ferret 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/MIT-LICENSE +20 -0
- data/README +109 -0
- data/Rakefile +275 -0
- data/TODO +9 -0
- data/TUTORIAL +197 -0
- data/ext/extconf.rb +3 -0
- data/ext/ferret.c +23 -0
- data/ext/ferret.h +85 -0
- data/ext/index_io.c +543 -0
- data/ext/priority_queue.c +227 -0
- data/ext/ram_directory.c +316 -0
- data/ext/segment_merge_queue.c +41 -0
- data/ext/string_helper.c +42 -0
- data/ext/tags +240 -0
- data/ext/term.c +261 -0
- data/ext/term_buffer.c +299 -0
- data/ext/util.c +12 -0
- data/lib/ferret.rb +41 -0
- data/lib/ferret/analysis.rb +11 -0
- data/lib/ferret/analysis/analyzers.rb +93 -0
- data/lib/ferret/analysis/standard_tokenizer.rb +65 -0
- data/lib/ferret/analysis/token.rb +79 -0
- data/lib/ferret/analysis/token_filters.rb +86 -0
- data/lib/ferret/analysis/token_stream.rb +26 -0
- data/lib/ferret/analysis/tokenizers.rb +107 -0
- data/lib/ferret/analysis/word_list_loader.rb +27 -0
- data/lib/ferret/document.rb +2 -0
- data/lib/ferret/document/document.rb +152 -0
- data/lib/ferret/document/field.rb +304 -0
- data/lib/ferret/index.rb +26 -0
- data/lib/ferret/index/compound_file_io.rb +343 -0
- data/lib/ferret/index/document_writer.rb +288 -0
- data/lib/ferret/index/field_infos.rb +259 -0
- data/lib/ferret/index/fields_io.rb +175 -0
- data/lib/ferret/index/index.rb +228 -0
- data/lib/ferret/index/index_file_names.rb +33 -0
- data/lib/ferret/index/index_reader.rb +462 -0
- data/lib/ferret/index/index_writer.rb +488 -0
- data/lib/ferret/index/multi_reader.rb +363 -0
- data/lib/ferret/index/multiple_term_doc_pos_enum.rb +105 -0
- data/lib/ferret/index/segment_infos.rb +130 -0
- data/lib/ferret/index/segment_merge_info.rb +47 -0
- data/lib/ferret/index/segment_merge_queue.rb +16 -0
- data/lib/ferret/index/segment_merger.rb +337 -0
- data/lib/ferret/index/segment_reader.rb +380 -0
- data/lib/ferret/index/segment_term_enum.rb +178 -0
- data/lib/ferret/index/segment_term_vector.rb +58 -0
- data/lib/ferret/index/term.rb +49 -0
- data/lib/ferret/index/term_buffer.rb +88 -0
- data/lib/ferret/index/term_doc_enum.rb +283 -0
- data/lib/ferret/index/term_enum.rb +52 -0
- data/lib/ferret/index/term_info.rb +41 -0
- data/lib/ferret/index/term_infos_io.rb +312 -0
- data/lib/ferret/index/term_vector_offset_info.rb +20 -0
- data/lib/ferret/index/term_vectors_io.rb +552 -0
- data/lib/ferret/query_parser.rb +274 -0
- data/lib/ferret/query_parser/query_parser.tab.rb +819 -0
- data/lib/ferret/search.rb +49 -0
- data/lib/ferret/search/boolean_clause.rb +100 -0
- data/lib/ferret/search/boolean_query.rb +303 -0
- data/lib/ferret/search/boolean_scorer.rb +294 -0
- data/lib/ferret/search/caching_wrapper_filter.rb +40 -0
- data/lib/ferret/search/conjunction_scorer.rb +99 -0
- data/lib/ferret/search/disjunction_sum_scorer.rb +203 -0
- data/lib/ferret/search/exact_phrase_scorer.rb +32 -0
- data/lib/ferret/search/explanation.rb +41 -0
- data/lib/ferret/search/field_cache.rb +216 -0
- data/lib/ferret/search/field_doc.rb +31 -0
- data/lib/ferret/search/field_sorted_hit_queue.rb +184 -0
- data/lib/ferret/search/filter.rb +11 -0
- data/lib/ferret/search/filtered_query.rb +130 -0
- data/lib/ferret/search/filtered_term_enum.rb +79 -0
- data/lib/ferret/search/fuzzy_query.rb +153 -0
- data/lib/ferret/search/fuzzy_term_enum.rb +244 -0
- data/lib/ferret/search/hit_collector.rb +34 -0
- data/lib/ferret/search/hit_queue.rb +11 -0
- data/lib/ferret/search/index_searcher.rb +173 -0
- data/lib/ferret/search/match_all_docs_query.rb +104 -0
- data/lib/ferret/search/multi_phrase_query.rb +204 -0
- data/lib/ferret/search/multi_term_query.rb +65 -0
- data/lib/ferret/search/non_matching_scorer.rb +22 -0
- data/lib/ferret/search/phrase_positions.rb +55 -0
- data/lib/ferret/search/phrase_query.rb +217 -0
- data/lib/ferret/search/phrase_scorer.rb +153 -0
- data/lib/ferret/search/prefix_query.rb +47 -0
- data/lib/ferret/search/query.rb +111 -0
- data/lib/ferret/search/query_filter.rb +51 -0
- data/lib/ferret/search/range_filter.rb +103 -0
- data/lib/ferret/search/range_query.rb +139 -0
- data/lib/ferret/search/req_excl_scorer.rb +125 -0
- data/lib/ferret/search/req_opt_sum_scorer.rb +70 -0
- data/lib/ferret/search/score_doc.rb +38 -0
- data/lib/ferret/search/score_doc_comparator.rb +114 -0
- data/lib/ferret/search/scorer.rb +91 -0
- data/lib/ferret/search/similarity.rb +278 -0
- data/lib/ferret/search/sloppy_phrase_scorer.rb +47 -0
- data/lib/ferret/search/sort.rb +105 -0
- data/lib/ferret/search/sort_comparator.rb +60 -0
- data/lib/ferret/search/sort_field.rb +87 -0
- data/lib/ferret/search/spans.rb +12 -0
- data/lib/ferret/search/spans/near_spans_enum.rb +304 -0
- data/lib/ferret/search/spans/span_first_query.rb +79 -0
- data/lib/ferret/search/spans/span_near_query.rb +108 -0
- data/lib/ferret/search/spans/span_not_query.rb +130 -0
- data/lib/ferret/search/spans/span_or_query.rb +176 -0
- data/lib/ferret/search/spans/span_query.rb +25 -0
- data/lib/ferret/search/spans/span_scorer.rb +74 -0
- data/lib/ferret/search/spans/span_term_query.rb +105 -0
- data/lib/ferret/search/spans/span_weight.rb +84 -0
- data/lib/ferret/search/spans/spans_enum.rb +44 -0
- data/lib/ferret/search/term_query.rb +128 -0
- data/lib/ferret/search/term_scorer.rb +181 -0
- data/lib/ferret/search/top_docs.rb +24 -0
- data/lib/ferret/search/top_field_docs.rb +17 -0
- data/lib/ferret/search/weight.rb +54 -0
- data/lib/ferret/search/wildcard_query.rb +26 -0
- data/lib/ferret/search/wildcard_term_enum.rb +61 -0
- data/lib/ferret/stemmers.rb +1 -0
- data/lib/ferret/stemmers/porter_stemmer.rb +218 -0
- data/lib/ferret/store.rb +5 -0
- data/lib/ferret/store/buffered_index_io.rb +191 -0
- data/lib/ferret/store/directory.rb +139 -0
- data/lib/ferret/store/fs_store.rb +338 -0
- data/lib/ferret/store/index_io.rb +259 -0
- data/lib/ferret/store/ram_store.rb +282 -0
- data/lib/ferret/utils.rb +7 -0
- data/lib/ferret/utils/bit_vector.rb +105 -0
- data/lib/ferret/utils/date_tools.rb +138 -0
- data/lib/ferret/utils/number_tools.rb +91 -0
- data/lib/ferret/utils/parameter.rb +41 -0
- data/lib/ferret/utils/priority_queue.rb +120 -0
- data/lib/ferret/utils/string_helper.rb +47 -0
- data/lib/ferret/utils/weak_key_hash.rb +51 -0
- data/rake_utils/code_statistics.rb +106 -0
- data/setup.rb +1551 -0
- data/test/benchmark/tb_ram_store.rb +76 -0
- data/test/benchmark/tb_rw_vint.rb +26 -0
- data/test/longrunning/tc_numbertools.rb +60 -0
- data/test/longrunning/tm_store.rb +19 -0
- data/test/test_all.rb +9 -0
- data/test/test_helper.rb +6 -0
- data/test/unit/analysis/tc_analyzer.rb +21 -0
- data/test/unit/analysis/tc_letter_tokenizer.rb +20 -0
- data/test/unit/analysis/tc_lower_case_filter.rb +20 -0
- data/test/unit/analysis/tc_lower_case_tokenizer.rb +27 -0
- data/test/unit/analysis/tc_per_field_analyzer_wrapper.rb +39 -0
- data/test/unit/analysis/tc_porter_stem_filter.rb +16 -0
- data/test/unit/analysis/tc_standard_analyzer.rb +20 -0
- data/test/unit/analysis/tc_standard_tokenizer.rb +20 -0
- data/test/unit/analysis/tc_stop_analyzer.rb +20 -0
- data/test/unit/analysis/tc_stop_filter.rb +14 -0
- data/test/unit/analysis/tc_white_space_analyzer.rb +21 -0
- data/test/unit/analysis/tc_white_space_tokenizer.rb +20 -0
- data/test/unit/analysis/tc_word_list_loader.rb +32 -0
- data/test/unit/document/tc_document.rb +47 -0
- data/test/unit/document/tc_field.rb +80 -0
- data/test/unit/index/tc_compound_file_io.rb +107 -0
- data/test/unit/index/tc_field_infos.rb +119 -0
- data/test/unit/index/tc_fields_io.rb +167 -0
- data/test/unit/index/tc_index.rb +140 -0
- data/test/unit/index/tc_index_reader.rb +622 -0
- data/test/unit/index/tc_index_writer.rb +57 -0
- data/test/unit/index/tc_multiple_term_doc_pos_enum.rb +80 -0
- data/test/unit/index/tc_segment_infos.rb +74 -0
- data/test/unit/index/tc_segment_term_docs.rb +17 -0
- data/test/unit/index/tc_segment_term_enum.rb +60 -0
- data/test/unit/index/tc_segment_term_vector.rb +71 -0
- data/test/unit/index/tc_term.rb +22 -0
- data/test/unit/index/tc_term_buffer.rb +57 -0
- data/test/unit/index/tc_term_info.rb +19 -0
- data/test/unit/index/tc_term_infos_io.rb +192 -0
- data/test/unit/index/tc_term_vector_offset_info.rb +18 -0
- data/test/unit/index/tc_term_vectors_io.rb +108 -0
- data/test/unit/index/th_doc.rb +244 -0
- data/test/unit/query_parser/tc_query_parser.rb +84 -0
- data/test/unit/search/tc_filter.rb +113 -0
- data/test/unit/search/tc_fuzzy_query.rb +136 -0
- data/test/unit/search/tc_index_searcher.rb +188 -0
- data/test/unit/search/tc_search_and_sort.rb +98 -0
- data/test/unit/search/tc_similarity.rb +37 -0
- data/test/unit/search/tc_sort.rb +48 -0
- data/test/unit/search/tc_sort_field.rb +27 -0
- data/test/unit/search/tc_spans.rb +153 -0
- data/test/unit/store/tc_fs_store.rb +84 -0
- data/test/unit/store/tc_ram_store.rb +35 -0
- data/test/unit/store/tm_store.rb +180 -0
- data/test/unit/store/tm_store_lock.rb +68 -0
- data/test/unit/ts_analysis.rb +16 -0
- data/test/unit/ts_document.rb +4 -0
- data/test/unit/ts_index.rb +18 -0
- data/test/unit/ts_query_parser.rb +3 -0
- data/test/unit/ts_search.rb +10 -0
- data/test/unit/ts_store.rb +6 -0
- data/test/unit/ts_utils.rb +10 -0
- data/test/unit/utils/tc_bit_vector.rb +65 -0
- data/test/unit/utils/tc_date_tools.rb +50 -0
- data/test/unit/utils/tc_number_tools.rb +59 -0
- data/test/unit/utils/tc_parameter.rb +40 -0
- data/test/unit/utils/tc_priority_queue.rb +62 -0
- data/test/unit/utils/tc_string_helper.rb +21 -0
- data/test/unit/utils/tc_weak_key_hash.rb +25 -0
- metadata +251 -0
@@ -0,0 +1,34 @@
|
|
1
|
+
module Ferret::Search
|
2
|
+
# Lower-level search API.
|
3
|
+
#
|
4
|
+
# HitCollectors are primarily meant to be used to implement queries, sorting
|
5
|
+
# and filtering.
|
6
|
+
#
|
7
|
+
# See Searcher#search(Query, HitCollector)
|
8
|
+
class HitCollector
|
9
|
+
# Called once for every non-zero scoring document, with the document number
|
10
|
+
# and its score.
|
11
|
+
#
|
12
|
+
# If, for example, an application wished to collect all of the hits for a
|
13
|
+
# query in a BitSet, then it might:
|
14
|
+
#
|
15
|
+
# searcher = IndexSearcher.new(index_reader)
|
16
|
+
# bits = BitSet.new(index_reader.max_doc())
|
17
|
+
# searcher.search(query, HitCollector.new()
|
18
|
+
# def collect(doc, score)
|
19
|
+
# bits.set(doc)
|
20
|
+
# end
|
21
|
+
# end
|
22
|
+
#
|
23
|
+
# NOTE: This is called in an inner search loop. For good search
|
24
|
+
# performance, implementations of this method should not call
|
25
|
+
# Searcher#doc(int) or IndexReader#document(int) on every document number
|
26
|
+
# encountered. Doing so can slow searches by an order of magnitude or more.
|
27
|
+
#
|
28
|
+
# NOTE: The +score+ passed to this method is a raw score. In other words,
|
29
|
+
# the score will not necessarily be a float whose value is between 0 and 1.
|
30
|
+
def collect(doc, score)
|
31
|
+
raise NotImplementedError
|
32
|
+
end
|
33
|
+
end
|
34
|
+
end
|
@@ -0,0 +1,173 @@
|
|
1
|
+
module Ferret::Search
|
2
|
+
|
3
|
+
# Implements search over a single IndexReader.
|
4
|
+
#
|
5
|
+
# Applications usually need only call the inherited @link #search(Query)end
|
6
|
+
# or @link #search(Query,Filter)endmethods. For performance reasons it is
|
7
|
+
# recommended to open only one IndexSearcher and use it for all of your searches.
|
8
|
+
class IndexSearcher
|
9
|
+
include Ferret::Index
|
10
|
+
|
11
|
+
attr_accessor :similarity, :reader
|
12
|
+
|
13
|
+
# Creates a searcher searching the index in the provided directory.
|
14
|
+
def initialize(arg)
|
15
|
+
if arg.is_a?(IndexReader)
|
16
|
+
@reader = arg
|
17
|
+
elsif arg.is_a?(Ferret::Store::Directory)
|
18
|
+
@reader = IndexReader.open(arg)
|
19
|
+
elsif arg.is_a?(String)
|
20
|
+
@dir = Ferret::Store::FSDirectory.open(arg)
|
21
|
+
@reader = IndexReader.open(@dir, true)
|
22
|
+
else
|
23
|
+
raise ArgumentError, "Unknown argument passed to initialize IndexReader"
|
24
|
+
end
|
25
|
+
|
26
|
+
@similarity = Similarity.default
|
27
|
+
end
|
28
|
+
|
29
|
+
# IndexSearcher was constructed with IndexSearcher(r).
|
30
|
+
# If the IndexReader was supplied implicitly by specifying a directory, then
|
31
|
+
# the IndexReader gets closed.
|
32
|
+
def close()
|
33
|
+
@reader.close()
|
34
|
+
end
|
35
|
+
|
36
|
+
# Expert: Returns the number of documents containing +term+.
|
37
|
+
# Called by search code to compute term weights.
|
38
|
+
# See IndexReader#doc_freq
|
39
|
+
def doc_freq(term)
|
40
|
+
return @reader.doc_freq(term)
|
41
|
+
end
|
42
|
+
|
43
|
+
# Expert: For each term in the terms array, calculates the number of
|
44
|
+
# documents containing +term+. Returns an array with these
|
45
|
+
# document frequencies. Used to minimize number of remote calls.
|
46
|
+
def doc_freqs(terms)
|
47
|
+
result = Array.new(terms.length)
|
48
|
+
terms.each_with_index {|term, i| result[i] = doc_freq(term)}
|
49
|
+
return result
|
50
|
+
end
|
51
|
+
|
52
|
+
# Expert: Returns the stored fields of document +i+.
|
53
|
+
# Called by HitCollector implementations.
|
54
|
+
# See IndexReader#get_document
|
55
|
+
def doc(i)
|
56
|
+
return @reader.document(i)
|
57
|
+
end
|
58
|
+
|
59
|
+
# Expert: Returns one greater than the largest possible document number.
|
60
|
+
# Called by search code to compute term weights.
|
61
|
+
# See IndexReader#max_doc
|
62
|
+
def max_doc()
|
63
|
+
return @reader.max_doc()
|
64
|
+
end
|
65
|
+
|
66
|
+
# Creates a weight for +query+
|
67
|
+
# returns:: new weight
|
68
|
+
def create_weight(query)
|
69
|
+
return query.weight(self)
|
70
|
+
end
|
71
|
+
|
72
|
+
# The main search method for the index. You need to create a query to
|
73
|
+
# pass to this method. You can also pass a hash with one or more of the
|
74
|
+
# following; {filter, num_docs, first_doc, sort}
|
75
|
+
#
|
76
|
+
# query:: the query to run on the index
|
77
|
+
# filter:: filters docs from the search result
|
78
|
+
# first_doc:: The index in the results of the first doc retrieved.
|
79
|
+
# Default is 0
|
80
|
+
# num_docs:: The number of results returned. Default is 10
|
81
|
+
# sort:: an array of SortFields describing how to sort the results.
|
82
|
+
def search(query, options = {})
|
83
|
+
filter = options[:filter]
|
84
|
+
first_doc = options[:first_doc]||0
|
85
|
+
num_docs = options[:num_docs]||10
|
86
|
+
sort = options[:sort]
|
87
|
+
|
88
|
+
if (num_docs <= 0) # nil might be returned from hq.top() below.
|
89
|
+
raise ArgumentError, "num_docs must be > 0 to run a search"
|
90
|
+
end
|
91
|
+
|
92
|
+
scorer = query.weight(self).scorer(@reader)
|
93
|
+
if (scorer == nil)
|
94
|
+
return TopDocs.new(0, [])
|
95
|
+
end
|
96
|
+
|
97
|
+
bits = (filter.nil? ? nil : filter.bits(@reader))
|
98
|
+
if (sort)
|
99
|
+
fields = sort.is_a?(Array) ? sort : sort.fields
|
100
|
+
hq = FieldSortedHitQueue.new(@reader, fields, num_docs + first_doc)
|
101
|
+
else
|
102
|
+
hq = HitQueue.new(num_docs + first_doc)
|
103
|
+
end
|
104
|
+
total_hits = 0
|
105
|
+
min_score = 0.0
|
106
|
+
scorer.each_hit() do |doc, score|
|
107
|
+
if score > 0.0 and (bits.nil? or bits.get(doc)) # skip docs not in bits
|
108
|
+
total_hits += 1
|
109
|
+
if hq.size < num_docs or score >= min_score
|
110
|
+
hq.insert(ScoreDoc.new(doc, score))
|
111
|
+
min_score = hq.top.score # maintain min_score
|
112
|
+
end
|
113
|
+
end
|
114
|
+
end
|
115
|
+
|
116
|
+
score_docs = Array.new(hq.size)
|
117
|
+
if (hq.size > first_doc)
|
118
|
+
score_docs = Array.new(hq.size - first_doc)
|
119
|
+
first_doc.times { hq.pop }
|
120
|
+
(hq.size - 1).downto(0) do |i|
|
121
|
+
score_docs[i] = hq.pop
|
122
|
+
end
|
123
|
+
else
|
124
|
+
score_docs = []
|
125
|
+
hq.clear
|
126
|
+
end
|
127
|
+
|
128
|
+
return TopDocs.new(total_hits, score_docs)
|
129
|
+
end
|
130
|
+
|
131
|
+
# Accepts a block and iterates through all of results yielding the doc
|
132
|
+
# number and the score for that hit. The hits are unsorted. This is the
|
133
|
+
# fastest way to get all of the hits from a search. However, you will
|
134
|
+
# usually want your hits sorted at least by score so you should use the
|
135
|
+
# #search method.
|
136
|
+
def search_each(query, filter = nil)
|
137
|
+
scorer = query.weight(self).scorer(@reader)
|
138
|
+
return if scorer == nil
|
139
|
+
bits = (filter.nil? ? nil : filter.bits(@reader))
|
140
|
+
scorer.each_hit() do |doc, score|
|
141
|
+
if score > 0.0 and (bits.nil? or bits.get(doc)) # skip docs not in bits
|
142
|
+
yield(doc, score)
|
143
|
+
end
|
144
|
+
end
|
145
|
+
end
|
146
|
+
|
147
|
+
# rewrites the query into a query that can be processed by the search
|
148
|
+
# methods. For example, a Fuzzy query is turned into a massive boolean
|
149
|
+
# query.
|
150
|
+
#
|
151
|
+
# original:: The original query to be rewritten.
|
152
|
+
def rewrite(original)
|
153
|
+
query = original
|
154
|
+
rewritten_query = query.rewrite(@reader)
|
155
|
+
while query != rewritten_query
|
156
|
+
query = rewritten_query
|
157
|
+
rewritten_query = query.rewrite(@reader)
|
158
|
+
end
|
159
|
+
return query
|
160
|
+
end
|
161
|
+
|
162
|
+
# Returns an Explanation that describes how +doc+ scored against
|
163
|
+
# +query+.
|
164
|
+
#
|
165
|
+
# This is intended to be used in developing Similarity implementations,
|
166
|
+
# and, for good performance, should not be displayed with every hit.
|
167
|
+
# Computing an explanation is as expensive as executing the query over the
|
168
|
+
# entire index.
|
169
|
+
def explain(query, doc)
|
170
|
+
return query.weight(self).explain(@reader, doc)
|
171
|
+
end
|
172
|
+
end
|
173
|
+
end
|
@@ -0,0 +1,104 @@
|
|
1
|
+
module Ferret::Search
|
2
|
+
# A query that matches all documents.
|
3
|
+
class MatchAllDocsQuery < Query
|
4
|
+
|
5
|
+
def initialize()
|
6
|
+
super
|
7
|
+
end
|
8
|
+
|
9
|
+
class MatchAllScorer < Scorer
|
10
|
+
|
11
|
+
def initialize(reader, similarity)
|
12
|
+
super(similarity)
|
13
|
+
@reader = reader
|
14
|
+
@count = -1
|
15
|
+
@max_doc = reader.max_doc
|
16
|
+
end
|
17
|
+
|
18
|
+
def doc()
|
19
|
+
return @count
|
20
|
+
end
|
21
|
+
|
22
|
+
def explain(doc)
|
23
|
+
return Explanation.new(1.0, "MatchAllDocsQuery")
|
24
|
+
end
|
25
|
+
|
26
|
+
def next?
|
27
|
+
while (@count < (@max_doc - 1))
|
28
|
+
@count += 1
|
29
|
+
if (!@reader.deleted?(@count))
|
30
|
+
return true
|
31
|
+
end
|
32
|
+
end
|
33
|
+
return false
|
34
|
+
end
|
35
|
+
|
36
|
+
def score()
|
37
|
+
return 1.0
|
38
|
+
end
|
39
|
+
|
40
|
+
def skip_to(target)
|
41
|
+
@count = target - 1
|
42
|
+
return next?
|
43
|
+
end
|
44
|
+
end
|
45
|
+
|
46
|
+
class MatchAllDocsWeight < Weight
|
47
|
+
attr_reader :query
|
48
|
+
def initialize(query, searcher)
|
49
|
+
@query = query
|
50
|
+
@searcher = searcher
|
51
|
+
end
|
52
|
+
|
53
|
+
def to_s()
|
54
|
+
return "weight(#{@query})"
|
55
|
+
end
|
56
|
+
|
57
|
+
def value()
|
58
|
+
return 1.0
|
59
|
+
end
|
60
|
+
|
61
|
+
def sum_of_squared_weights()
|
62
|
+
return 1.0
|
63
|
+
end
|
64
|
+
|
65
|
+
def normalize(query_norm)
|
66
|
+
end
|
67
|
+
|
68
|
+
def scorer(reader)
|
69
|
+
return MatchAllScorer.new(reader, @query.similarity(@searcher))
|
70
|
+
end
|
71
|
+
|
72
|
+
def explain(reader, doc)
|
73
|
+
# explain query weight
|
74
|
+
query_expl = Explanation.new(1.0, "MatchAllDocsQuery")
|
75
|
+
boost_expl = Explanation.new(@query.boost, "boost")
|
76
|
+
if (boost_expl.value != 1.0)
|
77
|
+
query_expl << boost_expl
|
78
|
+
query_expl.value = boost_expl.value
|
79
|
+
end
|
80
|
+
|
81
|
+
return query_expl
|
82
|
+
end
|
83
|
+
end
|
84
|
+
|
85
|
+
def create_weight(searcher)
|
86
|
+
return MatchAllDocsWeight.new(self, searcher)
|
87
|
+
end
|
88
|
+
|
89
|
+
def to_s(field)
|
90
|
+
buffer = "MatchAllDocsQuery"
|
91
|
+
buffer << "^#{boost}" if (boost() != 1.0)
|
92
|
+
return buffer
|
93
|
+
end
|
94
|
+
|
95
|
+
def eql?(o)
|
96
|
+
return (o.instance_of?(MatchAllDocsQuery) and boost == o.boost)
|
97
|
+
end
|
98
|
+
alias :== :eql?
|
99
|
+
|
100
|
+
def hash
|
101
|
+
return boost.hash
|
102
|
+
end
|
103
|
+
end
|
104
|
+
end
|
@@ -0,0 +1,204 @@
|
|
1
|
+
module Ferret::Search
|
2
|
+
# MultiPhraseQuery is a generalized version of PhraseQuery, with an added
|
3
|
+
# method #add(Term[]).
|
4
|
+
#
|
5
|
+
# To use this class, to search for the phrase "Microsoft app*" first use
|
6
|
+
# add(Term) on the term "Microsoft", then find all terms that have "app" as
|
7
|
+
# prefix using IndexReader.terms(Term), and use MultiPhraseQuery.add(Term[]
|
8
|
+
# terms) to add them to the query.
|
9
|
+
#
|
10
|
+
# Author Anders Nielsen
|
11
|
+
class MultiPhraseQuery < Query
|
12
|
+
include Ferret::Index
|
13
|
+
|
14
|
+
attr_accessor :slop
|
15
|
+
attr_reader :positions, :term_arrays, :field
|
16
|
+
|
17
|
+
def initialize()
|
18
|
+
super()
|
19
|
+
@slop = 0
|
20
|
+
@term_arrays = []
|
21
|
+
@positions = []
|
22
|
+
@field = nil
|
23
|
+
end
|
24
|
+
|
25
|
+
# Allows to specify the relative position of terms within the phrase.
|
26
|
+
#
|
27
|
+
# See PhraseQuery#add(Term, int)
|
28
|
+
# terms:: the array of terms to search for or a single term
|
29
|
+
# position:: the position to search for these terms
|
30
|
+
def add(terms, position = nil, pos_inc = 1)
|
31
|
+
if position.nil?
|
32
|
+
position = (@positions.size > 0) ? (@positions[-1] + pos_inc) : 0
|
33
|
+
end
|
34
|
+
|
35
|
+
if terms.instance_of?(Term)
|
36
|
+
terms = [terms]
|
37
|
+
end
|
38
|
+
|
39
|
+
if (@term_arrays.size == 0)
|
40
|
+
@field = terms[0].field
|
41
|
+
end
|
42
|
+
|
43
|
+
terms.each do |term|
|
44
|
+
if (term.field != @field)
|
45
|
+
raise ArgumentError,
|
46
|
+
"All phrase terms must be in the same field (#{@field}): #{term}"
|
47
|
+
end
|
48
|
+
end
|
49
|
+
|
50
|
+
@term_arrays << terms
|
51
|
+
@positions << position
|
52
|
+
end
|
53
|
+
alias :<< :add
|
54
|
+
|
55
|
+
class MultiPhraseWeight < Weight
|
56
|
+
include Ferret::Index
|
57
|
+
|
58
|
+
attr_reader :query, :value
|
59
|
+
|
60
|
+
def initialize(query, searcher)
|
61
|
+
@query = query
|
62
|
+
@term_arrays = query.term_arrays
|
63
|
+
@positions = query.positions
|
64
|
+
@similarity = query.similarity(searcher)
|
65
|
+
@idf = 0.0
|
66
|
+
|
67
|
+
# compute idf
|
68
|
+
query.term_arrays.each do |terms|
|
69
|
+
terms.each do |term|
|
70
|
+
@idf += @similarity.idf_term(term, searcher)
|
71
|
+
end
|
72
|
+
end
|
73
|
+
end
|
74
|
+
|
75
|
+
def sum_of_squared_weights()
|
76
|
+
@query_weight = @idf * @query.boost() # compute query weight
|
77
|
+
return @query_weight * @query_weight # square it
|
78
|
+
end
|
79
|
+
|
80
|
+
def normalize(query_norm)
|
81
|
+
@query_norm = query_norm
|
82
|
+
@query_weight *= query_norm # normalize query weight
|
83
|
+
@value = @query_weight * @idf # idf for document
|
84
|
+
end
|
85
|
+
|
86
|
+
def scorer(reader)
|
87
|
+
return nil if (@term_arrays.size == 0) # optimize zero-term case
|
88
|
+
tps = []
|
89
|
+
@term_arrays.each do |terms|
|
90
|
+
p = []
|
91
|
+
if (terms.length > 1)
|
92
|
+
p = MultipleTermDocPosEnum.new(reader, terms)
|
93
|
+
else
|
94
|
+
p = reader.term_positions_for(terms[0])
|
95
|
+
end
|
96
|
+
|
97
|
+
return nil if (p == nil)
|
98
|
+
|
99
|
+
tps << p
|
100
|
+
end
|
101
|
+
|
102
|
+
if (@query.slop == 0)
|
103
|
+
return ExactPhraseScorer.new(self, tps, @positions, @similarity,
|
104
|
+
reader.get_norms(@query.field))
|
105
|
+
else
|
106
|
+
return SloppyPhraseScorer.new(self, tps, @positions, @similarity,
|
107
|
+
@query.slop, reader.get_norms(@query.field))
|
108
|
+
end
|
109
|
+
end
|
110
|
+
|
111
|
+
def explain(reader, doc)
|
112
|
+
|
113
|
+
result = Explanation.new()
|
114
|
+
result.description = "weight(#{@query} in #{doc}), product of:"
|
115
|
+
|
116
|
+
idf_expl = Explanation.new(@idf, "idf(#{@query})")
|
117
|
+
|
118
|
+
# explain query weight
|
119
|
+
query_expl = Explanation.new()
|
120
|
+
query_expl.description = "query_weight(#{@query}), product of:"
|
121
|
+
|
122
|
+
boost_expl = Explanation.new(@query.boost(), "boost")
|
123
|
+
(query_expl << boost_expl) if (@query.boost() != 1.0)
|
124
|
+
|
125
|
+
query_expl << idf_expl
|
126
|
+
|
127
|
+
query_norm_expl = Explanation.new(@query_norm,"query_norm")
|
128
|
+
query_expl << query_norm_expl
|
129
|
+
|
130
|
+
query_expl.value =
|
131
|
+
boost_expl.value * idf_expl.value * query_norm_expl.value
|
132
|
+
|
133
|
+
result << query_expl
|
134
|
+
|
135
|
+
# explain field weight
|
136
|
+
field_expl = Explanation.new()
|
137
|
+
field_expl.description =
|
138
|
+
"field_weight(#{@query} in #{doc}), product of:"
|
139
|
+
|
140
|
+
tf_expl = scorer(reader).explain(doc)
|
141
|
+
field_expl << tf_expl
|
142
|
+
field_expl << idf_expl
|
143
|
+
|
144
|
+
field_norm_expl = Explanation.new()
|
145
|
+
field_norms = reader.get_norms(@query.field)
|
146
|
+
field_norm =
|
147
|
+
field_norms ? Similarity.decode_norm(field_norms[doc]) : 0.0
|
148
|
+
field_norm_expl.value = field_norm
|
149
|
+
field_norm_expl.description =
|
150
|
+
"field_norm(field=#{@query.field}, doc=#{doc})"
|
151
|
+
field_expl << field_norm_expl
|
152
|
+
|
153
|
+
field_expl.value = tf_expl.value * idf_expl.value * field_norm_expl.value
|
154
|
+
result << field_expl
|
155
|
+
|
156
|
+
if (query_expl.value == 1.0)
|
157
|
+
return field_expl
|
158
|
+
else
|
159
|
+
result.value = query_expl.value * field_expl.value
|
160
|
+
return result
|
161
|
+
end
|
162
|
+
end
|
163
|
+
end
|
164
|
+
|
165
|
+
def rewrite(reader)
|
166
|
+
if (@term_arrays.size() == 1) # optimize one-term case
|
167
|
+
terms = @term_arrays[0]
|
168
|
+
bq = BooleanQuery.new(true)
|
169
|
+
terms.each do |term|
|
170
|
+
bq.add(TermQuery.new(term), BooleanClause::Occur::SHOULD)
|
171
|
+
end
|
172
|
+
bq.boost = boost()
|
173
|
+
return boq
|
174
|
+
else
|
175
|
+
return self
|
176
|
+
end
|
177
|
+
end
|
178
|
+
|
179
|
+
def create_weight(searcher)
|
180
|
+
return MultiPhraseWeight.new(self, searcher)
|
181
|
+
end
|
182
|
+
|
183
|
+
# Prints a user-readable version of this query.
|
184
|
+
def to_s(f = nil)
|
185
|
+
buffer = ""
|
186
|
+
buffer << "#{@field}:" if @field != f
|
187
|
+
buffer << '"'
|
188
|
+
last_pos = -1
|
189
|
+
@term_arrays.each_index do |i|
|
190
|
+
terms = @term_arrays[i]
|
191
|
+
pos = @positions[i]
|
192
|
+
last_pos.upto(pos-2) {buffer << "<> "}
|
193
|
+
last_pos = pos
|
194
|
+
buffer << "#{terms.map {|term| term.text}.join("|")} "
|
195
|
+
end
|
196
|
+
buffer.rstrip!
|
197
|
+
buffer << '"'
|
198
|
+
|
199
|
+
buffer << "~#{@slop}" if (@slop != 0)
|
200
|
+
buffer << "^#{boost()}" if boost() != 1.0
|
201
|
+
return buffer
|
202
|
+
end
|
203
|
+
end
|
204
|
+
end
|