-
Notifications
You must be signed in to change notification settings - Fork 731
/
comparison.py
240 lines (204 loc) · 6.94 KB
/
comparison.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
"""Comparison tools for Zinnia"""
from math import sqrt
from django.contrib.sites.models import Site
from django.core.cache import InvalidCacheBackendError
from django.core.cache import caches
from django.utils import six
from django.utils.functional import cached_property
from django.utils.html import strip_tags
import regex as re
from zinnia.models.entry import Entry
from zinnia.settings import COMPARISON_FIELDS
from zinnia.settings import STOP_WORDS
PUNCTUATION = re.compile(r'\p{P}+')
def pearson_score(list1, list2):
"""
Compute the Pearson' score between 2 lists of vectors.
"""
size = len(list1)
sum1 = sum(list1)
sum2 = sum(list2)
sum_sq1 = sum([pow(l, 2) for l in list1])
sum_sq2 = sum([pow(l, 2) for l in list2])
prod_sum = sum([list1[i] * list2[i] for i in range(size)])
num = prod_sum - (sum1 * sum2 / float(size))
den = sqrt((sum_sq1 - pow(sum1, 2.0) / size) *
(sum_sq2 - pow(sum2, 2.0) / size))
return num / den
class ModelVectorBuilder(object):
"""
Build a list of vectors based on a Queryset.
"""
limit = None
fields = None
queryset = None
def __init__(self, **kwargs):
self.limit = kwargs.pop('limit', self.limit)
self.fields = kwargs.pop('fields', self.fields)
self.queryset = kwargs.pop('queryset', self.queryset)
def get_related(self, instance, number):
"""
Return a list of the most related objects to instance.
"""
related_pks = self.compute_related(instance.pk)[:number]
related_pks = [pk for pk, score in related_pks]
related_objects = sorted(
self.queryset.model.objects.filter(pk__in=related_pks),
key=lambda x: related_pks.index(x.pk))
return related_objects
def compute_related(self, object_id, score=pearson_score):
"""
Compute the most related pks to an object's pk.
"""
dataset = self.dataset
object_vector = dataset.get(object_id)
if not object_vector:
return []
object_related = {}
for o_id, o_vector in dataset.items():
if o_id != object_id:
try:
object_related[o_id] = score(object_vector, o_vector)
except ZeroDivisionError:
pass
related = sorted(object_related.items(),
key=lambda k_v: (k_v[1], k_v[0]), reverse=True)
return related
@cached_property
def raw_dataset(self):
"""
Generate a raw dataset based on the queryset
and the specified fields.
"""
dataset = {}
queryset = self.queryset.values_list(*(['pk'] + self.fields))
if self.limit:
queryset = queryset[:self.limit]
for item in queryset:
item = list(item)
item_pk = item.pop(0)
datas = ' '.join(map(six.text_type, item))
dataset[item_pk] = self.raw_clean(datas)
return dataset
def raw_clean(self, datas):
"""
Apply a cleaning on raw datas.
"""
datas = strip_tags(datas) # Remove HTML
datas = STOP_WORDS.rebase(datas, '') # Remove STOP WORDS
datas = PUNCTUATION.sub('', datas) # Remove punctuation
datas = datas.lower()
return [d for d in datas.split() if len(d) > 1]
@cached_property
def columns_dataset(self):
"""
Generate the columns and the whole dataset.
"""
data = {}
words_total = {}
for instance, words in self.raw_dataset.items():
words_item_total = {}
for word in words:
words_total.setdefault(word, 0)
words_item_total.setdefault(word, 0)
words_total[word] += 1
words_item_total[word] += 1
data[instance] = words_item_total
columns = sorted(words_total.keys(),
key=lambda w: words_total[w],
reverse=True)[:250]
columns = sorted(columns)
dataset = {}
for instance in data.keys():
dataset[instance] = [data[instance].get(word, 0)
for word in columns]
return columns, dataset
@property
def columns(self):
"""
Access to columns.
"""
return self.columns_dataset[0]
@property
def dataset(self):
"""
Access to dataset.
"""
return self.columns_dataset[1]
class CachedModelVectorBuilder(ModelVectorBuilder):
"""
Cached version of VectorBuilder.
"""
@property
def cache_backend(self):
"""
Try to access to ``comparison`` cache value,
if fail use the ``default`` cache backend config.
"""
try:
comparison_cache = caches['comparison']
except InvalidCacheBackendError:
comparison_cache = caches['default']
return comparison_cache
@property
def cache_key(self):
"""
Key for the cache.
"""
return self.__class__.__name__
def get_cache(self):
"""
Get the cache from cache.
"""
return self.cache_backend.get(self.cache_key, {})
def set_cache(self, value):
"""
Assign the cache in cache.
"""
value.update(self.cache)
return self.cache_backend.set(self.cache_key, value)
cache = property(get_cache, set_cache)
def cache_flush(self):
"""
Flush the cache for this instance.
"""
return self.cache_backend.delete(self.cache_key)
def get_related(self, instance, number):
"""
Implement high level cache system for get_related.
"""
cache = self.cache
cache_key = '%s:%s' % (instance.pk, number)
if cache_key not in cache:
related_objects = super(CachedModelVectorBuilder,
self).get_related(instance, number)
cache[cache_key] = related_objects
self.cache = cache
return cache[cache_key]
@property
def columns_dataset(self):
"""
Implement high level cache system for columns and dataset.
"""
cache = self.cache
cache_key = 'columns_dataset'
if cache_key not in cache:
columns_dataset = super(CachedModelVectorBuilder, self
).columns_dataset
cache[cache_key] = columns_dataset
self.cache = cache
return cache[cache_key]
class EntryPublishedVectorBuilder(CachedModelVectorBuilder):
"""
Vector builder for published entries.
"""
limit = 100
queryset = Entry.published
fields = COMPARISON_FIELDS
@property
def cache_key(self):
"""
Key for the cache handling current site.
"""
return '%s:%s' % (super(EntryPublishedVectorBuilder, self).cache_key,
Site.objects.get_current().pk)