-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_queries.py
262 lines (216 loc) · 8.2 KB
/
test_queries.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
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
# Wikipedia files used: AA - wiki_10 to wiki_29 (20 files)
# Group 3
# Lakshya Agarwal 2017B5A70904P
# Anuj Hydrabadi 2017A8PS0420P
# Samarth Gupta 2017B4A70467P
# Raj Shree Singh 2017B4A70808P
# Aditya Vishwakarma 2017B5A70954P
import pandas as pd
import numpy as np
import re
import pickle
import nltk
import os
import timeit
import string
import math
from nltk.tokenize import word_tokenize
from bs4 import BeautifulSoup
from collections import Counter
from collections import defaultdict
from nltk.corpus import wordnet
IDX_FILE = 'index.pickle'
MODIFIED_IDX_FILE = 'modified_index.pickle'
# One time download
nltk.download('stopwords')
nltk.download('punkt')
nltk.download('wordnet')
_trm_doc_wgt = None
_idf = None
_doc_id_ttl_map = None
_chmp_lst = None
_trm_ttl_wgt = None
_title_idf = None
query = None
_modified_trm_doc_wgt = None
_modified_idf = None
_modified_doc_id_ttl_map = None
_modified_chmp_lst = None
def load_index():
"""
Loads both normal index and modified index as saved by build_index.py
and assigns them as per requirements.
"""
global _trm_doc_wgt
global _idf
global _trm_ttl_wgt
global _chmp_lst
global _doc_id_ttl_map
global _title_idf
global _title_trm_freq
global _trm_freq
global _modified_trm_doc_wgt
global _modified_doc_id_ttl_map
global _modified_idf
global _modified_chmp_lst
with open(IDX_FILE, 'rb') as f:
ldd_obj = pickle.load(f)
_idf = ldd_obj['idf']
_title_idf = ldd_obj['title_idf']
_chmp_lst = ldd_obj['chmp_lst']
_trm_doc_wgt = ldd_obj['trm_doc_wgt']
_doc_id_ttl_map = ldd_obj['doc_id_ttl_map']
_title_trm_freq=ldd_obj['title_trm_freq']
_trm_freq=ldd_obj['trm_freq']
_trm_ttl_wgt = ldd_obj['trm_ttl_wgt']
with open(MODIFIED_IDX_FILE, 'rb') as f:
ldd_obj = pickle.load(f)
_modified_idf = ldd_obj['idf']
_modified_chmp_lst = ldd_obj['chmp_lst']
_modified_trm_doc_wgt = ldd_obj['trm_doc_wgt']
_modified_doc_id_ttl_map = ldd_obj['doc_id_ttl_map']
_title_idf = ldd_obj['title_idf']
_title_trm_freq=ldd_obj['title_trm_freq']
_trm_ttl_wgt = ldd_obj['trm_ttl_wgt']
def get_term_weights_for_query(query, idf, qry_typ = 0):
"""
This function tokenizes the query considering removal of the punctuations.
For queries we are using ltc model. ‘l’ represents term frequency is
considered using a logarithmic model that is the value of 1 plus logarithm
of frequency of occurrence of the term in the query. ‘t’ represents
the inverse document frequency which is already created in
get_term_document_weights() function. ‘c’ is cosine normalisation.
"""
raw_query_tokens = [token for token in word_tokenize(query) if token not in string.punctuation]
query_terms = [token.lower() for token in raw_query_tokens] #as a list for multiple occurence of same word
print(query_terms)
if (qry_typ == 1 or qry_typ == 4):
synList=[]
for q in query_terms:
synsets = wordnet.synsets(q)
for synset in synsets:
for syn in synset.lemma_names():
syn=syn.replace("_"," ")
synList.append(syn)
# print(synList)
# query_terms=synList
trm_qry_trm=[]
for w in synList:
if w in _trm_freq and w not in trm_qry_trm:
trm_qry_trm.append(w)
for q in query_terms:
if q in _trm_freq and q not in trm_qry_trm:
trm_qry_trm.append(q)
query_terms=trm_qry_trm
print("Query terms after including synonyms: ")
print(query_terms)
# print(query_terms)
result = defaultdict(int)
for term in query_terms:
result[term] += 1
for term in result:
result[term] = (1 + math.log10(result[term])) * (idf.get(term, 0))#tf_idf
vctr_len = 0
for weight in result.values():
vctr_len += weight ** 2
vctr_len = math.sqrt(vctr_len)#for normalisation of query
if vctr_len == 0:
vctr_len = 1 # To avoid Divide by Zero errors
# normalization
for term in result:
result[term] /= vctr_len
return query_terms, result
def search(qry_typ = 0):
"""
Calculates cosine similarity score on the basis of query types entered by
the user.
"""
k = int(input("\nEnter No. of documents to retrieve: "))
if (qry_typ == 0 or qry_typ == 1 or qry_typ == 3):
idf = _idf
trm_doc_wgt = _trm_doc_wgt
chmp_lst = _chmp_lst
doc_id_ttl_map = _doc_id_ttl_map
trm_ttl_wgt = _trm_ttl_wgt
elif (qry_typ == 2 or qry_typ == 4):
idf = _modified_idf
chmp_lst = _modified_chmp_lst
trm_doc_wgt = _modified_trm_doc_wgt
doc_id_ttl_map = _modified_doc_id_ttl_map
trm_freq = _trm_freq
trm_ttl_wgt = _trm_ttl_wgt
query_terms, trm_qry_wgts = get_term_weights_for_query(query, idf, qry_typ)
doc_id_score = defaultdict(int)
if (qry_typ == 2 or qry_typ == 4):
chmp_dcs = set()
for word in query_terms:
if word in _trm_freq:
chmp_dcs = chmp_dcs.union(set(chmp_lst[word]))
start1 = timeit.default_timer()
for term, qry_wgt in trm_qry_wgts.items():
for doc_id, doc_weight in trm_doc_wgt[term].items():
if (qry_typ == 2 or qry_typ == 4) and doc_id not in chmp_dcs: continue
doc_id_score[doc_id] += qry_wgt * doc_weight
stop1 = timeit.default_timer()
start2=0
stop2=0
if (qry_typ == 3 or qry_typ == 4):
# Zone indexing
for doc_id, val in doc_id_score.items():
doc_id_score[doc_id] *=0.8
start2 = timeit.default_timer()
for term, qry_wgt in trm_qry_wgts.items():
for doc_id, title_weight in trm_ttl_wgt[term].items():
if (qry_typ == 2 or qry_typ == 4) and doc_id not in chmp_dcs: continue
doc_id_score[doc_id] += qry_wgt * title_weight * 0.2
stop2 = timeit.default_timer()
print('Time: ', stop1 - start1 + stop2 - start2)
doc_scr_pr = [(doc_id, score) for doc_id, score in doc_id_score.items()]
doc_scr_pr.sort(key=lambda pair: pair[1], reverse=True)
k = min(k, len(doc_scr_pr))
print("\nNumber of retrieved documents:",min(k,len(doc_scr_pr)))
print("\n")
#df=pd.DataFrame(columns = ['Query', 'Top k documents', 'Score', 'Is the document relevant to the query?'])
for pair in doc_scr_pr[:k]:
doc_id = int(pair[0])
title = doc_id_ttl_map[str(doc_id)]
score = pair[1]
print("score = {0:.4f}, document_id = {1}, title = {2}".format(score, doc_id, title))
#df2 = {'Query': query, 'Top k documents': title, 'Score' : score}
#df = df.append(df2, ignore_index = True)
#df.to_csv('/content/drive/MyDrive/IRProject/'+query[:5]+'.xlsx', index=False, encoding='utf-8-sig')
def run_repl():
"""
Facilitates Main menu functioning. Takes input query from the user in
global variable ‘query’ and a choice from available options.
"""
global query
print('\nOptions:')
print('1. Normal search based on lnc.ltc scoring scheme')
print('2. Improvement using synonyms only')
print('3. Improvement using champion list')
print('4. Improvement using zone indexing')
print('5. Improvement using synonyms, championship list and zone indexing')
while True:
print('\nMain Menu\n')
query = input('Enter the query: ')
qry_typ = int(input('\nEnter your choice (either 1, 2, 3, 4 or 5): '))
if qry_typ>5 or qry_typ<1:
print('\nWrong choice, terminating... :-|')
return
qry_typ = qry_typ - 1
search(qry_typ)
again = int(input('\nEnter 1 to go back to main menu else 0 to quit: '))
if again!=1:
print('\nTerminating...')
return
def main():
choice =int(input('\nLoad index file from default location enter 0 else enter 1: '))
if(choice==1):
IDX_FILE = input('Enter path to basic index file: ')
MODIFIED_IDX_FILE = input('Enter path to modified index file: ')
print('\nLoading indices....')
load_index()
run_repl()
if __name__ == '__main__':
main()