context.rb
10.7 KB
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
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
#!/usr/bin/env ruby
require 'mirimiri'
require 'sanitize'
require 'lda-ruby'
require 'context/conceptual_element'
require 'context/concept_model'
require 'context/concept'
require 'context/query_context'
module Context
@@count = Hash.new { |h,k| h[k] = {} }
@@df = Hash.new { |h,k| h[k] = {} }
@@semaphore = Mutex.new
IndexPaths = {
:web_en => '/mnt/disk2/ClueWeb09_English_1_sDocs',
:web_fr => '/mnt/disk2/ClueWeb09_French_1_sDocs',
:web_nospam => '/mnt/disk1/ClueWeb09_English_1noSpam',
:robust => '/mnt/disk5/Robust04/',
:wt10g => '/mnt/disk3/WT10g_index',
:gov2 => '/mnt/disk3/GOV2_index',
:gigaword => '/local/data/GigaWord/index',
:nyt => '/local/data/NYT_index',
:wiki_en => '/local/data/WikiEn_index',
:wiki_en2012 => '/local/data/WikiEn2012_index',
:wiki_fr => '/local/data/WikiFr_index',
:wiki_tc2012 => '/local/data/INEXQA2012index',
:books => '/local/data/INEX/Books2011/indexedit',
:ent => '/home/sanjuan/works/nist_eval/csiro_indri.ind'
}
IndexPathsCaracole = {
:web_en => '/distant/index_clueweb/disk2/ClueWeb09_English_1_sDocs',
:web_nospam => '/distant/index_clueweb/disk1/ClueWeb09_English_1noSpam',
:robust => '/distant/data/Robust04',
:wt10g => '/distant/index_clueweb/disk3/WT10g_index',
:gov2 => '/distant/index_clueweb/disk3/GOV2_index'
}
# #
# From the SIGKDD 2007 paper : "Exploiting underrepresented query aspects for automatic query expansion"
def Context.query_aspects q
query = Mirimiri::Document.new q
2.upto(query.words.count) do |size|
query.ngrams(size).each do |s|
dp = Context.count_w Context::IndexPaths[:wiki_en2012],"#1(#{s})"
d = Context.count_w Context::IndexPaths[:wiki_en2012],"#{s}",100000000
p s
denum = s.split.permutation.inject(0.0) do |res,p|
tmp = (p == s.split) ? 0 : Context.count_w(Context::IndexPaths[:wiki_en2012],"#1(#{p.join(" ")})")
res + tmp
end
existence = dp.to_f/d
support = dp.to_f/denum
puts "#{s} ===> #{existence*support}"
end
end
end
# #
# From the CIKM 2007 paper : "Ranking Very Many Typed Entities on Wikipedia"
#
# The ``entities`` parameter is currently an array of strings. Could be moved
# to an array of Entity objects.
def Context.entity_web_ranking query,entities,nbdocs=100,index='web_nospam'
q = Indri::IndriQuery.new({:query => "#combine ( #{query} )", :count => nbdocs},"-trecFormat=true")
indri_index = Indri::IndriIndex.new IndexPaths[index.to_sym]
docs = indri_index.runquery(q).force_encoding("ISO-8859-1").encode("UTF-8")
query_list = docs.split("\n").collect { |p| p.scan(/\d+ Q0 (.+) \d+ -\d+.\d+ .+/).first }.flatten
res = entities.pmap(15) do |e|
eq = Indri::IndriQuery.new({:query => "#combine ( #{e.gsub(/[^a-zA-Z0-9\s]/,'')} )", :count => nbdocs},"-trecFormat=true")
edocs = indri_index.runquery(eq).force_encoding("ISO-8859-1").encode("UTF-8")
e_list = edocs.split("\n").collect { |p| p.scan(/\d+ Q0 (.+) \d+ -\d+.\d+ .+/).first }.flatten
rels = e_list&query_list
ave_p = 1.upto(nbdocs).inject(0.0) do |sum,k|
p = (e_list.first(k)&rels).count.to_f/k
rel = rels.include?(e_list[k-1]) ? 1.0 : 0.0
sum + p*rel
end
{:name => e, :score => ave_p}
end
res.sort { |a,b| b[:score] <=> a[:score] }
end
def Context.query_entities query,nb_docs=10
sources = ['wiki_en2012']
# sources = ['wiki_en2012','web_nospam','nyt','gigaword']
# sources = ['web_fr']
sc = Hash.new { |h,k| h[k] = 0.0 }
sources.each do |source|
puts " == source : #{source}"
c = ConceptModel.new query,source,nb_docs
p c.query
c.concepts.each do |concept|
querys = concept.words[0,4].join " "
d1 = Context::label_candidate querys.sequential_dependence_model,'wiki_en'
d2 = Context::label_candidate querys.sequential_dependence_model,'wiki_en2012'
d3 = Mirimiri::WikipediaPage.search_wikipedia_titles querys
d1 = [] if d1.nil?
d2 = [] if d2.nil?
d3 = [] if d3.nil?
d = d2 & d3
labels = d.collect { |c| c.downcase.gsub(/[^\w\d]/,' ') }
p d
mins = -10000000
lab = nil
scores = labels.collect do |l|
s = concept.score_label l
if s > mins
mins = s
lab = l
end
sc[l] += s*(concept.coherence/c.total_coherence)
{ :label => l, :score => s }
end
print (concept.coherence/c.total_coherence).to_s+" <= "
p concept.elements.collect { |c| c.word }
end
end
sc.sort { |a,b| b[1] <=> a[1] }
end
def Context.label_candidate query,index,nb_candidates,rm3=false
# Mirimiri::WikipediaPage.search_wikipedia_titles query
args = rm3 ? "-fbDocs=20 -fbTerms=30 -fbMu=2000 -fbOrigWeight=0.7" : ""
q = Indri::IndriQuery.new({:query => query, :count => nb_candidates},"-printDocuments=true -trecFormat=true #{args}")
index = Indri::IndriIndex.new IndexPaths[index.to_sym]
docs = index.runquery q
docs = docs.force_encoding("ISO-8859-1").encode("UTF-8") if ['web_fr','web_en','web_nospam'].include? index
idocs = Indri::IndriPrintedDocuments.new(docs)
wiki_titles = idocs.extract_docs.collect do |d|
t = Nokogiri::HTML d
t.xpath('//title').text
end
wiki_titles
end
def Context.lcm query
source = 'nyt'
a = Time.now
qc = QueryContext.new(1.upto(20).collect do |nb_docs|
beg = Time.now
c = ConceptModel.new query,source,nb_docs
puts "#{nb_docs} ==> Time elapsed: #{Time.now-beg} seconds"
c
end)
puts "All concepts : #{Time.now-a} seconds"
model = qc.best_concept_model
puts "Total : #{Time.now-a} seconds"
model
end
def Context.term_context index_path,query,size,num_page,args={}
terms = self.term_concepts index_path,query,size,num_page,args
args[:window] ||= 1
# context = "#weight ( #{terms.compact.sort{ |a,b| b[0] <=> a[0]}[0,size].collect { |e| "#{e[0]} #1(#{e[1]})" }.join " "} ) " unless terms.empty?
context = "#weight ( #{terms.collect { |c| "#{"%.10f" % c[:score]} #uw#{args[:window]}(#{c[:concept]})" }.join " "} ) " unless terms.empty?
context
end
# From SIGIR'06 paper : `Improving the estimation of relevance models using large external corpora`
#
def Context.morm index_path,query,size,num_page
docs,scores,names = self.feedback_docs index_path,query,num_page
terms = []
docs.each do |d|
r = Mirimiri::Document.new d
tmp = self.extract_ngrams r,:tf,1
terms += tmp.compact.collect { |t| [t[0]*Math.exp(scores[docs.index(d)].to_f),t[1]] }
end
final = terms.compact.sort{ |a,b| b[0] <=> a[0]}[0,size].collect { |e| { :score => e[0], :concept => e[1] } }
context = "#weight ( #{final.collect { |c| "#{"%.10f" % c[:score]} #1(#{c[:concept]})" }.join " "} ) " unless terms.empty?
context
end
def Context.term_concepts index_path,query,size,num_page,args={}
args[:func] ||= :entropy
args[:window] ||= 1
docs = self.feedback_docs index_path,query,num_page
resource = Mirimiri::Document.new docs.join(' ')
terms = self.extract_ngrams resource,args[:func].to_sym,args[:window]
terms.compact.sort{ |a,b| b[0] <=> a[0]}[0,size].collect { |e| { :score => e[0], :concept => e[1] } }
end
def Context.sentence_similarity s1,s2,index_path
q = s1.is_a?(String) ? s1.split : s1
r = s2.is_a?(String) ? s2.split : s2
inter = q & r
s = (inter.count/q.count.to_f) * inter.inject(0.0) { |sum,w| sum + Math.log(Context.df index_path,w) }
s
end
private
def Context.df index_path,w,window=1
if @@count[index_path]["total#{window}"].nil?
total = `dumpindex #{index_path} s`.match(/total terms:\t(.*)/)[1].to_f-(window-1).to_f
@@semaphore.synchronize {
@@count[index_path]["total#{window}"] = total
}
end
if @@df[index_path]["#uw#{window}(#{w})"].nil?
nb = `dumpindex #{index_path} e "#uw#{window}(#{w})" | awk ' { arr[$1]=$0 } END { for ( key in arr ) { print arr[key] } } ' | wc -l`.chomp.split(':').last.to_f - 1
@@semaphore.synchronize {
@@df[index_path]["#uw#{window}(#{w})"] = nb+1.0
}
end
begin
d = @@count[index_path]["total#{window}"]/@@df[index_path]["#uw#{window}(#{w})"]
rescue
puts w
exit
end
d
end
def Context.prob_w index_path,w,window=1
if @@count[index_path]["total#{window}"].nil?
total = `dumpindex #{index_path} s`.match(/total terms:\t(.*)/)[1].to_f-(window-1).to_f
@@semaphore.synchronize {
@@count[index_path]["total#{window}"] = total+1.0
}
end
nb = self.count_w index_path,w,window
nb/@@count[index_path]["total#{window}"]
end
def Context.count_w index_path,w,window=1
if @@count[index_path]["##{window}(#{w})"].nil?
nb = `dumpindex #{index_path} x "##{window}(#{w})"`.chomp.split(':').last.to_f
@@semaphore.synchronize {
@@count[index_path]["##{window}(#{w})"] = ("%.15f" % nb).to_f+1.0
}
end
@@count[index_path]["##{window}(#{w})"]
end
public
def Context.extract_ngrams resource,func,n
raw_terms = 1.upto(n).collect { |i| resource.ngrams(i) }.flatten
# raw_terms = resource.ngrams(n).flatten
terms = raw_terms.uniq.collect { |w| [resource.send(func.to_sym,w), w.unaccent] unless w.is_stopword? || w.split.any? { |e| e.length <= 2 } || w.split.any? { |e| e !~ /[a-zA-Z]/ } || w.include?(".") || (Mirimiri::Stoplist&w.unaccent.split).count >= 2 }
# terms = raw_terms.uniq.collect { |w| w=w.gsub(/\W/,' ').strip; [resource.send(func.to_sym,w), w.unaccent] unless w.is_stopword? || w.split.any? { |e| e.length <= 2 } || w.split.any? { |e| e !~ /[a-zA-Z]/ } || w.include?(".") || (Mirimiri::Stoplist&w.unaccent.split).count >= 1 }
# terms = raw_terms.uniq.collect { |w| [resource.send(func.to_sym,w), w.unaccent] unless w.is_stopword? || w.split.all? { |e| e.length <= 1 } || w.split.all? { |e| e !~ /[a-zA-Z]/ } || w.include?(".") }
terms
end
def Context.feedback_docs index_path,query,num_page
query = Indri::IndriQuery.new({:query => query, :count => num_page},"-printDocuments=true -trecFormat=true")
index = Indri::IndriIndex.new index_path
idocs = Indri::IndriPrintedDocuments.new(index.runquery(query).force_encoding("ISO-8859-1").encode("UTF-8"))
texts,scores,names = idocs.extract_docs_score
docs = texts.collect do |idoc|
begin
Sanitize.clean idoc,:remove_contents => ['script','style']
rescue
d = Nokogiri::HTML(idoc)
d.xpath('//text()').text
end
end
return docs,scores,names
end
end