STRLCPY/isp-data-pollution

Revision indexing in progress... (symbol navigation in revisions will be accurate after indexed)

Total 1 files

■ ■ ■ ■ ■ ■

isp_data_pollution.py

		skipped 454 lines
455	455		except Exception as e:
456	456		if self.debug: print('.update() exception:\n{}'.format(e))
457	457
458		-	def draw_link(self):
459		-	return self.draw_links(n=1)[0]
	458	+	def draw_link(self,log_sampling=True):
	459	+	return self.draw_links(n=1,log_sampling=log_sampling)[0]
460	460
461		-	def draw_links(self,n=1):
	461	+	def draw_links(self,n=1,log_sampling=False):
462	462		urls = []
463	463		domain_count = np.array([(dmn,len(self.domain_links[dmn])) for dmn in self.domain_links])
464	464		p = np.array([np.float(c) for d,c in domain_count])
465	465		count_total = p.sum()
466		-	# log-sampling [log(x+1)] to bias lower count domains
467		-	p = np.fromiter((np.log1p(x) for x in p), dtype=p.dtype)
	466	+	if log_sampling: # log-sampling [log(x+1)] to bias lower count domains
	467	+	p = np.fromiter((np.log1p(x) for x in p), dtype=p.dtype)
468	468		if count_total > 0:
469	469		p = p/p.sum()
470	470		cnts = npr.multinomial(n, pvals=p)
		skipped 212 lines