ferret 0.3.2 → 0.9.0
Sign up to get free protection for your applications and to get access to all the features.
- data/CHANGELOG +9 -0
- data/Rakefile +51 -25
- data/ext/analysis.c +553 -0
- data/ext/analysis.h +76 -0
- data/ext/array.c +83 -0
- data/ext/array.h +19 -0
- data/ext/bitvector.c +164 -0
- data/ext/bitvector.h +29 -0
- data/ext/compound_io.c +335 -0
- data/ext/document.c +336 -0
- data/ext/document.h +87 -0
- data/ext/ferret.c +88 -47
- data/ext/ferret.h +43 -109
- data/ext/field.c +395 -0
- data/ext/filter.c +103 -0
- data/ext/fs_store.c +352 -0
- data/ext/global.c +219 -0
- data/ext/global.h +73 -0
- data/ext/hash.c +446 -0
- data/ext/hash.h +80 -0
- data/ext/hashset.c +141 -0
- data/ext/hashset.h +37 -0
- data/ext/helper.c +11 -0
- data/ext/helper.h +5 -0
- data/ext/inc/lang.h +41 -0
- data/ext/ind.c +389 -0
- data/ext/index.h +884 -0
- data/ext/index_io.c +269 -415
- data/ext/index_rw.c +2543 -0
- data/ext/lang.c +31 -0
- data/ext/lang.h +41 -0
- data/ext/priorityqueue.c +228 -0
- data/ext/priorityqueue.h +44 -0
- data/ext/q_boolean.c +1331 -0
- data/ext/q_const_score.c +154 -0
- data/ext/q_fuzzy.c +287 -0
- data/ext/q_match_all.c +142 -0
- data/ext/q_multi_phrase.c +343 -0
- data/ext/q_parser.c +2180 -0
- data/ext/q_phrase.c +657 -0
- data/ext/q_prefix.c +75 -0
- data/ext/q_range.c +247 -0
- data/ext/q_span.c +1566 -0
- data/ext/q_term.c +308 -0
- data/ext/q_wildcard.c +146 -0
- data/ext/r_analysis.c +255 -0
- data/ext/r_doc.c +578 -0
- data/ext/r_index_io.c +996 -0
- data/ext/r_qparser.c +158 -0
- data/ext/r_search.c +2321 -0
- data/ext/r_store.c +263 -0
- data/ext/r_term.c +219 -0
- data/ext/ram_store.c +447 -0
- data/ext/search.c +524 -0
- data/ext/search.h +1065 -0
- data/ext/similarity.c +143 -39
- data/ext/sort.c +661 -0
- data/ext/store.c +35 -0
- data/ext/store.h +152 -0
- data/ext/term.c +704 -143
- data/ext/termdocs.c +599 -0
- data/ext/vector.c +594 -0
- data/lib/ferret.rb +9 -10
- data/lib/ferret/analysis/analyzers.rb +2 -2
- data/lib/ferret/analysis/standard_tokenizer.rb +1 -1
- data/lib/ferret/analysis/token.rb +14 -14
- data/lib/ferret/analysis/token_filters.rb +3 -3
- data/lib/ferret/document/field.rb +16 -17
- data/lib/ferret/index/document_writer.rb +4 -4
- data/lib/ferret/index/index.rb +39 -23
- data/lib/ferret/index/index_writer.rb +2 -2
- data/lib/ferret/index/multiple_term_doc_pos_enum.rb +1 -8
- data/lib/ferret/index/segment_term_vector.rb +4 -4
- data/lib/ferret/index/term.rb +5 -1
- data/lib/ferret/index/term_vector_offset_info.rb +6 -6
- data/lib/ferret/index/term_vectors_io.rb +5 -5
- data/lib/ferret/query_parser/query_parser.tab.rb +81 -77
- data/lib/ferret/search.rb +1 -1
- data/lib/ferret/search/boolean_query.rb +2 -1
- data/lib/ferret/search/field_sorted_hit_queue.rb +3 -3
- data/lib/ferret/search/fuzzy_query.rb +2 -1
- data/lib/ferret/search/index_searcher.rb +3 -0
- data/lib/ferret/search/{match_all_docs_query.rb → match_all_query.rb} +7 -7
- data/lib/ferret/search/multi_phrase_query.rb +6 -5
- data/lib/ferret/search/phrase_query.rb +3 -6
- data/lib/ferret/search/prefix_query.rb +4 -4
- data/lib/ferret/search/sort.rb +3 -1
- data/lib/ferret/search/sort_field.rb +9 -9
- data/lib/ferret/search/spans/near_spans_enum.rb +1 -1
- data/lib/ferret/search/spans/span_near_query.rb +1 -1
- data/lib/ferret/search/spans/span_weight.rb +1 -1
- data/lib/ferret/search/spans/spans_enum.rb +7 -7
- data/lib/ferret/store/fs_store.rb +10 -6
- data/lib/ferret/store/ram_store.rb +3 -3
- data/lib/rferret.rb +36 -0
- data/test/functional/thread_safety_index_test.rb +2 -2
- data/test/test_helper.rb +16 -2
- data/test/unit/analysis/c_token.rb +25 -0
- data/test/unit/analysis/tc_per_field_analyzer_wrapper.rb +1 -1
- data/test/unit/analysis/tc_standard_analyzer.rb +1 -1
- data/test/unit/document/{tc_document.rb → c_document.rb} +0 -0
- data/test/unit/document/c_field.rb +98 -0
- data/test/unit/document/tc_field.rb +0 -66
- data/test/unit/index/{tc_index.rb → c_index.rb} +62 -6
- data/test/unit/index/{tc_index_reader.rb → c_index_reader.rb} +51 -10
- data/test/unit/index/{tc_index_writer.rb → c_index_writer.rb} +0 -4
- data/test/unit/index/{tc_term.rb → c_term.rb} +1 -3
- data/test/unit/index/{tc_term_vector_offset_info.rb → c_term_voi.rb} +5 -5
- data/test/unit/index/tc_segment_term_vector.rb +2 -2
- data/test/unit/index/tc_term_vectors_io.rb +4 -4
- data/test/unit/query_parser/c_query_parser.rb +138 -0
- data/test/unit/search/{tc_filter.rb → c_filter.rb} +24 -24
- data/test/unit/search/{tc_fuzzy_query.rb → c_fuzzy_query.rb} +0 -0
- data/test/unit/search/{tc_index_searcher.rb → c_index_searcher.rb} +9 -26
- data/test/unit/search/{tc_search_and_sort.rb → c_search_and_sort.rb} +15 -15
- data/test/unit/search/{tc_sort.rb → c_sort.rb} +2 -1
- data/test/unit/search/c_sort_field.rb +27 -0
- data/test/unit/search/{tc_spans.rb → c_spans.rb} +0 -0
- data/test/unit/search/tc_sort_field.rb +7 -20
- data/test/unit/store/c_fs_store.rb +76 -0
- data/test/unit/store/c_ram_store.rb +35 -0
- data/test/unit/store/m_store.rb +34 -0
- data/test/unit/store/m_store_lock.rb +68 -0
- data/test/unit/store/tc_fs_store.rb +0 -53
- data/test/unit/store/tc_ram_store.rb +0 -20
- data/test/unit/store/tm_store.rb +0 -30
- data/test/unit/store/tm_store_lock.rb +0 -66
- metadata +84 -31
- data/ext/Makefile +0 -140
- data/ext/ferret_ext.so +0 -0
- data/ext/priority_queue.c +0 -232
- data/ext/ram_directory.c +0 -321
- data/ext/segment_merge_queue.c +0 -37
- data/ext/segment_term_enum.c +0 -326
- data/ext/string_helper.c +0 -42
- data/ext/tags +0 -344
- data/ext/term_buffer.c +0 -230
- data/ext/term_infos_reader.c +0 -54
- data/ext/terminfo.c +0 -160
- data/ext/token.c +0 -93
- data/ext/util.c +0 -12
data/ext/q_const_score.c
ADDED
@@ -0,0 +1,154 @@
|
|
1
|
+
#include "search.h"
|
2
|
+
#include <string.h>
|
3
|
+
|
4
|
+
/***************************************************************************
|
5
|
+
*
|
6
|
+
* Weight
|
7
|
+
*
|
8
|
+
***************************************************************************/
|
9
|
+
|
10
|
+
char *csw_to_s(Weight *self)
|
11
|
+
{
|
12
|
+
char dbuf[32];
|
13
|
+
dbl_to_s(dbuf, self->value);
|
14
|
+
return epstrdup("ConstantScoreWeight(%s)", strlen(dbuf), dbuf);
|
15
|
+
}
|
16
|
+
|
17
|
+
void csw_destroy(void *p)
|
18
|
+
{
|
19
|
+
free(p);
|
20
|
+
}
|
21
|
+
|
22
|
+
Explanation *csw_explain(Weight *self, IndexReader *ir, int doc_num)
|
23
|
+
{
|
24
|
+
Filter *filter = (Filter *)self->query->data;
|
25
|
+
Explanation *expl;
|
26
|
+
char *filter_str = filter->to_s(filter);
|
27
|
+
BitVector *bv = filt_get_bv(filter, ir);
|
28
|
+
|
29
|
+
if (bv_get(bv, doc_num)) {
|
30
|
+
expl = expl_create(self->value,
|
31
|
+
epstrdup("ConstantScoreQuery(%s), product of:",
|
32
|
+
strlen(filter_str), filter_str));
|
33
|
+
expl_add_detail(expl, expl_create(self->query->boost, estrdup("boost")));
|
34
|
+
expl_add_detail(expl, expl_create(self->qnorm, estrdup("query_norm")));
|
35
|
+
} else {
|
36
|
+
expl = expl_create(self->value,
|
37
|
+
epstrdup("ConstantScoreQuery(%s), does not match id %d",
|
38
|
+
strlen(filter_str) + 20, filter_str, doc_num));
|
39
|
+
}
|
40
|
+
free(filter_str);
|
41
|
+
return expl;
|
42
|
+
}
|
43
|
+
|
44
|
+
Weight *csw_create(Query *query, Searcher *searcher)
|
45
|
+
{
|
46
|
+
Weight *self = ALLOC(Weight);
|
47
|
+
ZEROSET(self, Weight, 1);
|
48
|
+
self->get_query = &w_get_query;
|
49
|
+
self->get_value = &w_get_value;
|
50
|
+
self->normalize = &w_normalize;
|
51
|
+
self->scorer = &cssc_create;
|
52
|
+
self->explain = &csw_explain;
|
53
|
+
self->to_s = &csw_to_s;
|
54
|
+
self->destroy = &csw_destroy;
|
55
|
+
self->sum_of_squared_weights = &w_sum_of_squared_weights;
|
56
|
+
|
57
|
+
self->similarity = query->get_similarity(query, searcher);
|
58
|
+
self->idf = 1.0;
|
59
|
+
self->query = query;
|
60
|
+
self->value = 0.0;
|
61
|
+
|
62
|
+
return self;
|
63
|
+
}
|
64
|
+
|
65
|
+
/***************************************************************************
|
66
|
+
*
|
67
|
+
* ConstantScoreQuery
|
68
|
+
*
|
69
|
+
***************************************************************************/
|
70
|
+
|
71
|
+
char *csq_to_s(Query *self, char *field)
|
72
|
+
{
|
73
|
+
Filter *filter = (Filter *)self->data;
|
74
|
+
char *filter_str = filter->to_s(filter);
|
75
|
+
char *buffer;
|
76
|
+
if (self->boost == 1.0) {
|
77
|
+
buffer = epstrdup("ConstantScore(%s)", strlen(filter_str), filter_str);
|
78
|
+
} else {
|
79
|
+
char dbuf[32];
|
80
|
+
dbl_to_s(dbuf, self->boost);
|
81
|
+
buffer = epstrdup("ConstantScore(%s)^%s",
|
82
|
+
strlen(filter_str) + strlen(dbuf), filter_str, dbuf);
|
83
|
+
}
|
84
|
+
free(filter_str);
|
85
|
+
return buffer;;
|
86
|
+
}
|
87
|
+
|
88
|
+
void csq_destroy(void *p)
|
89
|
+
{
|
90
|
+
Query *self = (Query *)p;
|
91
|
+
if (self->destroy_all) {
|
92
|
+
Filter *filter = (Filter *)self->data;
|
93
|
+
filter->destroy(filter);
|
94
|
+
}
|
95
|
+
q_destroy(self);
|
96
|
+
}
|
97
|
+
|
98
|
+
Query *csq_create(Filter *filter)
|
99
|
+
{
|
100
|
+
Query *self = q_create();
|
101
|
+
self->type = CONSTANT_QUERY;
|
102
|
+
self->data = filter;
|
103
|
+
self->create_weight = &csw_create;
|
104
|
+
self->to_s = &csq_to_s;
|
105
|
+
self->destroy = &csq_destroy;
|
106
|
+
|
107
|
+
return self;
|
108
|
+
}
|
109
|
+
|
110
|
+
/***************************************************************************
|
111
|
+
*
|
112
|
+
* ConstantScoreScorer
|
113
|
+
*
|
114
|
+
***************************************************************************/
|
115
|
+
|
116
|
+
float cssc_score(Scorer *self)
|
117
|
+
{
|
118
|
+
return ((ConstantScoreScorer *)self->data)->score;
|
119
|
+
}
|
120
|
+
|
121
|
+
bool cssc_next(Scorer *self)
|
122
|
+
{
|
123
|
+
BitVector *bv = ((ConstantScoreScorer *)self->data)->bv;
|
124
|
+
return ((self->doc = bv_scan_next(bv)) >= 0);
|
125
|
+
}
|
126
|
+
|
127
|
+
bool cssc_skip_to(Scorer *self, int doc_num)
|
128
|
+
{
|
129
|
+
BitVector *bv = ((ConstantScoreScorer *)self->data)->bv;
|
130
|
+
return ((self->doc = bv_scan_next_from(bv, doc_num)) >= 0);
|
131
|
+
}
|
132
|
+
|
133
|
+
Explanation *cssc_explain(Scorer *self, int doc_num)
|
134
|
+
{
|
135
|
+
return expl_create(1.0, estrdup("ConstantScoreScorer"));
|
136
|
+
}
|
137
|
+
|
138
|
+
Scorer *cssc_create(Weight *weight, IndexReader *ir)
|
139
|
+
{
|
140
|
+
Scorer *self = scorer_create(weight->similarity);
|
141
|
+
Filter *filter = (Filter *)weight->query->data;
|
142
|
+
ConstantScoreScorer *cssc = ALLOC(ConstantScoreScorer);
|
143
|
+
ZEROSET(cssc, ConstantScoreScorer, 1);
|
144
|
+
self->data = cssc;
|
145
|
+
cssc->score = weight->value;
|
146
|
+
cssc->bv = filt_get_bv(filter, ir);
|
147
|
+
|
148
|
+
self->score = &cssc_score;
|
149
|
+
self->next = &cssc_next;
|
150
|
+
self->skip_to = &cssc_skip_to;
|
151
|
+
self->explain = &cssc_explain;
|
152
|
+
self->destroy = &scorer_destroy;
|
153
|
+
return self;
|
154
|
+
}
|
data/ext/q_fuzzy.c
ADDED
@@ -0,0 +1,287 @@
|
|
1
|
+
#include <string.h>
|
2
|
+
#include "search.h"
|
3
|
+
|
4
|
+
/****************************************************************************
|
5
|
+
*
|
6
|
+
* FuzzyStuff
|
7
|
+
*
|
8
|
+
* The main method here is the fuzq_score method which scores a term against
|
9
|
+
* another term. The other methods all act in support.
|
10
|
+
*
|
11
|
+
****************************************************************************/
|
12
|
+
|
13
|
+
|
14
|
+
int fuzq_calculate_max_distance(FuzzyQuery *fuzq, int m)
|
15
|
+
{
|
16
|
+
return (int)((1.0 - fuzq->min_sim) * (MIN(fuzq->text_len, m) + fuzq->pre_len));
|
17
|
+
}
|
18
|
+
|
19
|
+
void fuzq_initialize_max_distances(FuzzyQuery *fuzq)
|
20
|
+
{
|
21
|
+
int i;
|
22
|
+
for (i = 0; i < TYPICAL_LONGEST_WORD; i++) {
|
23
|
+
fuzq->max_distances[i] = fuzq_calculate_max_distance(fuzq, i);
|
24
|
+
}
|
25
|
+
}
|
26
|
+
|
27
|
+
float fuzq_get_max_distance(FuzzyQuery *fuzq, int m)
|
28
|
+
{
|
29
|
+
return (m < TYPICAL_LONGEST_WORD) ? fuzq->max_distances[m]
|
30
|
+
: fuzq_calculate_max_distance(fuzq, m);
|
31
|
+
}
|
32
|
+
|
33
|
+
float fuzq_score(FuzzyQuery *fuzq, char *target)
|
34
|
+
{
|
35
|
+
int i, j;
|
36
|
+
int max_distance;
|
37
|
+
int m = strlen(target);
|
38
|
+
int n = fuzq->text_len;
|
39
|
+
int *d = fuzq->da;
|
40
|
+
char *text = fuzq->text;
|
41
|
+
if (n == 0) {
|
42
|
+
/* we don't have anything to compare. That means if we just add
|
43
|
+
* the letters for m we get the new word */
|
44
|
+
return fuzq->pre_len == 0 ? 0.0f : 1.0f - ((float) m / fuzq->pre_len);
|
45
|
+
}
|
46
|
+
if (m == 0) {
|
47
|
+
return fuzq->pre_len == 0 ? 0.0f : 1.0f - ((float) n / fuzq->pre_len);
|
48
|
+
}
|
49
|
+
|
50
|
+
max_distance = fuzq_get_max_distance(fuzq, m);
|
51
|
+
|
52
|
+
//printf("n%dm%dmd%ddiff%d<%s><%s>\n", n, m, max_distance, m-n, fuzq->text, target);
|
53
|
+
if (max_distance < ((m > n) ? (m-n) : (n-m))) { /* abs */
|
54
|
+
/* Just adding the characters of m to n or vice-versa results in too many
|
55
|
+
* edits for example "pre" length is 3 and "prefixes" length is 8. We can
|
56
|
+
* see that given this optimal circumstance, the edit distance cannot be
|
57
|
+
* less than 5 which is 8-3 or more precisesly Math.abs(3-8). If our
|
58
|
+
* maximum edit distance is 4, then we can discard this word without
|
59
|
+
* looking at it. */
|
60
|
+
return 0.0f;
|
61
|
+
}
|
62
|
+
|
63
|
+
/* Let's make sure we have enough room in our array to do the distance
|
64
|
+
* calculations. */
|
65
|
+
if (((m+1) * (n+1)) >= fuzq->da_capa) {
|
66
|
+
fuzq->da_capa = (m * (fuzq->text_len+1)) * 2;
|
67
|
+
REALLOC_N(fuzq->da, int, fuzq->da_capa);
|
68
|
+
d = fuzq->da;
|
69
|
+
}
|
70
|
+
|
71
|
+
/* init matrix d */
|
72
|
+
for (i = 0; i <= n; i++) d[i + m * 0] = i;
|
73
|
+
for (j = 0; j <= m; j++) d[0 + m * j] = j;
|
74
|
+
|
75
|
+
/* start computing edit distance */
|
76
|
+
for (i = 1; i <= n; i++) {
|
77
|
+
int best_pos_ed_dist = m;
|
78
|
+
char s_i = text[i - 1];
|
79
|
+
for (j = 1; j <= m; j++) {
|
80
|
+
if (s_i != target[j-1]) {
|
81
|
+
d[i + m*j] = min3(d[i-1 + m*j], d[i + m*(j-1)], d[i-1 + m*(j-1)])+1;
|
82
|
+
} else {
|
83
|
+
d[i + m*j] = min3(d[i-1 + m*j]+1, d[i + m*(j-1)]+1, d[i-1 + m*(j-1)]);
|
84
|
+
}
|
85
|
+
best_pos_ed_dist = min(best_pos_ed_dist, d[i + m*j]);
|
86
|
+
}
|
87
|
+
//printf("(bped = %d, i = %d, md = %d)", best_pos_ed_dist, i, max_distance);
|
88
|
+
|
89
|
+
/* After calculating row i, the best possible edit distance can be found
|
90
|
+
* by found by finding the smallest value in a given column. If the
|
91
|
+
* best_pos_ed_dist is greater than the max distance, abort.
|
92
|
+
*/
|
93
|
+
|
94
|
+
if ((i > max_distance) && (best_pos_ed_dist > max_distance)) {
|
95
|
+
/* equal is okay, but not greater
|
96
|
+
* the closest the target can be to the text is just too far away.
|
97
|
+
* this target is leaving the party early. */
|
98
|
+
return 0.0f;
|
99
|
+
}
|
100
|
+
}
|
101
|
+
//printf("<%f, d[n + m*m] = %d min_len = %d>", 1.0f - ((float)d[n + m*m] / (float) (fuzq->pre_len + min(n, m))), d[n + m*m], fuzq->pre_len + min(n, m));
|
102
|
+
|
103
|
+
/* this will return less than 0.0 when the edit distance is greater than the
|
104
|
+
* number of characters in the shorter word. but this was the formula that
|
105
|
+
* was previously used in FuzzyTermEnum, so it has not been changed (even
|
106
|
+
* though min_sim must be greater than 0.0) */
|
107
|
+
return 1.0f - ((float)d[n + m*m] / (float) (fuzq->pre_len + min(n, m)));
|
108
|
+
}
|
109
|
+
|
110
|
+
/****************************************************************************
|
111
|
+
*
|
112
|
+
* FuzzyQuery
|
113
|
+
*
|
114
|
+
****************************************************************************/
|
115
|
+
|
116
|
+
char *fuzq_to_s(Query *self, char *field)
|
117
|
+
{
|
118
|
+
char *buffer, *bptr;
|
119
|
+
FuzzyQuery *fuzq = (FuzzyQuery *)self->data;
|
120
|
+
Term *term = fuzq->term;
|
121
|
+
int tlen = strlen(term->text);
|
122
|
+
int flen = strlen(term->field);
|
123
|
+
bptr = buffer = ALLOC_N(char, tlen + flen + 35);
|
124
|
+
|
125
|
+
if (strcmp(term->field, field) != 0) {
|
126
|
+
sprintf(bptr, "%s:", term->field);
|
127
|
+
bptr += strlen(term->field) + 1;
|
128
|
+
}
|
129
|
+
sprintf(bptr, "%s~", term->text);
|
130
|
+
bptr += strlen(bptr);
|
131
|
+
if (fuzq->min_sim != 0.5) {
|
132
|
+
dbl_to_s(bptr, fuzq->min_sim);
|
133
|
+
bptr += strlen(bptr);
|
134
|
+
}
|
135
|
+
if (self->boost != 1.0) {
|
136
|
+
*bptr = '^';
|
137
|
+
dbl_to_s(++bptr, self->boost);
|
138
|
+
}
|
139
|
+
return buffer;
|
140
|
+
}
|
141
|
+
|
142
|
+
typedef struct ScoredTerm {
|
143
|
+
Term *term;
|
144
|
+
float score;
|
145
|
+
} ScoredTerm;
|
146
|
+
|
147
|
+
bool scored_term_less_than(void *p1, void *p2)
|
148
|
+
{
|
149
|
+
ScoredTerm *st1 = (ScoredTerm *)p1;
|
150
|
+
ScoredTerm *st2 = (ScoredTerm *)p2;
|
151
|
+
|
152
|
+
if (st1->score == st2->score)
|
153
|
+
return (strcmp(st1->term->text, st2->term->text) < 0);
|
154
|
+
|
155
|
+
return (st1->score < st2->score);
|
156
|
+
}
|
157
|
+
|
158
|
+
void scored_term_destroy(void *p)
|
159
|
+
{
|
160
|
+
ScoredTerm *st = (ScoredTerm *)p;
|
161
|
+
term_destroy(st->term);
|
162
|
+
free(st);
|
163
|
+
}
|
164
|
+
|
165
|
+
ScoredTerm *scored_term_create(Term *term, float score)
|
166
|
+
{
|
167
|
+
ScoredTerm *self = ALLOC(ScoredTerm);
|
168
|
+
self->term = term;
|
169
|
+
self->score = score;
|
170
|
+
return self;
|
171
|
+
}
|
172
|
+
|
173
|
+
Query *fuzq_rewrite(Query *self, IndexReader *ir)
|
174
|
+
{
|
175
|
+
Query *q;
|
176
|
+
Query *tq;
|
177
|
+
FuzzyQuery *fuzq = (FuzzyQuery *)self->data;
|
178
|
+
|
179
|
+
Term *term = fuzq->term;
|
180
|
+
char *text = term->text;
|
181
|
+
char *field = term->field;
|
182
|
+
Term prefix_term;
|
183
|
+
prefix_term.field = field;
|
184
|
+
if (fuzq->pre_len >= strlen(text)) {
|
185
|
+
q = tq_create(term_clone(term));
|
186
|
+
} else {
|
187
|
+
PriorityQueue *term_pq;
|
188
|
+
TermEnum *te;
|
189
|
+
Term prefix_term;
|
190
|
+
char *prefix = NULL;
|
191
|
+
int pre_len = fuzq->pre_len;
|
192
|
+
ScoredTerm *scored_term;
|
193
|
+
|
194
|
+
q = bq_create(true);
|
195
|
+
|
196
|
+
term_pq = pq_create(((BooleanQuery *)q->data)->max_clause_cnt,
|
197
|
+
&scored_term_less_than);
|
198
|
+
term_pq->free_elem = &scored_term_destroy;
|
199
|
+
|
200
|
+
prefix_term.field = field;
|
201
|
+
prefix_term.text = (char *)EMPTY_STRING;
|
202
|
+
if (pre_len >= 0) {
|
203
|
+
prefix = ALLOC_N(char, pre_len + 1);
|
204
|
+
strncpy(prefix, text, pre_len);
|
205
|
+
prefix_term.text = prefix;
|
206
|
+
prefix_term.text[pre_len] = '\0';
|
207
|
+
}
|
208
|
+
te = ir->terms_from(ir, &prefix_term);
|
209
|
+
|
210
|
+
fuzq->scale_factor = 1.0 / (1.0 - fuzq->min_sim);
|
211
|
+
fuzq->text = fuzq->term->text + pre_len;
|
212
|
+
fuzq->text_len = strlen(fuzq->text);
|
213
|
+
fuzq_initialize_max_distances(fuzq);
|
214
|
+
|
215
|
+
if (te) {
|
216
|
+
TermBuffer *tb = te->tb_curr;
|
217
|
+
float score = 0.0, min_score = fuzq->min_sim;
|
218
|
+
|
219
|
+
do {
|
220
|
+
if (strcmp(tb->field, field) != 0 ||
|
221
|
+
(prefix && strncmp(tb->text, prefix, pre_len) != 0))
|
222
|
+
break;
|
223
|
+
|
224
|
+
score = fuzq_score(fuzq, tb->text + pre_len);
|
225
|
+
//printf("%s:%s:%f\n", tb->text, fuzq->text, score);
|
226
|
+
|
227
|
+
if (score > min_score) {
|
228
|
+
pq_insert(term_pq, scored_term_create(tb_get_term(tb), score));
|
229
|
+
if (pq_full(term_pq))
|
230
|
+
min_score = ((ScoredTerm *)pq_top(term_pq))->score;
|
231
|
+
}
|
232
|
+
} while ((tb = te->next(te)) != NULL);
|
233
|
+
te->close(te);
|
234
|
+
}
|
235
|
+
free(prefix);
|
236
|
+
|
237
|
+
while ((scored_term = pq_pop(term_pq)) != NULL) {
|
238
|
+
tq = tq_create(scored_term->term); /* found match */
|
239
|
+
tq->boost = self->boost; /* set boost */
|
240
|
+
bq_add_query(q, tq, BC_SHOULD); /* add query */
|
241
|
+
free(scored_term); /* no need to free the term as it's in the query */
|
242
|
+
}
|
243
|
+
pq_destroy(term_pq);
|
244
|
+
}
|
245
|
+
|
246
|
+
if (self->rewritten) self->rewritten->destroy(self->rewritten);
|
247
|
+
return self->rewritten = q;
|
248
|
+
}
|
249
|
+
|
250
|
+
void fuzq_destroy(void *p)
|
251
|
+
{
|
252
|
+
Query *self = (Query *)p;
|
253
|
+
FuzzyQuery *fuzq = (FuzzyQuery *)self->data;
|
254
|
+
if (self->destroy_all) term_destroy((Term *)fuzq->term);
|
255
|
+
free(fuzq->da);
|
256
|
+
free(fuzq);
|
257
|
+
q_destroy(self);
|
258
|
+
}
|
259
|
+
|
260
|
+
Query *fuzq_create(Term *term)
|
261
|
+
{
|
262
|
+
Query *self = q_create();
|
263
|
+
FuzzyQuery *fq = ALLOC(FuzzyQuery);
|
264
|
+
ZEROSET(fq, FuzzyQuery, 1);
|
265
|
+
|
266
|
+
fq->term = term;
|
267
|
+
fq->pre_len = DEF_PRE_LEN;
|
268
|
+
fq->min_sim = DEF_MIN_SIM;
|
269
|
+
self->data = fq;
|
270
|
+
self->type = FUZZY_QUERY;
|
271
|
+
self->create_weight = NULL;
|
272
|
+
self->to_s = &fuzq_to_s;
|
273
|
+
self->rewrite = &fuzq_rewrite;
|
274
|
+
self->destroy = &fuzq_destroy;
|
275
|
+
self->rewritten = NULL;
|
276
|
+
|
277
|
+
return self;
|
278
|
+
}
|
279
|
+
|
280
|
+
Query *fuzq_create_mp(Term *term, float min_sim, int pre_len)
|
281
|
+
{
|
282
|
+
Query *self = fuzq_create(term);
|
283
|
+
FuzzyQuery *fuzq = (FuzzyQuery *)self->data;
|
284
|
+
if (pre_len) fuzq->pre_len = pre_len;
|
285
|
+
if (min_sim) fuzq->min_sim = min_sim;
|
286
|
+
return self;
|
287
|
+
}
|
data/ext/q_match_all.c
ADDED
@@ -0,0 +1,142 @@
|
|
1
|
+
#include "search.h"
|
2
|
+
#include <string.h>
|
3
|
+
|
4
|
+
/***************************************************************************
|
5
|
+
*
|
6
|
+
* Weight
|
7
|
+
*
|
8
|
+
***************************************************************************/
|
9
|
+
|
10
|
+
char *maw_to_s(Weight *self)
|
11
|
+
{
|
12
|
+
char dbuf[32];
|
13
|
+
dbl_to_s(dbuf, self->value);
|
14
|
+
return epstrdup("MatchAllWeight(%s)", strlen(dbuf), dbuf);
|
15
|
+
}
|
16
|
+
|
17
|
+
void maw_destroy(void *p)
|
18
|
+
{
|
19
|
+
free(p);
|
20
|
+
}
|
21
|
+
|
22
|
+
Explanation *maw_explain(Weight *self, IndexReader *ir, int doc_num)
|
23
|
+
{
|
24
|
+
Explanation *expl;
|
25
|
+
if (!ir->is_deleted(ir, doc_num)) {
|
26
|
+
expl = expl_create(self->value, estrdup("MatchAllQuery: product of:"));
|
27
|
+
expl_add_detail(expl, expl_create(self->query->boost, estrdup("boost")));
|
28
|
+
expl_add_detail(expl, expl_create(self->qnorm, estrdup("query_norm")));
|
29
|
+
} else {
|
30
|
+
expl = expl_create(self->value,
|
31
|
+
epstrdup("MatchAllQuery: doc %d was deleted", 20, doc_num));
|
32
|
+
}
|
33
|
+
|
34
|
+
return expl;
|
35
|
+
}
|
36
|
+
|
37
|
+
Weight *maw_create(Query *query, Searcher *searcher)
|
38
|
+
{
|
39
|
+
Weight *self = ALLOC(Weight);
|
40
|
+
ZEROSET(self, Weight, 1);
|
41
|
+
self->get_query = &w_get_query;
|
42
|
+
self->get_value = &w_get_value;
|
43
|
+
self->normalize = &w_normalize;
|
44
|
+
self->scorer = &masc_create;
|
45
|
+
self->explain = &maw_explain;
|
46
|
+
self->to_s = &maw_to_s;
|
47
|
+
self->destroy = &maw_destroy;
|
48
|
+
self->sum_of_squared_weights = &w_sum_of_squared_weights;
|
49
|
+
|
50
|
+
self->similarity = query->get_similarity(query, searcher);
|
51
|
+
self->idf = 1.0;
|
52
|
+
self->query = query;
|
53
|
+
self->value = 0.0;
|
54
|
+
|
55
|
+
return self;
|
56
|
+
}
|
57
|
+
|
58
|
+
/***************************************************************************
|
59
|
+
*
|
60
|
+
* MatchAllQuery
|
61
|
+
*
|
62
|
+
***************************************************************************/
|
63
|
+
|
64
|
+
char *maq_to_s(Query *self, char *field)
|
65
|
+
{
|
66
|
+
if (self->boost == 1.0) {
|
67
|
+
return estrdup("MatchAll");
|
68
|
+
} else {
|
69
|
+
char dbuf[32];
|
70
|
+
dbl_to_s(dbuf, self->boost);
|
71
|
+
return epstrdup("MatchAll^%s", strlen(dbuf), dbuf);
|
72
|
+
}
|
73
|
+
}
|
74
|
+
|
75
|
+
void maq_destroy(void *p)
|
76
|
+
{
|
77
|
+
Query *self = (Query *)p;
|
78
|
+
q_destroy(self);
|
79
|
+
}
|
80
|
+
|
81
|
+
Query *maq_create()
|
82
|
+
{
|
83
|
+
Query *self = q_create();
|
84
|
+
self->type = MATCH_ALL_QUERY;
|
85
|
+
self->create_weight = &maw_create;
|
86
|
+
self->to_s = &maq_to_s;
|
87
|
+
self->destroy = &maq_destroy;
|
88
|
+
|
89
|
+
return self;
|
90
|
+
}
|
91
|
+
|
92
|
+
/***************************************************************************
|
93
|
+
*
|
94
|
+
* MatchAllScorer
|
95
|
+
*
|
96
|
+
***************************************************************************/
|
97
|
+
|
98
|
+
float masc_score(Scorer *self)
|
99
|
+
{
|
100
|
+
return ((MatchAllScorer *)self->data)->score;
|
101
|
+
}
|
102
|
+
|
103
|
+
bool masc_next(Scorer *self)
|
104
|
+
{
|
105
|
+
MatchAllScorer *mas = (MatchAllScorer *)self->data;
|
106
|
+
while (self->doc < (mas->max_doc - 1)) {
|
107
|
+
self->doc++;
|
108
|
+
if (!mas->ir->is_deleted(mas->ir, self->doc)) {
|
109
|
+
return true;
|
110
|
+
}
|
111
|
+
}
|
112
|
+
return false;
|
113
|
+
}
|
114
|
+
|
115
|
+
bool masc_skip_to(Scorer *self, int doc_num)
|
116
|
+
{
|
117
|
+
self->doc = doc_num - 1;
|
118
|
+
return masc_next(self);
|
119
|
+
}
|
120
|
+
|
121
|
+
Explanation *masc_explain(Scorer *self, int doc_num)
|
122
|
+
{
|
123
|
+
return expl_create(1.0, estrdup("MatchAllScorer"));
|
124
|
+
}
|
125
|
+
|
126
|
+
Scorer *masc_create(Weight *weight, IndexReader *ir)
|
127
|
+
{
|
128
|
+
Scorer *self = scorer_create(weight->similarity);
|
129
|
+
MatchAllScorer *mas = ALLOC(MatchAllScorer);
|
130
|
+
mas->ir = ir;
|
131
|
+
mas->max_doc = ir->max_doc(ir);
|
132
|
+
mas->score = weight->value;
|
133
|
+
self->data = mas;
|
134
|
+
|
135
|
+
self->doc = -1;
|
136
|
+
self->score = &masc_score;
|
137
|
+
self->next = &masc_next;
|
138
|
+
self->skip_to = &masc_skip_to;
|
139
|
+
self->explain = &masc_explain;
|
140
|
+
self->destroy = &scorer_destroy;
|
141
|
+
return self;
|
142
|
+
}
|