit in qubits:
# rotate qubit with an angle from parameters
theta = params[i] * 2 * np.pi
qc.rz(theta, qubit)
i += 1
for qubit in qubits:
# rotate qubit with an angle from parameters
theta = params[i] * 2 * np.pi
qc.rx(theta, qubit)
i += 1
for qubit_pair in pairwise(qubits):
# do control-Z rotations for neighbour qubits using angles from the params array
theta = params[i] * 2 * np.pi
qc.crz(theta, *list(qubit_pair))
i += 1
parameters_are_exhausted = i < len(params)
return qc
def build_barkan(ppr, background=None):
"""Build a PSSM based on Barkan coding"""
if background == None:
background = (1,1,1,1,)
code = get_code(ppr)
#print "Building model for:\n\t{}".format(
# string_code(code).replace('\n','\n\t'))
PSSM = []
for i,(a,b) in enumerate(pairwise(code)):
s = a[1] + b[0]
if a[2] == 'P' and s in Ptype:
emit = Ptype[s]
elif a[2] == 'S' and s in Stype:
emit = Stype[s]
else:
emit = equal
if sum(emit) == 0:
emit = equal
#convert to log odds
tot = sum(emit)
emit = tuple(log(float(i) / float(tot*b)) for i,b in zip(emit,background))
#print "{:2}: \"{}\" [{}] -> {}".format(i,s,a[2],emit)
PSSM.append(emit)
return PSSM
def information(activation):
data_t = activation
if bins == -1:
data_t = [
bin_array(t, bins=30, low=t.min(), high=t.max())
for t in data_t
]
else:
data_t = [
bin_array(t, bins=bins, low=t.min(), high=t.max())
for t in data_t
]
#data_t = [binarize(t) for t in data_t]
data_t = [hash_data(t) for t in data_t]
h_t = np.array([entropy_of_data(t) for t in data_t])
#h_t_x = np.array([__conditional_entropy(t, data_x) for t in data_t])
h_t_y = np.array([__conditional_entropy(t, data_y) for t in data_t])
h_t_t = np.array(
[__conditional_entropy(t1, t2) for (t1, t2) in pairwise(data_t)])
i_x_t = h_t # - h_t_x # H(T, X) is 0 since every element in X is unique
i_y_t = h_t - h_t_y
i_t_t = h_t[:-1] - h_t_t
return i_x_t, i_y_t, i_t_t
def get_cost_score(self) -> float:
self.initial_port()
for source, dest in utils.pairwise([0] + self.route + [0]):
# проверка на наличие потребности каждого вида топлива
if dest!=0:
if not self.check_demand(dest):
continue
if self.check_capacity(dest):
# автомобиль удовлетворяет условиям multi-compartment
if not self.check_time_and_go(source, dest):
# автомобиль не успевает прибыть к новому потребителю, тогда + авто
# Текущий автомобиль возвращается в депо
self.move_vehicle(source, 0)
# новый атомобиль стартует с депо
self.add_vehicle()
self.move_vehicle(0, dest)
else:
# автомобиль не проходит по ограничениям груза
# возвращается в депо
self.move_vehicle(source, 0)
# расчет дистанции из депо к азс
distance = self.get_distance(0, dest) # just for speeding up (caching)
self.total_travel_dist +=distance
if not self.check_time_and_go(0, dest, distance):
#этот автомобиль не успевает, тогда выезжает новый
self.add_vehicle()
self.move_vehicle(0, dest, distance)
total_travel_cost = Chromosome.get_travel_cost(self.total_travel_dist)
total_vehicles_and_deport_working_hours_cost = self.get_vehicle_count_preference_cost(
vehicles_count=self.vehicles_count,
deport_working_hours=self.max_elapsed_time)
return total_travel_cost + total_vehicles_and_deport_working_hours_cost
def hist_pdf(x, bin_edges, bin_weights):
y = np.zeros_like(x)
for w, (lo, hi) in zip(bin_weights, pairwise(bin_edges)):
idx = (lo <= x) & (x < hi)
y[idx] = w
return y
def hist_err(x, weights, bin_edges, bin_weights):
var = np.zeros_like(bin_weights)
for i, (lo, hi) in enumerate(pairwise(bin_edges)):
idx = (lo <= x) & (x < hi)
var[i] = np.sum(weights[idx]**2)
return np.sqrt(var)
def _build_layers(self):
units = self._get_units()
layers = []
for in_units, out_units in pairwise(units):
layers += [self._build_hidden_layer(in_units, out_units)]
layers += [self._build_final_layer(units[-1])]
return nn.Sequential(*layers)
def hist_err(x, weights, bin_edges, bin_weights):
var = np.zeros_like(bin_weights)
for i, (lo, hi) in enumerate(pairwise(bin_edges)):
idx = (lo <= x) & (x < hi)
var[i] = np.sum(weights[idx]**2)
return np.sqrt(var)
def move(self, how_many, where, delay):
sess = Session.object_session(self)
already = sess.query(MarchingOrder).filter_by(leader=self).first()
if already:
raise InProgressException(already)
fighting = (sess.query(SkirmishAction).
filter_by(participant=self).first())
if fighting:
raise InProgressException(fighting)
if how_many > self.loyalists:
# TODO: Attempt to pick up loyalists
raise InsufficientException(how_many, self.loyalists, "loyalists")
# TODO: Drop off loyalists
where = forcelist(where)
locations = [self.region] + where
for src, dest in pairwise(locations):
if not dest in src.borders:
raise NonAdjacentException(src, dest)
if not dest.enterable_by(self.team):
raise TeamException(dest)
orders = []
if(delay > 0):
orders = []
step = 0
for src, dest in pairwise(locations):
step += 1
mo = MarchingOrder(arrival=time.mktime(time.localtime())
+ delay * step,
leader=self,
source=src,
dest=dest)
orders.append(mo)
sess.add(mo)
else:
self.region = where[-1]
# TODO: Change number of loyalists
self.defectable = False
sess.commit()
return orders
def get_rejection_count_from_status(self, searched_status):
'''
Get number of times an item has transitioned from searched_status to "In Progress"
'''
return len([
a for a, b in pairwise(self.get_status_flow())
if a == searched_status and b == 'In Progress'
])
def compute_edges(self, shop_lists: List[Shop]):
edges = list()
for shop_list1, shop_list2 in pairwise(shop_lists):
for shop1, shop2 in product(shop_list1, shop_list2):
edges.append(
Edge(shop1.identifier, shop2.identifier,
self.compute_distance(shop1, shop2)))
return edges
def hist_pdf(x, bin_edges, bin_weights):
y = np.zeros_like(x)
for w, (lo, hi) in zip(bin_weights, pairwise(bin_edges)):
idx = (lo <= x) & (x < hi)
y[idx] = w
return y
def get_parameters():
parser = argparse.ArgumentParser()
parameters_data(parser)
network_parameters(parser)
parser.add_argument(
'--bins',
'-b',
dest='bins',
default=30,
type=int,
help="select the number of bins to use for bnning defaults to 30")
parser.add_argument('--dest',
dest='dest',
default="output",
help="destination folder for output files")
parser.add_argument(
'--epoch_list',
'-el',
dest='epoch_list',
default="1-19,20-90",
help=
"list of ranges for which to compute mi. ex: 1-10,15-30, eanges assumed to be disjoint"
)
parser.add_argument(
'--saved_epochs',
'-se',
dest='no_saved_epochs',
default=100,
type=int,
help="no of epochs to consider when calculating mutual information")
args = parser.parse_args()
def convert_to_range(rrange):
split = rrange.split('-')
return int(split[0]), int(split[1])
args.epoch_list = args.epoch_list
args.epoch_list = args.epoch_list.split(',')
args.epoch_list = [convert_to_range(e) for e in args.epoch_list]
for s, e in args.epoch_list:
if s > e:
raise ValueError("invalid range {}".format((s, e)))
if s == e:
raise ValueError("range is 0, start {}, end {}".format(s, e))
for a, b in pairwise(args.epoch_list):
_, e = a
s, _ = b
if s < e:
raise ValueError("ranges {} and {} overlap".format(a, b))
return args
def get_path_attr_list(self, path, attrs):
attr_map = defaultdict(list)
for (n1, n2) in U.pairwise(path):
v1 = self.vd[n1]
v2 = self.vd[n2]
e = self.G.edge(v1, v2, add_missing=False)
for attr in attrs:
attr_map[attr].append(self.edge_attr[e][attr])
return attr_map
def get_path_attr_list(self, path, attrs):
attr_map = defaultdict(list)
for (n1, n2) in U.pairwise(path):
v1 = self.vd[n1]
v2 = self.vd[n2]
e = self.G.edge(v1, v2, add_missing=False)
for attr in attrs:
attr_map[attr].append(self.edge_attr[e][attr])
return attr_map
def parse_driver(line):
driver = [int(x) for x in line.split()]
if len(driver) % 6 != 0:
raise Exception(
'All users should have origin ox oy and destination dx dy')
idxs, points = get_indexes_and_points_separated_lists(driver)
point_tuples = tuple([(x, y) for x, y in utils.pairwise(points)])
idxs_and_points = [(idx, point) for idx, point in zip(idxs, point_tuples)]
return idxs_and_points
def get_yagi_NSRs(ppr): """Return a list of 3-tuples referring to the NSRs""" o = 3 * 1 NSR = [] for a,b in pairwise(ppr.features): s1 = ppr.seq[a.location.start+o:o+a.location.start + 1 + 3 * 4].translate() s2 = ppr.seq[b.location.start - 3*2+o:o+b.location.start].translate() NSR.append((s1[0], s1[3], s2[0],)) return NSR
def __init__(self, category_factors=(1, 1, 1, 1, 1)):
if category_factors[0] <= 0:
raise ValueError("ScoringScheme: category factors must be > 0")
for f1, f2 in pairwise(category_factors):
if f1 > f2:
raise ValueError(
"ScoringScheme: category factors must in non-decreasing order"
)
self.category_factors = category_factors
def tracks(self):
_tracks = self.tmx['track']
segments = []
for track in _tracks:
for (x1, y1), (x2, y2) in pairwise(track.points):
segments.append(
LineSegment(
x1 + track.x, -y1 + track.y,
x2 + track.x, -y2 + track.y))
return segments
def compositions(n, m):
"""
>>> len(list(compositions(4, 3)))
15
>>> list(compositions(2, 1))
[(2,)]
>>> list(compositions(2, 2))
[(0, 2), (1, 1), (2, 0)]
"""
for c in combinations(range(n + m - 1), m - 1):
yield tuple(b-a-1 for a, b in pairwise((-1,) + c + (n+m-1,)))
def warp(self):
if not self.convexOnly:
pairs = [pair for pair in pairwise(self.hullVertices)]
for p1, p2 in pairs:
# TODO: implement min distance check?
mdpt = midpoint(p1, p2)
angle = random.randint(self.minWarpAngle, self.maxWarpAngle)
newPoint = endpoint(p1, mdpt, angle, 'd')
self.addedPoints.append(newPoint)
# interleave self.hullVertices and newPoints
self.hullVertices = [point for point in roundrobin(self.hullVertices, self.addedPoints)]
def get_wiki_table(cls, soup_object):
trs = soup_object.find('table', class_='infobox').find_all('tr')
table = {}
for tr in trs:
# Look at each pair of td tags, since infobox rows that we care about have two columns
for h_text, v_text in [(replace_xa0(h.text), v.text) for h, v in pairwise(tr.find_all('td'))]:
if h_text in WikipediaExtractor.headers_to_terms:
comp_term = WikipediaExtractor.headers_to_terms[h_text]
# removes blank lines and blank indexes
table[comp_term] = [x for x in re.split('\n+', v_text) if x != u'']
return table
def print_cluster_sim_mat(self): print "CLUSTER SIM MATRICES:" centers = map(lambda cluster: cluster.center, self.clusters) feature_set_names = self.docs[0].get_feature_set_names() for name in feature_set_names: print print "Similarity Type: %s" % name mat = utils.pairwise(centers, lambda doc1, doc2: doc1.global_sim(doc2, name)) mat = utils.apply_mat(mat, lambda x: "%3.2f" % x) utils.insert_indices(mat) utils.print_mat(mat) print print "Similarity Type: Cluster sim by CONFIRM" sub_mat = utils.pairwise(self.clusters, lambda c1, c2: self.confirm.cluster_similarity(c1, c2)) sub_mat = utils.apply_mat(sub_mat, lambda x: "%3.2f" % x) utils.insert_indices(sub_mat) utils.print_mat(sub_mat) print print
def path_to_dict(path):
"""Converts the path data (which is a list of node IDs) from point A to point B,
to list of edges (i.e. pairs of nodes) corresponding to the path. That data is then
written to the path_dict observable for consumption by the websockets frontend.
Format:
{'path': list of node pairs,
'endpoints': tuple: (start point ID, end point ID)}
"""
global path_dict
path_dict.overwrite({
'path': [sorted(x) for x in pairwise(path)],
'endpoints': (path[0], path[-1])
})
def find_used_labels(cfg):
labels = set()
for addr, nxt in pairwise(cfg.iter_rev_postorder()):
info = cfg[addr]
bblock = info["val"]
succs = cfg.sorted_succ(addr)
if len(succs) > 1 and nxt == succs[0]:
swap_if_branches(cfg, addr)
succs = cfg.sorted_succ(addr)
for succ in succs:
cond = cfg.edge(addr, succ).get("cond")
if not cond and nxt and succ == nxt:
continue
labels.add(succ)
return labels
def pitches_to_intervals(chord):
"""
>>> pitches_to_intervals([12, 16, 31])
[4, 3, 5]
"""
# TODO rewrite this
pcs = sorted(list(set([p % 12 for p in chord])))
pcs = [p - pcs[0] for p in pcs]
pcs.reverse()
pcs.insert(0, 12)
intervals = []
for a, b in pairwise(pcs):
intervals.append(a - b)
intervals.reverse()
return intervals
def _init_global_thresh(self): sub_docs = self.docs[:20] sim_mat = utils.pairwise(sub_docs, lambda x, y: max(self.doc_similarity(x, y), self.doc_similarity(y, x))) for x in xrange(len(sim_mat)): del sim_mat[x][x] self.global_thresh = .7 #utils.avg(map(max, sim_mat)) print print "INITIAL GLOBAL THRESH: ", self.global_thresh tmp = utils.flatten(sim_mat) tmp.sort() print map(lambda x: "%.2f" % x, tmp) print map(lambda x: "%.2f" % x, map(max, sim_mat)) print self.sim_sum = self.global_thresh * self.initial_thresh_weight self.num_counted = self.initial_thresh_weight
def generate(self, streams):
newstreams = []
output = []
i = 0
for s1, s2 in pairwise(streams):
newstream = self._outstreamprefix
if i > 0: newstream += str(i)
output.append(
"[{b}][{o}]{e}[{ns}]".format(
b = s1,
o = s2,
e = self._expressions_accessor.get_expression(i),
ns = newstream)
)
newstreams.append(newstream)
i += 1
return output, newstreams
def simulate_entire_trajectories(self, sim_states, T):
comm = self.communicator
comm.put_value('initial_continuous_states', sim_states.cont_states)
comm.put_value('initial_discrete_states', sim_states.discrete_states)
comm.put_value('initial_pvt_states', sim_states.pvt_states)
comm.put_value('T', np.array([T]))
ret_val_str_list = ['data_mat', 'idx_arr']
arg_str_list = [
'sim_function',
'initial_continuous_states',
'initial_discrete_states',
'initial_pvt_states',
'inputs',
'T',
]
fun_name_str = 'simulate_entire_trajectories'
if self.parallel:
[data_arr, idx_arr] = comm.call_function_retVal(ret_val_str_list, fun_name_str, arg_str_list)
else:
logger.warning('non-parallel simulate_entire_trajectories isunimplemented'
)
raise NotImplementedError('single threaded implementation missing')
idx_arr = self.matlab2py_indices(idx_arr)
list_of_trajs = []
# quick sanity check
if len(idx_arr.shape) != 1 or idx_arr[0] != 0:
print 'idx_arr: ', idx_arr
raise err.Fatal('sanity check on idx_arr fails!')
# get a pairwise iterator for the index array
for (i, j) in utils.pairwise(idx_arr):
list_of_trajs.append(data_arr[i:j, :])
# append the last trajectory
logger.info('Populating trajectories...')
list_of_trajs.append(data_arr[j:, :])
# ##!!##logger.debug('===Trajectories===\n%s', str(list_of_trajs))
return list_of_trajs
def parse_line(line, fields):
"""Parses a line of fixed width data.
Fields should be a list of tuples of form:
(field_start, field_name), field_start being the column number at which the
field begins.
"""
data = {}
field_starts = [field[0] for field in fields]
# Construct slices ranging from the start of each field to the start of the
# next field.
slices = [slice(a, b) for a, b in pairwise(field_starts + [None])]
for slice_, field in zip(slices, fields):
field_name = field[1]
value = line[slice_].strip()
data[field_name] = value
return data
def _init_clusters(self): sub_docs = self.docs[:self.num_instances] sim_mat = utils.pairwise(sub_docs, lambda x, y: max(self.doc_similarity(x, y), self.doc_similarity(y, x))) edges = utils.minimum_spanning_tree(sim_mat) ccs = utils.get_ccs(range(self.num_instances), edges) biggest_cc = max(map(len, ccs)) while biggest_cc > self.num_init: edge_to_remove = random.sample(edges, 1)[0] edges.remove(edge_to_remove) ccs = utils.get_ccs(range(self.num_instances), edges) biggest_cc = max(map(len, ccs)) cc = ccs[utils.argmax(map(len, ccs))] for idx in cc: self._add_cluster(self.docs[idx], member=False)
def _init_clusters(self): super(MaxCliqueInitCONFIRM, self)._init_clusters() sub_docs = self.docs[:self.num_instances] sim_mat = utils.pairwise(sub_docs, lambda x, y: max(self.doc_similarity(x, y), self.doc_similarity(y, x))) print print "Doc Sim Mat" utils.print_mat(utils.apply_mat(sim_mat, lambda x: "%3.2f" % x)) idxs = utils.find_best_clique(sim_mat, self.num_clust) print print "Cluster Labels:" for idx in idxs: self._add_cluster(self.docs[idx], member=False) print idx, self.docs[idx].label
def is_good(ornament, biggest_interval=3):
# Put the target note at the end of the ornament
# to prevent repeated notes or large intervals
# between the last note of the ornament and the target note
ornament = list(ornament)
ornament.append(0)
pairs = pairwise(ornament)
for a, b in pairs:
# Don't allow repeated pitches
if a == b:
return False
# Don't allow transitions greater than `biggest_interval` steps
if abs(a - b) > biggest_interval:
return False
return True
def maxFlowOpt(reads, maxReadsPerPos, minReadsPerPos):
"""
reads:
maxReadsPerPos: maximum coverage (aka k)
minReadsPerPos: minimum coverage (aka t)
"""
# flatten read start and end into one list
positions = sorted(set([x for sublist in reads for x in sublist]))
backboneCapacity = maxReadsPerPos - minReadsPerPos
sourceSinkCapacity = maxReadsPerPos
readIntervalCapacity = 1
superSource = -1
superSink = positions[-1] + 1
g = nx.DiGraph()
# we remove reads directly from the list
reads = copy.copy(reads)
# add backbone
g.add_edge(superSource, positions[0], {'capacity': sourceSinkCapacity})
g.add_edge(positions[-1], superSink, {'capacity': sourceSinkCapacity})
for (p0, p1) in utils.pairwise(positions):
g.add_edge(p0, p1, {'capacity': backboneCapacity})
# add intervals
for read in reads:
g.add_edge(read[0], read[1], {'capacity': readIntervalCapacity})
# TODO: implement flow algorithm
val, mincostFlow = nx.maximum_flow(g, superSource, superSink)
pruned = []
# TODO: this is in O(n) -> we can do it in O(log n)
for key, values in mincostFlow.items():
for value, weight in values.items():
if weight is 0:
toRemove = (key, value)
if toRemove in reads:
pruned.append((key, value))
reads.remove((key, value))
return reads, pruned, mincostFlow
def find_gaps(ppr, mingap=30, maxgap=None, skip_introns=True):
"""Find all the gaps between PPR motifs which are gte maxgap"""
loc = []
feats = sorted(ppr.features, key = lambda(p): int(p.location.start))
for a,b in pairwise(feats):
#ignore any gaps if a and b aren't in the same frame
if a.qualifiers['frame'] != b.qualifiers['frame']:
continue
#if the size is within the range
l = int(b.location.start) - int(a.location.end)
if l >= (mingap or -float('inf')) and l <= (maxgap or float('inf')):
#We've found a gap
g = FeatureLocation(int(a.location.end),
int(b.location.start), strand=1)
g.prev = a
g.next = b
loc.append(g)
return loc
def dump_c(cfg, stream=sys.stdout):
labels = find_used_labels(cfg)
func_start = True
for addr, nxt in pairwise(cfg.iter_rev_postorder()):
info = cfg[addr]
bblock = info["val"]
if func_start:
label = cfg.props["name"]
if not label:
label = cfg.parser.label_from_addr(bblock.addr)
if label[0].isdigit():
label = "fun_" + label
if ("estimated_params" in cfg.props):
print("// Estimated params: %s" % sorted(list(cfg.props["estimated_params"])), file=stream)
if cfg.props["trailing_jumps"]:
print("// Trailing jumps not removed, not rendering CFG edges as jumps", file=stream)
func_props = progdb.FUNC_DB.get(label, {})
params = ""
if "params" in func_props:
params = sorted(func_props["params"], key=natural_sort_key)
params = ", ".join(["u32 " + str(r) + "_0" for r in params])
print("void %s(%s)\n{" % (label, params), file=stream)
func_start = False
if addr in labels:
print("\nl%s:" % addr, file=stream)
bblock.dump(stream, indent=1, printer=print_inst)
if not cfg.props["trailing_jumps"]:
for succ in cfg.succ(addr):
cond = cfg.edge(addr, succ).get("cond")
if not cond and nxt and succ == nxt:
continue
stream.write(" ")
if cond:
stream.write("if %s " % cond)
print("goto l%s;" % succ, file=stream)
print("}", file=stream)
def pseudoDistance(clustering):
# for each cluster, returns the estimated pairwise distances of every point to every point in the cluster
distances = []
total = 0.0
count = 0
startTime = time.time()
for cluster in clustering:
# print "Cluster Size:", len(cluster.members)
'''
print "Starting True Similarities"
startTime = time.time()
trueSimilarities = map(lambda r: map(lambda c: 1-c, r) ,utils.pairwise(cluster.members, lambda x,y: x.similarity(y)))
endTime = time.time()
print "Done. Elapsed Time:", endTime-startTime
'''
#Order Matters here.
'''
namedSims = map(lambda x: cluster.center.similarities_by_name(x).items(), [cluster.center] + cluster.members)
distToCenter = map(lambda x: [1-i[1] for i in x], namedSims)
euclid = utils.pairwise(distToCenter, lambda x,y: utils.euclideanDistance(x,y))
distances.append(euclid)
'''
vectors = map(lambda x: cluster.center.similarity_vector(x), [cluster.center] + cluster.members)
distToCenter = map(lambda x: map(lambda i: 1-i, x), vectors)
euclid = utils.pairwise(distToCenter, lambda x,y: utils.euclideanDistance(x,y))
distances.append(euclid)
endTime = time.time()
print "Done. Elapsed Time:", endTime-startTime
return distances
def refine_rel_model_based(
AA, error_paths, pi_seq_list, sp, sys_sim, opts, sys, prop):
'''does not handle pi_seq_list yet'''
# abs_state relations: maps an abs_state to other abs_states
# reachable in one time step
abs_relations = defaultdict(set)
for path in error_paths:
# abs_state_1 -> abs_state_2
for a1, a2 in U.pairwise(path):
abs_relations[a1].add(a2)
flat_relations = []
for abs_state, rch_states in abs_relations.iteritems():
flat_relation = list(itertools.product([abs_state], rch_states))
flat_relations.extend(flat_relation)
pwa_model = build_pwa_model(
AA, flat_relations, sp, opts.max_model_error,
opts.model_err, 'rel')
if __debug__:
sim_n_plot(error_paths, pwa_model, AA, sp, opts)
check4CE(pwa_model, error_paths, sys.sys_name, 'rel', AA, sp, opts.bmc_engine)
def move(self, how_many, where, delay, sector=0, conf=None):
sess = Session.object_session(self)
where = forcelist(where)
already = sess.query(MarchingOrder).filter_by(leader=self).first()
if already:
raise InProgressException(already)
fighting = (sess.query(SkirmishAction).
filter_by(participant=self).first())
if fighting:
allow = False
if conf:
if len(where) == 1 and where[0] == self.region:
allow = conf["game"].get("allow_sector_retreat", False)
if not allow:
raise InProgressException(fighting)
if how_many > self.loyalists:
# TODO: Attempt to pick up loyalists
raise InsufficientException(how_many, self.loyalists, "loyalists")
# Is that sector even real?
if conf:
num_sectors = conf["game"].get("num_sectors", 1)
if sector < 0 or sector > num_sectors:
raise NoSuchSectorException(sector, num_sectors)
elif sector == 0: # Assign a random sector
sector = random.randint(1, num_sectors)
# TODO: Drop off loyalists
locations = [self.region] + where
for src, dest in pairwise(locations):
if src == dest:
continue
if dest not in src.borders:
raise NonAdjacentException(src, dest)
traverse_neutrals = False
if conf:
traverse_neutrals = conf["game"].get("traversable_neutrals", False)
if not dest.enterable_by(self.team,
traverse_neutrals=traverse_neutrals):
raise TeamException(dest)
orders = []
if delay > 0:
orders = []
total_delay = 0
for src, dest in pairwise(locations):
travel_mult = max(src.travel_multiplier, dest.travel_multiplier)
if src != dest:
total_delay += (delay * travel_mult)
else:
if conf:
intrasector = conf["game"].get("intrasector_travel",
900)
# Travel multiplier doesn't apply to intrasector
total_delay += intrasector
mo = MarchingOrder(arrival=time.mktime(time.localtime())
+ total_delay,
leader=self,
source=src,
dest=dest,
dest_sector=sector)
orders.append(mo)
sess.add(mo)
else:
self.region = where[-1]
self.sector = sector
# TODO: Change number of loyalists
self.defectable = False
sess.commit()
return orders
def _handle_openflow_PacketIn(self, event):
# TODO: Refactor this method
packet = event.parsed
source = packet.src
destination = packet.dst
# self.log.info("SRC: %s" % source)
# self.log.info("DST: %s" % destination)
if destination.is_multicast:
# Note: pick between 2 flooding techniques here
# Flood the packet
# ** Manual flooding **
# TODO: Install new flow instead of crafting new packet (hold down?)
for mac, entry in core.host_tracker.entryByMAC.iteritems():
if source == mac:
continue
message = of.ofp_packet_out()
message.actions.append(of.ofp_action_output(port=entry.port))
message.data = event.data
# message.data = event.ofp
core.overseer_topology.graph.node[entry.dpid]['connection'].send(message)
self.log.debug("FLOODING: MAC %s is connected to switch %s at port %s" % (mac, entry.dpid, entry.port))
# ** Use OFPP_FLOOD **
# message = of.ofp_packet_out()
# message.actions.append(of.ofp_action_output(port=of.OFPP_FLOOD))
# message.buffer_id = event.ofp.buffer_id
# # message.data = event.ofp
# message.in_port = event.port
# event.connection.send(message)
return
entryByMAC = core.host_tracker.entryByMAC
known_hosts = entryByMAC.keys()
if (source not in known_hosts) or (destination not in known_hosts):
# Ignore non-end-to-end packet
self.log.info("There is no path from host %s to host %s" % (source, destination))
return
self.log.info("Finding path from host %s to host %s" % (source, destination))
from_host = entryByMAC[source]
to_host = entryByMAC[destination]
path = self.get_path(from_host.dpid, to_host.dpid, packet)
match = of.ofp_match.from_packet(packet)
match.in_port = None
# match.dl_src = None
# match.dl_dst = None
#match.dl_vlan = None
#match.dl_vlan_pcp = None
# match.nw_proto = None
#match.nw_tos = None
self.log.info("Installing path from host %s to host %s" % (source, destination))
# Install flows
# TODO: Handle buffer_id properly
# first = True
for from_switch, to_switch in utils.pairwise(path):
self.log.info("Installing flow from switch %x to switch %x" % (from_switch, to_switch))
portByDpid = core.overseer_topology.graph.get_edge_data(from_switch, to_switch)["portByDpid"]
message = of.ofp_flow_mod()
message.match = match
message.idle_timeout = self.flow_idle_timeout
message.hard_timeout = self.flow_hard_timeout
message.actions.append(of.ofp_action_output(port=portByDpid[from_switch]))
# self.log.info("DEBUG: %s %d" % (type(portByDpid[from_switch]), portByDpid[from_switch]))
# if first:
# message.buffer_id = event.ofp.buffer_id
# first = False
core.overseer_topology.graph.node[from_switch]['connection'].send(message)
# Install final flow
self.log.info("Installing final flow from switch %x to host %s" % (path[-1], destination))
message = of.ofp_flow_mod()
message.match = match
message.idle_timeout = self.flow_idle_timeout
message.hard_timeout = self.flow_hard_timeout
message.actions.append(of.ofp_action_output(port=to_host.port))
core.overseer_topology.graph.node[path[-1]]['connection'].send(message)
