class memoized(object):
"""A memoized decorator with reset functionality for instance methods
"""
memo = makehash()
n_call = 0
n_hit = 0
n_reset = 0
def __init__(self, f):
self.f = f
functools.update_wrapper(self, f)
@staticmethod
def key(*args, **kwargs):
args = flatten(list(args))
return tuple(args), frozenset(kwargs.items())
def __call__(self, obj, *args, **kwargs):
memoized.n_call += 1
key = memoized.key(*args, **kwargs)
fn = self.f.__name__
if key not in memoized.memo[obj][fn]:
memoized.memo[obj][fn][key] = self.f(obj, *args, **kwargs)
else:
memoized.n_hit += 1
return memoized.memo[obj][fn][key]
def __get__(self, obj, objtype):
return functools.partial(self.__call__, obj)
@staticmethod
def reset(name=None):
def wrapper(f):
@functools.wraps(f)
def helper(obj, *args, **kwargs):
if name and name in memoized.memo[obj]:
del memoized.memo[obj][name]
elif obj in memoized.memo:
del memoized.memo[obj]
memoized.n_reset += 1
return f(obj, *args, **kwargs)
return helper
return wrapper
@classmethod
def statistics(cls):
if cls.n_call:
print >> sys.stderr, 'memoized hitting rate: %d/%d=%f' % \
(cls.n_hit, cls.n_call,
cls.n_hit / cls.n_call)
if cls.n_reset:
print >> sys.stderr, 'memoized reset times: %d' % cls.n_reset
def load(opened_file):
codes = makehash()
loaded = makehash()
for line in opened_file:
splitted = line.strip().split('\t')
if len(splitted) == 4:
k1, k2, idx, code = splitted
codes[k1][k2][int(idx)] = code
for k1 in codes:
for k2 in codes[k1]:
joined = ''.join(codes[k1][k2][i]
for i in sorted(codes[k1][k2].keys()))
loaded[k1][k2] = decode(joined)
return loaded
def mat_to_dict(self, charlist, ignore_this={}):
"""Return a dict from the mat"""
mat_val = utils.makehash(1, list)
for node in self.tree.traverse():
for i, val in enumerate(charlist):
for st in self.state_mat[i, node.ind]:
ignore_corr = (ignore_this.get(val, "") == st)
if not ignore_corr:
mat_val[node.name][st].append(val)
return mat_val
def flip_rea_forward(clc, nodestates):
"""Flip rea data dict"""
new_dt = utils.makehash(1, set)
state_map = defaultdict(set)
for (genome, aarea) in nodestates.items():
nl = [(c, aa) for aa, codons in aarea.items() for c in codons]
for (c, aa) in nl:
new_dt[genome][c].add(aa)
state_map[c].add(aa)
return new_dt, state_map
def main():
loaded = stream.load(sys.stdin)
merged = makehash()
for k1 in loaded:
merged[k1] = SemiMarkov()
for k2 in loaded[k1]:
merged[k1] += loaded[k1][k2]
stream.dump(k1, merged[k1])
def create_login(user, password, update=False, superuser=None):
hash = makehash(password)
result = web.ctx.db.where('user', login=user).first()
data = {}
if password:
data['password'] = hash
if superuser is not None:
data['superuser'] = superuser
if result and update:
web.ctx.db.update(
'user', where='login=$user', vars=locals(), **data)
else:
web.ctx.db.insert('user', login=user, **data)
def evaluate(samples, models,
use_duration=True,
window=semi_markov.WINDOW):
confusion_matrix = makehash()
for klass, file_names in samples.items():
counters = statistics(
file_names, models,
use_duration=use_duration,
window=window)
for threshold in counters:
for term, counter in counters[threshold].items():
key = '_'.join([klass, term])
confusion_matrix[threshold][key] = counter
for threshold, matrix in sorted(confusion_matrix.items()):
pv('threshold', stdout=True)
print_statistics(matrix, threshold)
print
