def sort_hits(self, local_pp_scores, urlparams):
"""
Combine IR (Information Relevant) score given by Elasticsearch,
with PP (Page Popularity) score, to sort the results
"""
if not self.object_list:
return
hits = self.object_list.hits
ir_scores = []
pp_globl_scores = []
pp_local_scores = []
for h in hits:
ir_scores.append(h.get('score', 0))
pp_globl_scores.append(PagePopScore.objects.get_score(onion=h['domain']))
pp_local_scores.append(local_pp_scores[h['domain']])
ir_scores_norm = utils.normalize_on_max(ir_scores)
pp_globl_scores_norm = utils.normalize_on_max(pp_globl_scores)
pp_local_scores_norm = utils.normalize_on_max(pp_local_scores)
if settings.DEBUG:
assert len(ir_scores_norm) == len(pp_globl_scores_norm) == \
len(pp_local_scores_norm) == len(hits)
for h, ir, pp, pl in zip(hits, ir_scores_norm, pp_globl_scores_norm,
pp_local_scores_norm):
h['score'] = heuristic_score(ir, pp, pl, urlparams)
self.object_list.hits = sorted(hits, key=lambda k: k['score'], reverse=True)
def sort_hits(self, local_pp_scores, urlparams):
"""
Combine IR (Information Relevant) score given by Elasticsearch,
with PP (Page Popularity) score, to sort the results
"""
if not self.object_list:
return
hits = self.object_list.hits
ir_scores = []
pp_globl_scores = []
pp_local_scores = []
for h in hits:
ir_scores.append(h.get('score', 0))
pp_globl_scores.append(
PagePopScore.objects.get_score(onion=h['domain']))
pp_local_scores.append(local_pp_scores[h['domain']])
ir_scores_norm = utils.normalize_on_max(ir_scores)
pp_globl_scores_norm = utils.normalize_on_max(pp_globl_scores)
pp_local_scores_norm = utils.normalize_on_max(pp_local_scores)
if settings.DEBUG:
assert len(ir_scores_norm) == len(pp_globl_scores_norm) == \
len(pp_local_scores_norm) == len(hits)
for h, ir, pp, pl in zip(hits, ir_scores_norm, pp_globl_scores_norm,
pp_local_scores_norm):
h['score'] = heuristic_score(ir, pp, pl, urlparams)
self.object_list.hits = sorted(hits,
key=lambda k: k['score'],
reverse=True)
def sort_results(self):
"""
Combine IR (Information Relevant) score given by Elasticsearch,
with PP (Page Popularity) score, to sort the results
"""
hits = self.object_list[1] # object_list is tuple(int, list)
if not hits:
return
ir_scores = [h.get('score', 0) for h in hits]
pp_scores = [PagePopScore.objects.get_score(onion=h['domain'])
for h in hits]
ir_scores_norm = utils.normalize_on_max(ir_scores)
pp_scores_norm = utils.normalize_on_max(pp_scores)
assert len(ir_scores_norm) == len(pp_scores_norm) == len(hits)
for h, ir, pp in zip(hits, ir_scores_norm, pp_scores_norm):
h['score'] = heuristic_score(ir, pp)
self.object_list[1] = sorted(
hits,
key=lambda k: k['score'],
reverse=True)
