def bin_size(orig_exp, ci_method, nib_lim=0.01, max_ci_diff=0.25, min_corcoef=0.3):
for bin_size in itertools.count(1):
exp = orig_exp.aggregate_bins(times_bin_size=bin_size)
df = logit.ci_for_df(exp.as_data_frame(), ci_method, ci_min=max_ci_diff)
ratio_nib = logit.get_nib_ratio(df)
pvar = corrcoeff(df)
log.notice('testing bin size %d, nib ratio: %.4f, spearmanr: %.3f' % (bin_size, ratio_nib, pvar))
if ratio_nib <= nib_lim and pvar > min_corcoef:
return exp.bin_size
assert bin_size < 100, "Could not find a suitable bin size."
def bin_size(orig_exp,
ci_method,
nib_lim=0.01,
max_ci_diff=0.25,
min_corcoef=0.3):
for bin_size in itertools.count(1):
exp = orig_exp.aggregate_bins(times_bin_size=bin_size)
df = logit.ci_for_df(exp.as_data_frame(),
ci_method,
ci_min=max_ci_diff)
ratio_nib = logit.get_nib_ratio(df)
pvar = corrcoeff(df)
log.notice('testing bin size %d, nib ratio: %.4f, spearmanr: %.3f' %
(bin_size, ratio_nib, pvar))
if ratio_nib <= nib_lim and pvar > min_corcoef:
return exp.bin_size
assert bin_size < 100, "Could not find a suitable bin size."
def __adjust_bin_size_and_get_df(self, exp):
if self.bin_size is None:
self.bin_size = estimate.bin_size(exp, self.ci_method,
max_ci_diff=self.ci_lim,
nib_lim=self.nib_lim)
log.notice('Optimal bin size: %d' % self.bin_size)
else:
log.notice('Using preset bin size of: %d' % self.bin_size)
odf = exp.aggregate_bins(new_bin_size=self.bin_size).as_data_frame()
df = logit.ci_for_df(odf, self.ci_method, ci_min=self.ci_lim)
ratio_nib = logit.get_nib_ratio(df)
log.notice('''Manually specified bin size of %dKB gives %.2f%% informative bins. \
The required amount is %.2f%%.''' % (self.bin_size / 1000, (1-ratio_nib)*100, (1-self.nib_lim)*100))
assert ratio_nib < self.nib_lim, '''\
The selected bin size results in less informative bins that what specified by the\
parameter required_fraction_of_informative_bins. Please try a bigger bin size or let \
EDD auto-estimate a bin size for you. Please consult the EDD manual for more information'''
return df
def __adjust_bin_size_and_get_df(self, exp):
if self.bin_size is None:
self.bin_size = estimate.bin_size(exp,
self.ci_method,
max_ci_diff=self.ci_lim,
nib_lim=self.nib_lim)
log.notice('Optimal bin size: %d' % self.bin_size)
else:
log.notice('Using preset bin size of: %d' % self.bin_size)
odf = exp.aggregate_bins(new_bin_size=self.bin_size).as_data_frame()
df = logit.ci_for_df(odf, self.ci_method, ci_min=self.ci_lim)
ratio_nib = logit.get_nib_ratio(df)
log.notice(
'''Manually specified bin size of %dKB gives %.2f%% informative bins. \
The required amount is %.2f%%.''' % (self.bin_size / 1000,
(1 - ratio_nib) * 100,
(1 - self.nib_lim) * 100))
assert ratio_nib < self.nib_lim, '''\
The selected bin size results in less informative bins that what specified by the\
parameter required_fraction_of_informative_bins. Please try a bigger bin size or let \
EDD auto-estimate a bin size for you. Please consult the EDD manual for more information'''
return df
