forked from suvodeep-pyne/gitbook
/
recommender.py
214 lines (172 loc) · 8.25 KB
/
recommender.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
'''
Created on Apr 26, 2013
@author: Suvodeep Pyne
Edited by Garima Agarwal
'''
import os
import cmd
import pickle
import pickle
import marshal
import operator
import pprint as pp
from page_rank import pagerank
from collections import defaultdict
from read_data import DataRetriever
from nb_classifier import NaiveBayesClassifier
from resource_manager import ResourceManager
rm = ResourceManager()
class ProjectVectorBuilder():
projects = {}
def __init__(self, project_data):
self.project_data = project_data
self.nb = NaiveBayesClassifier(rm.TRAINDATA_VOCAB, rm.TRAINDATA_DATASET)
self.nb.train()
def build_projects_vector(self):
print "In build projects"
for name, project in self.project_data.iteritems():
readme = project['readme']
# Bad case: When readme is not found. It returns empty lists.
if isinstance(readme, list):
readme = ""
else:
readme = unicode(readme, 'utf-8', errors = 'ignore')
if project['description'] != None:
readme += project['description']
if readme == "": continue
self.projects[name] = {}
prob_data = self.nb.classify(readme)[0]
self.projects[name]['class_prob'] = prob_data
self.projects[name]['description'] = project['description']
if len(prob_data) > 0:
self.projects[name]['category'] = max(prob_data.iteritems(), key=operator.itemgetter(1))[0]
self.projects[name]['prob'] = max(prob_data.iteritems(), key=operator.itemgetter(1))[1]
return self.projects
class Recommender():
"""Initialize the recommender"""
def __init__(self):
print 'Initializing Recommender..'
directory_name = rm.CACHE
self.data_retriever = DataRetriever(directory_name)
self.project_data = self.data_retriever.parseProjectData()
self.user_data, self.user_follower_map = self.data_retriever.parseUserFollowers()
self.language_proj = defaultdict()
def get_languages(self):
lang_dict = {}
for lang in self.language_proj.keys():
_lang = lang.replace(' ','$')
lang_dict[_lang] = lang
return lang_dict
def get_aoi(self):
return self.categories
"""Get different scores for each project"""
def build_project_features(self):
try:
with open(rm.NB_PROB, 'rb') as f:
print "Reading probabilities from:", rm.NB_PROB
self.project_vector = pickle.load(f)
self.categories = pickle.load(f)
print 'done.'
print '#Projects:', len(self.project_vector)
print '#Categories:', len(self.categories)
except:
print "Generating a new Naive Base classifier"
self.project_vector_builder = ProjectVectorBuilder(self.project_data)
self.project_vector = self.project_vector_builder.build_projects_vector()
self.categories = list(self.project_vector_builder.nb.clf.classes_)
with open(rm.NB_PROB, 'wb') as f:
pickle.dump(self.project_vector, f)
with open(rm.NB_PROB, 'ab') as f:
pickle.dump(self.categories, f)
self.user_ranking = pagerank(self.user_data)
with open(os.path.join(rm.CACHE, 'lang_to_projects.p'), 'rb') as f:
self.language_proj = pickle.load(f)
with open(os.path.join(rm.CACHE, 'new_LOC.p'),'rb') as f:
self.difficulty_score = pickle.load(f)
def recommend_projects(self, languages, area_interest, difficulty):
print "Calling recommender"
projects = set()
#Filter based on languages
for language in languages:
projects = projects.union(self.language_proj[language])
similar_projects = []
for project in projects:
if project not in self.project_vector: continue
if self.project_vector[project]['category'] in area_interest:
project_desc = self.project_vector[project]
project_desc['html_url'] = self.project_data[project]['html_url']
project_desc['full_name'] = self.project_data[project]['full_name']
similar_projects.append(project_desc)
sorted_similar_projects = sorted(similar_projects, key=lambda k: k['prob'], reverse=True)
#pp.pprint(sorted_similar_projects)
zipped = map(list, zip(*self.user_ranking))
userLists = zipped[0]
PRs = zipped[1]
sortedProjsLength = len(sorted_similar_projects)
for i in range(0,len(sorted_similar_projects)):
proj = sorted_similar_projects[i]
project = self.project_data[proj[u'full_name']]
owner = project[u'owner']
if owner[u'login'] in userLists:
userIndex = userLists.index(owner[u'login'])
sorted_similar_projects[i]['page_rank_of_owner'] = PRs[userIndex]
sorted_similar_projects[i]['owner'] = owner[u'login']
#sorted_similar_projects[i]['contributors'] = self.project_data[proj['full_name']]['contributors'][0]['login']
if len(self.project_data[proj['full_name']]['contributors']) >=1:
sorted_similar_projects[i]['contributors'] = self.project_data[proj['full_name']]['contributors'][0]['login']
sorted_similar_projects[i]['contributors_url'] = self.project_data[proj['full_name']]['contributors'][0]['html_url']
else:
sorted_similar_projects[i]['contributors'] = ''
sorted_similar_projects[i]['contributors_url'] =''
else:
sorted_similar_projects[i]['page_rank_of_owner'] = 0
sorted_similar_projects[i]['owner'] = owner[u'login']
if len(self.project_data[proj['full_name']]['contributors']) >=1:
sorted_similar_projects[i]['contributors'] = self.project_data[proj['full_name']]['contributors'][0]['login']
sorted_similar_projects[i]['contributors_url'] = self.project_data[proj['full_name']]['contributors'][0]['html_url']
else:
sorted_similar_projects[i]['contributors'] = ''
sorted_similar_projects[i]['contributors_url'] = ''
# sort the sorted_similar_projects based on the key 'page_rank_of_owner' value
# have the contributors tag with the first contributor for the server side handling
#"""
if len(sorted_similar_projects) > 10:
firstListToSort = sorted_similar_projects[0:sortedProjsLength/2]
secListToSort = sorted_similar_projects[sortedProjsLength/2 + 1 : sortedProjsLength*4/5 ]
thirListToSort = sorted_similar_projects[sortedProjsLength*4/5 + 1 : ]
#"""
sorted1 = sorted(firstListToSort, key=lambda k: k['page_rank_of_owner'], reverse=True)
sorted2 = sorted(secListToSort, key=lambda k: k['page_rank_of_owner'], reverse=True)
sorted3 = sorted(thirListToSort, key=lambda k: k['page_rank_of_owner'], reverse=True)
#print len(sorted1)
#print len(sorted2)
#print len(sorted3)
sorted1.extend(sorted2)
sorted3.extend(sorted1)
#pp.pprint(sorted3)
#print 'lenght after merging all: ',len(sorted3)
return sorted3
return sorted_similar_projects
class CommandLineInterface(cmd.Cmd):
def __init__(self):
cmd.Cmd.__init__(self)
self.obj = Recommender()
self.obj.build_project_features()
self.name = "garima"
print self.obj.get_aoi()
self.prompt = ">> "
def do_recommend_projects(self, args):
languages, aoi, level = args.split("\" \"")
languages = languages.replace("\"", "")
level = level.replace("\"", "")
self.obj.recommend_projects(languages.split(), [aoi], level)
def do_EOF(self, args):
return self.do_exit(args)
def do_exit(self, args):
"""Exit"""
return -1
def do_help(self, args):
cmd.Cmd.do_help(self, args)
if __name__ == '__main__':
cmi = CommandLineInterface()
cmi.cmdloop()