ferret 0.9.6 → 0.10.0
Sign up to get free protection for your applications and to get access to all the features.
- data/MIT-LICENSE +1 -1
- data/README +12 -24
- data/Rakefile +38 -54
- data/TODO +14 -17
- data/ext/analysis.c +982 -823
- data/ext/analysis.h +133 -76
- data/ext/array.c +96 -58
- data/ext/array.h +40 -13
- data/ext/bitvector.c +476 -118
- data/ext/bitvector.h +264 -22
- data/ext/compound_io.c +217 -229
- data/ext/defines.h +49 -0
- data/ext/document.c +107 -317
- data/ext/document.h +31 -65
- data/ext/except.c +81 -36
- data/ext/except.h +117 -55
- data/ext/extconf.rb +2 -9
- data/ext/ferret.c +211 -104
- data/ext/ferret.h +22 -11
- data/ext/filter.c +97 -82
- data/ext/fs_store.c +348 -367
- data/ext/global.c +226 -188
- data/ext/global.h +44 -26
- data/ext/hash.c +474 -391
- data/ext/hash.h +441 -68
- data/ext/hashset.c +124 -96
- data/ext/hashset.h +169 -20
- data/ext/helper.c +56 -5
- data/ext/helper.h +7 -0
- data/ext/inc/lang.h +29 -49
- data/ext/inc/threading.h +31 -0
- data/ext/ind.c +288 -278
- data/ext/ind.h +68 -0
- data/ext/index.c +5688 -0
- data/ext/index.h +663 -616
- data/ext/lang.h +29 -49
- data/ext/libstemmer.c +3 -3
- data/ext/mem_pool.c +84 -0
- data/ext/mem_pool.h +35 -0
- data/ext/posh.c +1006 -0
- data/ext/posh.h +1007 -0
- data/ext/priorityqueue.c +117 -194
- data/ext/priorityqueue.h +135 -39
- data/ext/q_boolean.c +1305 -1108
- data/ext/q_const_score.c +106 -93
- data/ext/q_filtered_query.c +138 -135
- data/ext/q_fuzzy.c +206 -242
- data/ext/q_match_all.c +94 -80
- data/ext/q_multi_term.c +663 -0
- data/ext/q_parser.c +667 -593
- data/ext/q_phrase.c +992 -555
- data/ext/q_prefix.c +72 -61
- data/ext/q_range.c +235 -210
- data/ext/q_span.c +1480 -1166
- data/ext/q_term.c +273 -246
- data/ext/q_wildcard.c +127 -114
- data/ext/r_analysis.c +1720 -711
- data/ext/r_index.c +3049 -0
- data/ext/r_qparser.c +433 -146
- data/ext/r_search.c +2934 -1993
- data/ext/r_store.c +372 -143
- data/ext/r_utils.c +941 -0
- data/ext/ram_store.c +330 -326
- data/ext/search.c +1291 -668
- data/ext/search.h +403 -702
- data/ext/similarity.c +91 -113
- data/ext/similarity.h +45 -30
- data/ext/sort.c +721 -484
- data/ext/stopwords.c +361 -273
- data/ext/store.c +556 -58
- data/ext/store.h +706 -126
- data/ext/tags +3578 -2780
- data/ext/term_vectors.c +352 -0
- data/ext/threading.h +31 -0
- data/ext/win32.h +54 -0
- data/lib/ferret.rb +5 -17
- data/lib/ferret/document.rb +130 -2
- data/lib/ferret/index.rb +577 -26
- data/lib/ferret/number_tools.rb +157 -0
- data/lib/ferret_version.rb +3 -0
- data/test/test_helper.rb +5 -13
- data/test/unit/analysis/tc_analyzer.rb +513 -1
- data/test/unit/analysis/{ctc_tokenstream.rb → tc_token_stream.rb} +23 -0
- data/test/unit/index/tc_index.rb +183 -240
- data/test/unit/index/tc_index_reader.rb +312 -479
- data/test/unit/index/tc_index_writer.rb +397 -13
- data/test/unit/index/th_doc.rb +269 -206
- data/test/unit/query_parser/tc_query_parser.rb +40 -33
- data/test/unit/search/tc_filter.rb +59 -71
- data/test/unit/search/tc_fuzzy_query.rb +24 -16
- data/test/unit/search/tc_index_searcher.rb +23 -201
- data/test/unit/search/tc_multi_searcher.rb +78 -226
- data/test/unit/search/tc_search_and_sort.rb +93 -81
- data/test/unit/search/tc_sort.rb +23 -23
- data/test/unit/search/tc_sort_field.rb +7 -7
- data/test/unit/search/tc_spans.rb +51 -47
- data/test/unit/search/tm_searcher.rb +339 -0
- data/test/unit/store/tc_fs_store.rb +1 -1
- data/test/unit/store/tm_store_lock.rb +3 -3
- data/test/unit/tc_document.rb +81 -0
- data/test/unit/ts_analysis.rb +1 -1
- data/test/unit/ts_utils.rb +1 -1
- data/test/unit/utils/tc_bit_vector.rb +288 -0
- data/test/unit/utils/tc_number_tools.rb +117 -0
- data/test/unit/utils/tc_priority_queue.rb +106 -0
- metadata +140 -301
- data/CHANGELOG +0 -9
- data/ext/dummy.exe +0 -0
- data/ext/field.c +0 -408
- data/ext/frtio.h +0 -13
- data/ext/inc/except.h +0 -90
- data/ext/index_io.c +0 -382
- data/ext/index_rw.c +0 -2658
- data/ext/lang.c +0 -41
- data/ext/nix_io.c +0 -134
- data/ext/q_multi_phrase.c +0 -380
- data/ext/r_doc.c +0 -582
- data/ext/r_index_io.c +0 -1021
- data/ext/r_term.c +0 -219
- data/ext/term.c +0 -820
- data/ext/termdocs.c +0 -611
- data/ext/vector.c +0 -637
- data/ext/w32_io.c +0 -150
- data/lib/ferret/analysis.rb +0 -11
- data/lib/ferret/analysis/analyzers.rb +0 -112
- data/lib/ferret/analysis/standard_tokenizer.rb +0 -71
- data/lib/ferret/analysis/token.rb +0 -100
- data/lib/ferret/analysis/token_filters.rb +0 -86
- data/lib/ferret/analysis/token_stream.rb +0 -26
- data/lib/ferret/analysis/tokenizers.rb +0 -112
- data/lib/ferret/analysis/word_list_loader.rb +0 -27
- data/lib/ferret/document/document.rb +0 -152
- data/lib/ferret/document/field.rb +0 -312
- data/lib/ferret/index/compound_file_io.rb +0 -338
- data/lib/ferret/index/document_writer.rb +0 -289
- data/lib/ferret/index/field_infos.rb +0 -279
- data/lib/ferret/index/fields_io.rb +0 -181
- data/lib/ferret/index/index.rb +0 -675
- data/lib/ferret/index/index_file_names.rb +0 -33
- data/lib/ferret/index/index_reader.rb +0 -503
- data/lib/ferret/index/index_writer.rb +0 -534
- data/lib/ferret/index/multi_reader.rb +0 -377
- data/lib/ferret/index/multiple_term_doc_pos_enum.rb +0 -98
- data/lib/ferret/index/segment_infos.rb +0 -130
- data/lib/ferret/index/segment_merge_info.rb +0 -49
- data/lib/ferret/index/segment_merge_queue.rb +0 -16
- data/lib/ferret/index/segment_merger.rb +0 -358
- data/lib/ferret/index/segment_reader.rb +0 -412
- data/lib/ferret/index/segment_term_enum.rb +0 -169
- data/lib/ferret/index/segment_term_vector.rb +0 -58
- data/lib/ferret/index/term.rb +0 -53
- data/lib/ferret/index/term_buffer.rb +0 -83
- data/lib/ferret/index/term_doc_enum.rb +0 -291
- data/lib/ferret/index/term_enum.rb +0 -52
- data/lib/ferret/index/term_info.rb +0 -37
- data/lib/ferret/index/term_infos_io.rb +0 -321
- data/lib/ferret/index/term_vector_offset_info.rb +0 -20
- data/lib/ferret/index/term_vectors_io.rb +0 -553
- data/lib/ferret/query_parser.rb +0 -312
- data/lib/ferret/query_parser/query_parser.tab.rb +0 -928
- data/lib/ferret/search.rb +0 -50
- data/lib/ferret/search/boolean_clause.rb +0 -100
- data/lib/ferret/search/boolean_query.rb +0 -299
- data/lib/ferret/search/boolean_scorer.rb +0 -294
- data/lib/ferret/search/caching_wrapper_filter.rb +0 -40
- data/lib/ferret/search/conjunction_scorer.rb +0 -99
- data/lib/ferret/search/disjunction_sum_scorer.rb +0 -205
- data/lib/ferret/search/exact_phrase_scorer.rb +0 -32
- data/lib/ferret/search/explanation.rb +0 -41
- data/lib/ferret/search/field_cache.rb +0 -215
- data/lib/ferret/search/field_doc.rb +0 -31
- data/lib/ferret/search/field_sorted_hit_queue.rb +0 -184
- data/lib/ferret/search/filter.rb +0 -11
- data/lib/ferret/search/filtered_query.rb +0 -130
- data/lib/ferret/search/filtered_term_enum.rb +0 -79
- data/lib/ferret/search/fuzzy_query.rb +0 -154
- data/lib/ferret/search/fuzzy_term_enum.rb +0 -247
- data/lib/ferret/search/hit_collector.rb +0 -34
- data/lib/ferret/search/hit_queue.rb +0 -11
- data/lib/ferret/search/index_searcher.rb +0 -200
- data/lib/ferret/search/match_all_query.rb +0 -104
- data/lib/ferret/search/multi_phrase_query.rb +0 -216
- data/lib/ferret/search/multi_searcher.rb +0 -261
- data/lib/ferret/search/multi_term_query.rb +0 -65
- data/lib/ferret/search/non_matching_scorer.rb +0 -22
- data/lib/ferret/search/phrase_positions.rb +0 -55
- data/lib/ferret/search/phrase_query.rb +0 -214
- data/lib/ferret/search/phrase_scorer.rb +0 -152
- data/lib/ferret/search/prefix_query.rb +0 -54
- data/lib/ferret/search/query.rb +0 -140
- data/lib/ferret/search/query_filter.rb +0 -51
- data/lib/ferret/search/range_filter.rb +0 -103
- data/lib/ferret/search/range_query.rb +0 -139
- data/lib/ferret/search/req_excl_scorer.rb +0 -125
- data/lib/ferret/search/req_opt_sum_scorer.rb +0 -70
- data/lib/ferret/search/score_doc.rb +0 -38
- data/lib/ferret/search/score_doc_comparator.rb +0 -114
- data/lib/ferret/search/scorer.rb +0 -91
- data/lib/ferret/search/similarity.rb +0 -278
- data/lib/ferret/search/sloppy_phrase_scorer.rb +0 -47
- data/lib/ferret/search/sort.rb +0 -112
- data/lib/ferret/search/sort_comparator.rb +0 -60
- data/lib/ferret/search/sort_field.rb +0 -91
- data/lib/ferret/search/spans.rb +0 -12
- data/lib/ferret/search/spans/near_spans_enum.rb +0 -304
- data/lib/ferret/search/spans/span_first_query.rb +0 -79
- data/lib/ferret/search/spans/span_near_query.rb +0 -108
- data/lib/ferret/search/spans/span_not_query.rb +0 -130
- data/lib/ferret/search/spans/span_or_query.rb +0 -176
- data/lib/ferret/search/spans/span_query.rb +0 -25
- data/lib/ferret/search/spans/span_scorer.rb +0 -74
- data/lib/ferret/search/spans/span_term_query.rb +0 -105
- data/lib/ferret/search/spans/span_weight.rb +0 -84
- data/lib/ferret/search/spans/spans_enum.rb +0 -44
- data/lib/ferret/search/term_query.rb +0 -128
- data/lib/ferret/search/term_scorer.rb +0 -183
- data/lib/ferret/search/top_docs.rb +0 -36
- data/lib/ferret/search/top_field_docs.rb +0 -17
- data/lib/ferret/search/weight.rb +0 -54
- data/lib/ferret/search/wildcard_query.rb +0 -26
- data/lib/ferret/search/wildcard_term_enum.rb +0 -61
- data/lib/ferret/stemmers.rb +0 -1
- data/lib/ferret/stemmers/porter_stemmer.rb +0 -218
- data/lib/ferret/store.rb +0 -5
- data/lib/ferret/store/buffered_index_io.rb +0 -190
- data/lib/ferret/store/directory.rb +0 -141
- data/lib/ferret/store/fs_store.rb +0 -381
- data/lib/ferret/store/index_io.rb +0 -245
- data/lib/ferret/store/ram_store.rb +0 -286
- data/lib/ferret/utils.rb +0 -8
- data/lib/ferret/utils/bit_vector.rb +0 -123
- data/lib/ferret/utils/date_tools.rb +0 -138
- data/lib/ferret/utils/number_tools.rb +0 -91
- data/lib/ferret/utils/parameter.rb +0 -41
- data/lib/ferret/utils/priority_queue.rb +0 -120
- data/lib/ferret/utils/string_helper.rb +0 -47
- data/lib/ferret/utils/thread_local.rb +0 -28
- data/lib/ferret/utils/weak_key_hash.rb +0 -60
- data/lib/rferret.rb +0 -37
- data/rake_utils/code_statistics.rb +0 -106
- data/test/benchmark/tb_ram_store.rb +0 -76
- data/test/benchmark/tb_rw_vint.rb +0 -26
- data/test/functional/thread_safety_index_test.rb +0 -81
- data/test/functional/thread_safety_test.rb +0 -137
- data/test/longrunning/tc_numbertools.rb +0 -60
- data/test/longrunning/tm_store.rb +0 -19
- data/test/unit/analysis/ctc_analyzer.rb +0 -532
- data/test/unit/analysis/data/wordfile +0 -6
- data/test/unit/analysis/rtc_letter_tokenizer.rb +0 -20
- data/test/unit/analysis/rtc_lower_case_filter.rb +0 -20
- data/test/unit/analysis/rtc_lower_case_tokenizer.rb +0 -27
- data/test/unit/analysis/rtc_per_field_analyzer_wrapper.rb +0 -39
- data/test/unit/analysis/rtc_porter_stem_filter.rb +0 -16
- data/test/unit/analysis/rtc_standard_analyzer.rb +0 -20
- data/test/unit/analysis/rtc_standard_tokenizer.rb +0 -20
- data/test/unit/analysis/rtc_stop_analyzer.rb +0 -20
- data/test/unit/analysis/rtc_stop_filter.rb +0 -14
- data/test/unit/analysis/rtc_white_space_analyzer.rb +0 -21
- data/test/unit/analysis/rtc_white_space_tokenizer.rb +0 -20
- data/test/unit/analysis/rtc_word_list_loader.rb +0 -32
- data/test/unit/analysis/tc_token.rb +0 -25
- data/test/unit/document/rtc_field.rb +0 -28
- data/test/unit/document/tc_document.rb +0 -47
- data/test/unit/document/tc_field.rb +0 -98
- data/test/unit/index/rtc_compound_file_io.rb +0 -107
- data/test/unit/index/rtc_field_infos.rb +0 -127
- data/test/unit/index/rtc_fields_io.rb +0 -167
- data/test/unit/index/rtc_multiple_term_doc_pos_enum.rb +0 -83
- data/test/unit/index/rtc_segment_infos.rb +0 -74
- data/test/unit/index/rtc_segment_term_docs.rb +0 -17
- data/test/unit/index/rtc_segment_term_enum.rb +0 -60
- data/test/unit/index/rtc_segment_term_vector.rb +0 -71
- data/test/unit/index/rtc_term_buffer.rb +0 -57
- data/test/unit/index/rtc_term_info.rb +0 -19
- data/test/unit/index/rtc_term_infos_io.rb +0 -192
- data/test/unit/index/rtc_term_vectors_io.rb +0 -108
- data/test/unit/index/tc_term.rb +0 -27
- data/test/unit/index/tc_term_voi.rb +0 -18
- data/test/unit/search/rtc_similarity.rb +0 -37
- data/test/unit/search/rtc_sort_field.rb +0 -14
- data/test/unit/search/tc_multi_searcher2.rb +0 -126
- data/test/unit/store/rtc_fs_store.rb +0 -62
- data/test/unit/store/rtc_ram_store.rb +0 -15
- data/test/unit/store/rtm_store.rb +0 -150
- data/test/unit/store/rtm_store_lock.rb +0 -2
- data/test/unit/ts_document.rb +0 -2
- data/test/unit/utils/rtc_bit_vector.rb +0 -73
- data/test/unit/utils/rtc_date_tools.rb +0 -50
- data/test/unit/utils/rtc_number_tools.rb +0 -59
- data/test/unit/utils/rtc_parameter.rb +0 -40
- data/test/unit/utils/rtc_priority_queue.rb +0 -62
- data/test/unit/utils/rtc_string_helper.rb +0 -21
- data/test/unit/utils/rtc_thread.rb +0 -61
- data/test/unit/utils/rtc_weak_key_hash.rb +0 -25
- data/test/utils/number_to_spoken.rb +0 -132
data/ext/r_term.c
DELETED
@@ -1,219 +0,0 @@
|
|
1
|
-
#include "ferret.h"
|
2
|
-
#include "index.h"
|
3
|
-
|
4
|
-
VALUE cTerm;
|
5
|
-
/****************************************************************************
|
6
|
-
*
|
7
|
-
* Term Methods
|
8
|
-
*
|
9
|
-
****************************************************************************/
|
10
|
-
|
11
|
-
typedef struct RTerm {
|
12
|
-
VALUE field;
|
13
|
-
VALUE text;
|
14
|
-
} RTerm;
|
15
|
-
|
16
|
-
void
|
17
|
-
frt_term_mark(void *p)
|
18
|
-
{
|
19
|
-
RTerm *term = (RTerm *)p;
|
20
|
-
rb_gc_mark(term->field);
|
21
|
-
rb_gc_mark(term->text);
|
22
|
-
}
|
23
|
-
|
24
|
-
static VALUE
|
25
|
-
frt_term_alloc(VALUE klass)
|
26
|
-
{
|
27
|
-
RTerm *term = ALLOC(RTerm);
|
28
|
-
term->field = Qnil;
|
29
|
-
term->text = Qnil;
|
30
|
-
return Data_Wrap_Struct(klass, &frt_term_mark, &free, term);
|
31
|
-
}
|
32
|
-
|
33
|
-
#define GET_TERM RTerm *term; Data_Get_Struct(self, RTerm, term)
|
34
|
-
static VALUE
|
35
|
-
frt_term_init(VALUE self, VALUE rfield, VALUE rtext)
|
36
|
-
{
|
37
|
-
GET_TERM;
|
38
|
-
term->field = rb_obj_as_string(rfield);
|
39
|
-
term->text = rb_obj_as_string(rtext);
|
40
|
-
return self;
|
41
|
-
}
|
42
|
-
|
43
|
-
VALUE
|
44
|
-
frt_get_rterm(char *field, char *text)
|
45
|
-
{
|
46
|
-
RTerm *rterm = ALLOC(RTerm);
|
47
|
-
rterm->field = rb_str_new2(field);
|
48
|
-
rterm->text = rb_str_new2(text);
|
49
|
-
return Data_Wrap_Struct(cTerm, &frt_term_mark, &free, rterm);
|
50
|
-
}
|
51
|
-
|
52
|
-
Term *
|
53
|
-
frt_set_term(VALUE self, Term *t)
|
54
|
-
{
|
55
|
-
GET_TERM;
|
56
|
-
t->field = RSTRING(term->field)->ptr;
|
57
|
-
t->text = RSTRING(term->text)->ptr;
|
58
|
-
return t;
|
59
|
-
}
|
60
|
-
|
61
|
-
Term *
|
62
|
-
frt_get_term(VALUE self)
|
63
|
-
{
|
64
|
-
Term *t = NULL;
|
65
|
-
if (self != Qnil) {
|
66
|
-
GET_TERM;
|
67
|
-
t = ALLOC(Term);
|
68
|
-
/* store text and field in text so that field will be freed with text */
|
69
|
-
t->text = ALLOC_N(char, RSTRING(term->text)->len +
|
70
|
-
RSTRING(term->field)->len + 2);
|
71
|
-
sprintf(t->text, "%s %s", RSTRING(term->text)->ptr,
|
72
|
-
RSTRING(term->field)->ptr);
|
73
|
-
t->text[RSTRING(term->text)->len] = '\0';
|
74
|
-
t->field = t->text + RSTRING(term->text)->len + 1;
|
75
|
-
}
|
76
|
-
return t;
|
77
|
-
}
|
78
|
-
|
79
|
-
static VALUE
|
80
|
-
frt_term_get_text(VALUE self)
|
81
|
-
{
|
82
|
-
GET_TERM;
|
83
|
-
return term->text;
|
84
|
-
}
|
85
|
-
|
86
|
-
static VALUE
|
87
|
-
frt_term_set_text(VALUE self, VALUE rtext)
|
88
|
-
{
|
89
|
-
GET_TERM;
|
90
|
-
term->text = rb_obj_as_string(rtext);
|
91
|
-
return Qnil;
|
92
|
-
}
|
93
|
-
|
94
|
-
static VALUE
|
95
|
-
frt_term_get_field(VALUE self)
|
96
|
-
{
|
97
|
-
GET_TERM;
|
98
|
-
return term->field;
|
99
|
-
}
|
100
|
-
|
101
|
-
static VALUE
|
102
|
-
frt_term_set_field(VALUE self, VALUE rfield)
|
103
|
-
{
|
104
|
-
GET_TERM;
|
105
|
-
term->field = rb_obj_as_string(rfield);
|
106
|
-
return Qnil;
|
107
|
-
}
|
108
|
-
|
109
|
-
VALUE
|
110
|
-
frt_term_to_s(VALUE self)
|
111
|
-
{
|
112
|
-
int tlen, flen;
|
113
|
-
char *res;
|
114
|
-
GET_TERM;
|
115
|
-
tlen = RSTRING(term->text)->len;
|
116
|
-
flen = RSTRING(term->field)->len;
|
117
|
-
res = alloca(flen + tlen + 1);
|
118
|
-
|
119
|
-
MEMCPY(res, StringValuePtr(term->field), char, flen);
|
120
|
-
res[flen] = ':';
|
121
|
-
MEMCPY(res + flen + 1, StringValuePtr(term->text), char, tlen);
|
122
|
-
return rb_str_new(res, tlen + flen + 1 );
|
123
|
-
}
|
124
|
-
|
125
|
-
inline int
|
126
|
-
frt_term_cmp(RTerm *t1, RTerm *t2)
|
127
|
-
{
|
128
|
-
int comp = rb_str_cmp(t1->field, t2->field);
|
129
|
-
if (comp == 0) {
|
130
|
-
comp = rb_str_cmp(t1->text, t2->text);
|
131
|
-
}
|
132
|
-
return comp;
|
133
|
-
}
|
134
|
-
|
135
|
-
int
|
136
|
-
frt_term_compare_to_int(VALUE self, VALUE rother)
|
137
|
-
{
|
138
|
-
RTerm *other;
|
139
|
-
GET_TERM;
|
140
|
-
Data_Get_Struct(rother, RTerm, other);
|
141
|
-
return frt_term_cmp(term, other);
|
142
|
-
}
|
143
|
-
|
144
|
-
VALUE
|
145
|
-
frt_term_lt(VALUE self, VALUE rother)
|
146
|
-
{
|
147
|
-
return frt_term_compare_to_int(self, rother) < 0 ? Qtrue : Qfalse;
|
148
|
-
}
|
149
|
-
|
150
|
-
VALUE
|
151
|
-
frt_term_gt(VALUE self, VALUE rother)
|
152
|
-
{
|
153
|
-
return frt_term_compare_to_int(self, rother) > 0 ? Qtrue : Qfalse;
|
154
|
-
}
|
155
|
-
|
156
|
-
VALUE
|
157
|
-
frt_term_le(VALUE self, VALUE rother)
|
158
|
-
{
|
159
|
-
return frt_term_compare_to_int(self, rother) <= 0 ? Qtrue : Qfalse;
|
160
|
-
}
|
161
|
-
|
162
|
-
VALUE
|
163
|
-
frt_term_ge(VALUE self, VALUE rother)
|
164
|
-
{
|
165
|
-
return frt_term_compare_to_int(self, rother) >= 0 ? Qtrue : Qfalse;
|
166
|
-
}
|
167
|
-
|
168
|
-
VALUE
|
169
|
-
frt_term_eq(VALUE self, VALUE rother)
|
170
|
-
{
|
171
|
-
if (rother == Qnil)
|
172
|
-
return Qfalse;
|
173
|
-
return frt_term_compare_to_int(self, rother) == 0 ? Qtrue : Qfalse;
|
174
|
-
}
|
175
|
-
|
176
|
-
|
177
|
-
static VALUE
|
178
|
-
frt_term_compare_to(VALUE self, VALUE other)
|
179
|
-
{
|
180
|
-
return INT2FIX(frt_term_compare_to_int(self, other));
|
181
|
-
}
|
182
|
-
|
183
|
-
static VALUE
|
184
|
-
frt_term_hash(VALUE self)
|
185
|
-
{
|
186
|
-
GET_TERM;
|
187
|
-
return INT2FIX(rb_str_hash(term->field) + rb_str_hash(term->text));
|
188
|
-
}
|
189
|
-
|
190
|
-
/****************************************************************************
|
191
|
-
*
|
192
|
-
* Init Function
|
193
|
-
*
|
194
|
-
****************************************************************************/
|
195
|
-
|
196
|
-
void
|
197
|
-
Init_term(void)
|
198
|
-
{
|
199
|
-
/* Term */
|
200
|
-
cTerm = rb_define_class_under(mIndex, "Term", rb_cObject);
|
201
|
-
rb_define_alloc_func(cTerm, frt_term_alloc);
|
202
|
-
rb_include_module(cTerm, rb_mComparable);
|
203
|
-
|
204
|
-
rb_define_method(cTerm, "initialize", frt_term_init, 2);
|
205
|
-
rb_define_method(cTerm, "set!", frt_term_init, 2);
|
206
|
-
rb_define_method(cTerm, "to_s", frt_term_to_s, 0);
|
207
|
-
rb_define_method(cTerm, "<=>", frt_term_compare_to, 1);
|
208
|
-
rb_define_method(cTerm, "<", frt_term_lt, 1);
|
209
|
-
rb_define_method(cTerm, ">", frt_term_gt, 1);
|
210
|
-
rb_define_method(cTerm, "<=", frt_term_le, 1);
|
211
|
-
rb_define_method(cTerm, ">=", frt_term_ge, 1);
|
212
|
-
rb_define_method(cTerm, "eql?", frt_term_eq, 1);
|
213
|
-
rb_define_method(cTerm, "==", frt_term_eq, 1);
|
214
|
-
rb_define_method(cTerm, "text", frt_term_get_text, 0);
|
215
|
-
rb_define_method(cTerm, "text=", frt_term_set_text, 1);
|
216
|
-
rb_define_method(cTerm, "field", frt_term_get_field, 0);
|
217
|
-
rb_define_method(cTerm, "field=", frt_term_set_field, 1);
|
218
|
-
rb_define_method(cTerm, "hash", frt_term_hash, 0);
|
219
|
-
}
|
data/ext/term.c
DELETED
@@ -1,820 +0,0 @@
|
|
1
|
-
#include "index.h"
|
2
|
-
#include "helper.h"
|
3
|
-
#include "hash.h"
|
4
|
-
#include <string.h>
|
5
|
-
|
6
|
-
static char * const FORMAT_VERSION_ERROR_MSG = "Unknown format version";
|
7
|
-
static char * const TERM_ORDER_ERROR_MSG = "term out of order";
|
8
|
-
static char * const FP_ORDER_ERROR_MSG = "freq pointer out of order";
|
9
|
-
static char * const PP_ORDER_ERROR_MSG = "prox pointer out of order";
|
10
|
-
|
11
|
-
/****************************************************************************
|
12
|
-
*
|
13
|
-
* Term
|
14
|
-
*
|
15
|
-
****************************************************************************/
|
16
|
-
|
17
|
-
Term *term_clone(Term *term)
|
18
|
-
{
|
19
|
-
Term *t = ALLOC(Term);
|
20
|
-
|
21
|
-
t->field = term->field;
|
22
|
-
t->text = estrdup(term->text);
|
23
|
-
return t;
|
24
|
-
}
|
25
|
-
|
26
|
-
Term *term_create(const char *field, char *text)
|
27
|
-
{
|
28
|
-
Term *t = ALLOC(Term);
|
29
|
-
|
30
|
-
t->field = (char *)field;
|
31
|
-
t->text = estrdup(text);
|
32
|
-
return t;
|
33
|
-
}
|
34
|
-
|
35
|
-
void term_destroy(Term *self)
|
36
|
-
{
|
37
|
-
free(self->text);
|
38
|
-
free(self);
|
39
|
-
}
|
40
|
-
|
41
|
-
int term_cmp(void *t1, void *t2)
|
42
|
-
{
|
43
|
-
int res = strcmp(((Term *)t1)->field, ((Term *)t2)->field);
|
44
|
-
if (res != 0) {
|
45
|
-
return res;
|
46
|
-
} else {
|
47
|
-
return strcmp(((Term *)t1)->text, ((Term *)t2)->text);
|
48
|
-
}
|
49
|
-
}
|
50
|
-
|
51
|
-
int term_eq(const void *t1, const void *t2)
|
52
|
-
{
|
53
|
-
return (strcmp(((Term *)t1)->text, ((Term *)t2)->text)) == 0 &&
|
54
|
-
(strcmp(((Term *)t1)->field, ((Term *)t2)->field) == 0);
|
55
|
-
}
|
56
|
-
|
57
|
-
unsigned int term_hash(const void *t)
|
58
|
-
{
|
59
|
-
return str_hash(((Term *)t)->text) * str_hash(((Term *)t)->field);
|
60
|
-
}
|
61
|
-
|
62
|
-
char *term_to_s(Term *term)
|
63
|
-
{
|
64
|
-
char *string = ALLOC_N(char, strlen(term->field) + strlen(term->text) + 2);
|
65
|
-
sprintf(string, "%s:%s", term->field, term->text);
|
66
|
-
return string;
|
67
|
-
}
|
68
|
-
|
69
|
-
/****************************************************************************
|
70
|
-
*
|
71
|
-
* TermBuffer
|
72
|
-
*
|
73
|
-
****************************************************************************/
|
74
|
-
|
75
|
-
void tb_reset(TermBuffer *tb)
|
76
|
-
{
|
77
|
-
tb->field = (char *)EMPTY_STRING;
|
78
|
-
tb->text[0] = '\0';
|
79
|
-
}
|
80
|
-
|
81
|
-
TermBuffer *tb_create()
|
82
|
-
{
|
83
|
-
TermBuffer *tb = ALLOC(TermBuffer);
|
84
|
-
tb->field = (char *)EMPTY_STRING;
|
85
|
-
tb->text[0] = '\0';
|
86
|
-
return tb;
|
87
|
-
}
|
88
|
-
|
89
|
-
void tb_destroy(TermBuffer *tb)
|
90
|
-
{
|
91
|
-
free(tb);
|
92
|
-
}
|
93
|
-
|
94
|
-
TermBuffer *tb_set_term(TermBuffer *tb, Term *t)
|
95
|
-
{
|
96
|
-
tb->field = t->field;
|
97
|
-
strcpy(tb->text, t->text);
|
98
|
-
return tb;
|
99
|
-
}
|
100
|
-
|
101
|
-
Term *tb_get_term(TermBuffer *tb)
|
102
|
-
{
|
103
|
-
return term_create(tb->field, tb->text);
|
104
|
-
}
|
105
|
-
|
106
|
-
int tb_cmp(TermBuffer *tb1, TermBuffer *tb2)
|
107
|
-
{
|
108
|
-
int res;
|
109
|
-
if ((tb1->field != tb2->field) &&
|
110
|
-
(0 != (res = strcmp(tb1->field, tb2->field)))) {
|
111
|
-
return res;
|
112
|
-
} else {
|
113
|
-
return strcmp(tb1->text, tb2->text);
|
114
|
-
}
|
115
|
-
}
|
116
|
-
|
117
|
-
int tb_term_cmp(TermBuffer *tb, Term *t)
|
118
|
-
{
|
119
|
-
int res = strcmp(tb->field, t->field);
|
120
|
-
if (res != 0) {
|
121
|
-
return res;
|
122
|
-
} else {
|
123
|
-
return strcmp(tb->text, t->text);
|
124
|
-
}
|
125
|
-
}
|
126
|
-
|
127
|
-
TermBuffer *tb_cpy(TermBuffer *tb1, TermBuffer *tb2)
|
128
|
-
{
|
129
|
-
tb1->field = tb2->field;
|
130
|
-
strcpy(tb1->text, tb2->text);
|
131
|
-
return tb1;
|
132
|
-
}
|
133
|
-
|
134
|
-
TermBuffer *tb_read(TermBuffer *tb, InStream *is, FieldInfos *fis)
|
135
|
-
{
|
136
|
-
signed int fnum;
|
137
|
-
int start = (int)is_read_vint(is);
|
138
|
-
int length = (int)is_read_vint(is);
|
139
|
-
int total_length = start + length;
|
140
|
-
is_read_bytes(is, (uchar *)tb->text, start, length);
|
141
|
-
tb->text[total_length] = '\0';
|
142
|
-
fnum = (signed int)is_read_vint(is);
|
143
|
-
if (fnum < 0)
|
144
|
-
tb->field = (char *)EMPTY_STRING;
|
145
|
-
else
|
146
|
-
tb->field = fis->by_number[fnum]->name;
|
147
|
-
return tb;
|
148
|
-
}
|
149
|
-
|
150
|
-
/****************************************************************************
|
151
|
-
*
|
152
|
-
* TermInfo
|
153
|
-
*
|
154
|
-
****************************************************************************/
|
155
|
-
|
156
|
-
TermInfo *ti_create(int doc_freq, int freq_pointer, int prox_pointer, int skip_offset)
|
157
|
-
{
|
158
|
-
TermInfo *ti = ALLOC(TermInfo);
|
159
|
-
ti->doc_freq = doc_freq;
|
160
|
-
ti->freq_pointer = freq_pointer;
|
161
|
-
ti->prox_pointer = prox_pointer;
|
162
|
-
ti->skip_offset = skip_offset;
|
163
|
-
return ti;
|
164
|
-
}
|
165
|
-
|
166
|
-
TermInfo *ti_set(TermInfo *ti, int doc_freq, int freq_pointer, int prox_pointer, int skip_offset)
|
167
|
-
{
|
168
|
-
ti->doc_freq = doc_freq;
|
169
|
-
ti->freq_pointer = freq_pointer;
|
170
|
-
ti->prox_pointer = prox_pointer;
|
171
|
-
ti->skip_offset = skip_offset;
|
172
|
-
return ti;
|
173
|
-
}
|
174
|
-
|
175
|
-
void ti_destroy(TermInfo *ti)
|
176
|
-
{
|
177
|
-
free(ti);
|
178
|
-
}
|
179
|
-
|
180
|
-
TermInfo *ti_cpy(TermInfo *ti, TermInfo *other)
|
181
|
-
{
|
182
|
-
memcpy(ti, other, sizeof(TermInfo));
|
183
|
-
return ti;
|
184
|
-
}
|
185
|
-
|
186
|
-
TermInfo *ti_clone(TermInfo *other)
|
187
|
-
{
|
188
|
-
return ti_create(other->doc_freq,
|
189
|
-
other->freq_pointer, other->prox_pointer, other->skip_offset);
|
190
|
-
}
|
191
|
-
|
192
|
-
int ti_eq(TermInfo *ti, TermInfo *other)
|
193
|
-
{
|
194
|
-
return (memcmp(ti, other, sizeof(TermInfo)) == 0);
|
195
|
-
}
|
196
|
-
|
197
|
-
/****************************************************************************
|
198
|
-
*
|
199
|
-
* TermEnum
|
200
|
-
*
|
201
|
-
****************************************************************************/
|
202
|
-
|
203
|
-
TermEnum *te_create()
|
204
|
-
{
|
205
|
-
TermEnum *te = ALLOC(TermEnum);
|
206
|
-
te->tb_curr = tb_create();
|
207
|
-
te->tb_prev = tb_create();
|
208
|
-
te->ti_curr = ti_create(0, 0, 0, 0);
|
209
|
-
return te;
|
210
|
-
}
|
211
|
-
|
212
|
-
void te_destroy(TermEnum *te)
|
213
|
-
{
|
214
|
-
tb_destroy(te->tb_curr);
|
215
|
-
tb_destroy(te->tb_prev);
|
216
|
-
ti_destroy(te->ti_curr);
|
217
|
-
free(te);
|
218
|
-
}
|
219
|
-
|
220
|
-
Term *te_get_term(TermEnum *te)
|
221
|
-
{
|
222
|
-
return tb_get_term(te->tb_curr);
|
223
|
-
}
|
224
|
-
|
225
|
-
TermInfo *te_get_ti(TermEnum *te)
|
226
|
-
{
|
227
|
-
TermInfo *ti = te->ti_curr;
|
228
|
-
return ti_create(ti->doc_freq, ti->freq_pointer, ti->prox_pointer, ti->skip_offset);
|
229
|
-
}
|
230
|
-
|
231
|
-
TermBuffer *te_skip_to(TermEnum *te, Term *t)
|
232
|
-
{
|
233
|
-
TermBuffer *tb_curr;
|
234
|
-
if (tb_term_cmp(te->tb_curr, t) == 0)
|
235
|
-
return te->tb_curr;
|
236
|
-
|
237
|
-
while (((tb_curr = te->next(te)) != NULL) &&
|
238
|
-
(tb_term_cmp(tb_curr, t) < 0)) {
|
239
|
-
}
|
240
|
-
return tb_curr;
|
241
|
-
}
|
242
|
-
|
243
|
-
/****************************************************************************
|
244
|
-
*
|
245
|
-
* SegmentTermEnum
|
246
|
-
*
|
247
|
-
****************************************************************************/
|
248
|
-
|
249
|
-
#define GET_STE SegmentTermEnum *ste = (SegmentTermEnum *)te->data
|
250
|
-
|
251
|
-
TermBuffer *ste_next(TermEnum *te)
|
252
|
-
{
|
253
|
-
GET_STE;
|
254
|
-
TermInfo *ti;
|
255
|
-
InStream *is = ste->is;
|
256
|
-
ste->pos++;
|
257
|
-
if (ste->pos > ste->size - 1) {
|
258
|
-
tb_reset(te->tb_curr);
|
259
|
-
return NULL;
|
260
|
-
}
|
261
|
-
|
262
|
-
tb_cpy(te->tb_prev, te->tb_curr);
|
263
|
-
tb_read(te->tb_curr, is, ste->fis);
|
264
|
-
|
265
|
-
ti = te->ti_curr;
|
266
|
-
ti->doc_freq = (int)is_read_vint(is); /* read doc freq */
|
267
|
-
ti->freq_pointer += (int)is_read_vint(is);/* read freq pointer */
|
268
|
-
ti->prox_pointer += (int)is_read_vint(is);/* read prox pointer */
|
269
|
-
|
270
|
-
if (ste->format == -1) {
|
271
|
-
/* just read skip_offset in order to increment file pointer
|
272
|
-
* value is never used since skip_to is switched off */
|
273
|
-
if (!ste->is_index) {
|
274
|
-
if (ti->doc_freq > ste->format_m1skip_interval) {
|
275
|
-
ti->skip_offset = (int)is_read_vint(is);
|
276
|
-
}
|
277
|
-
}
|
278
|
-
} else {
|
279
|
-
if (ti->doc_freq >= ste->skip_interval) {
|
280
|
-
ti->skip_offset = (int)is_read_vint(is);
|
281
|
-
}
|
282
|
-
}
|
283
|
-
|
284
|
-
if (ste->is_index) {
|
285
|
-
ste->index_pointer += (int)is_read_vint(is); /* read index pointer */
|
286
|
-
}
|
287
|
-
|
288
|
-
return te->tb_curr;
|
289
|
-
}
|
290
|
-
|
291
|
-
TermEnum *ste_clone(TermEnum *other_te);
|
292
|
-
|
293
|
-
TermEnum *ste_allocate()
|
294
|
-
{
|
295
|
-
TermEnum *te = te_create();
|
296
|
-
SegmentTermEnum *ste;
|
297
|
-
|
298
|
-
te->next = &ste_next;
|
299
|
-
te->close = &ste_close;
|
300
|
-
te->clone = &ste_clone;
|
301
|
-
ste = ALLOC(SegmentTermEnum);
|
302
|
-
te->data = ste;
|
303
|
-
return te;
|
304
|
-
}
|
305
|
-
|
306
|
-
TermEnum *ste_clone(TermEnum *other_te)
|
307
|
-
{
|
308
|
-
SegmentTermEnum *other_ste = (SegmentTermEnum *)other_te->data;
|
309
|
-
TermEnum *te = ste_allocate();
|
310
|
-
SegmentTermEnum *ste = (SegmentTermEnum *)te->data;
|
311
|
-
|
312
|
-
memcpy(ste, other_ste, sizeof(SegmentTermEnum));
|
313
|
-
ste->is = is_clone(other_ste->is);
|
314
|
-
tb_cpy(te->tb_curr, other_te->tb_curr);
|
315
|
-
tb_cpy(te->tb_prev, other_te->tb_prev);
|
316
|
-
ti_cpy(te->ti_curr, other_te->ti_curr);
|
317
|
-
return te;
|
318
|
-
}
|
319
|
-
|
320
|
-
void ste_close(TermEnum *te)
|
321
|
-
{
|
322
|
-
GET_STE;
|
323
|
-
is_close(ste->is);
|
324
|
-
free(ste);
|
325
|
-
te->data = NULL;
|
326
|
-
te_destroy(te);
|
327
|
-
}
|
328
|
-
|
329
|
-
TermEnum *ste_create(InStream *is, FieldInfos *fis, int is_index)
|
330
|
-
{
|
331
|
-
TermEnum *te = ste_allocate();
|
332
|
-
GET_STE;
|
333
|
-
int first_int;
|
334
|
-
|
335
|
-
ste->fis = fis;
|
336
|
-
ste->is_index = is_index;
|
337
|
-
ste->is = is;
|
338
|
-
ste->pos = -1;
|
339
|
-
ste->index_pointer = 0;
|
340
|
-
ste->format_m1skip_interval = -1;
|
341
|
-
|
342
|
-
first_int = (int)is_read_int(is);
|
343
|
-
|
344
|
-
if (first_int >= 0) {
|
345
|
-
/* original-format file, without explicit format version number */
|
346
|
-
ste->format = 0;
|
347
|
-
ste->size = first_int;
|
348
|
-
|
349
|
-
/* back-compatible settings */
|
350
|
-
ste->index_interval = 128;
|
351
|
-
ste->skip_interval = INT_MAX; /* switch off skip_to optimization */
|
352
|
-
|
353
|
-
} else {
|
354
|
-
/* check that it is a format we can understand */
|
355
|
-
if (first_int < TERM_INFO_FORMAT)
|
356
|
-
RAISE(EXCEPTION, FORMAT_VERSION_ERROR_MSG);
|
357
|
-
|
358
|
-
/* we have a format version number */
|
359
|
-
ste->format = first_int;
|
360
|
-
|
361
|
-
|
362
|
-
ste->size = (int)is_read_long(is); /* read the size */
|
363
|
-
|
364
|
-
if (ste->format == -1) {
|
365
|
-
if (!ste->is_index) {
|
366
|
-
ste->index_interval = is_read_int(is);
|
367
|
-
ste->format_m1skip_interval = is_read_int(is);
|
368
|
-
}
|
369
|
-
/* switch off skip_to optimization for file format prior to
|
370
|
-
* 1.4rc2 in order to avoid a bug in skip_to implementation
|
371
|
-
* of these versions */
|
372
|
-
ste->skip_interval = INT_MAX;
|
373
|
-
} else {
|
374
|
-
ste->index_interval = is_read_int(is);
|
375
|
-
ste->skip_interval = is_read_int(is);
|
376
|
-
}
|
377
|
-
}
|
378
|
-
return te;
|
379
|
-
}
|
380
|
-
|
381
|
-
void ste_seek(TermEnum *te, int pointer, int pos, Term *t, TermInfo *ti)
|
382
|
-
{
|
383
|
-
GET_STE;
|
384
|
-
is_seek(ste->is, pointer);
|
385
|
-
ste->pos = pos;
|
386
|
-
tb_set_term(te->tb_curr, t);
|
387
|
-
tb_reset(te->tb_prev);
|
388
|
-
ti_cpy(te->ti_curr, ti);
|
389
|
-
}
|
390
|
-
|
391
|
-
TermInfo *ste_scan_for_term_info(TermEnum *te, Term *t)
|
392
|
-
{
|
393
|
-
te_skip_to(te, t);
|
394
|
-
|
395
|
-
if (tb_term_cmp(te->tb_curr, t) == 0) {
|
396
|
-
return te_get_ti(te);
|
397
|
-
} else {
|
398
|
-
return NULL;
|
399
|
-
}
|
400
|
-
}
|
401
|
-
|
402
|
-
Term *ste_scan_for_term(TermEnum *te, int pos)
|
403
|
-
{
|
404
|
-
GET_STE;
|
405
|
-
while (ste->pos < pos) {
|
406
|
-
if (ste_next(te) == NULL)
|
407
|
-
return NULL;
|
408
|
-
}
|
409
|
-
|
410
|
-
return te_get_term(te);
|
411
|
-
}
|
412
|
-
|
413
|
-
/****************************************************************************
|
414
|
-
*
|
415
|
-
* MultiTermEnum
|
416
|
-
*
|
417
|
-
****************************************************************************/
|
418
|
-
|
419
|
-
#define GET_MTE MultiTermEnum *mte = (MultiTermEnum *)te->data
|
420
|
-
|
421
|
-
TermBuffer *mte_next(TermEnum *te)
|
422
|
-
{
|
423
|
-
GET_MTE;
|
424
|
-
SegmentMergeInfo *top = (SegmentMergeInfo *)pq_top(mte->smi_queue);
|
425
|
-
|
426
|
-
if (top == NULL) {
|
427
|
-
tb_reset(te->tb_curr);
|
428
|
-
return false;
|
429
|
-
}
|
430
|
-
|
431
|
-
tb_cpy(te->tb_prev, te->tb_curr);
|
432
|
-
tb_cpy(te->tb_curr, top->tb);
|
433
|
-
|
434
|
-
te->ti_curr->doc_freq = 0;
|
435
|
-
|
436
|
-
while ((top != NULL) && (tb_cmp(te->tb_curr, top->tb) == 0)) {
|
437
|
-
pq_pop(mte->smi_queue);
|
438
|
-
te->ti_curr->doc_freq += top->te->ti_curr->doc_freq;/* increment freq */
|
439
|
-
if (smi_next(top)) {
|
440
|
-
pq_push(mte->smi_queue, top); /* restore queue */
|
441
|
-
} else {
|
442
|
-
smi_destroy(top); /* done with a segment */
|
443
|
-
}
|
444
|
-
top = (SegmentMergeInfo *)pq_top(mte->smi_queue);
|
445
|
-
}
|
446
|
-
return te->tb_curr;
|
447
|
-
}
|
448
|
-
|
449
|
-
void mte_close(TermEnum *te)
|
450
|
-
{
|
451
|
-
GET_MTE;
|
452
|
-
pq_clear(mte->smi_queue);
|
453
|
-
pq_destroy(mte->smi_queue);
|
454
|
-
free(mte);
|
455
|
-
te_destroy(te);
|
456
|
-
}
|
457
|
-
|
458
|
-
TermEnum *mte_clone(TermEnum *te)
|
459
|
-
{
|
460
|
-
RAISE(UNSUPPORTED_ERROR, UNSUPPORTED_ERROR_MSG);
|
461
|
-
return NULL;
|
462
|
-
}
|
463
|
-
|
464
|
-
TermEnum *mte_create(IndexReader **readers, int *starts, int rcnt, Term *t)
|
465
|
-
{
|
466
|
-
int i;
|
467
|
-
IndexReader *reader;
|
468
|
-
TermEnum *sub_te;
|
469
|
-
MultiTermEnum *mte = ALLOC(MultiTermEnum);
|
470
|
-
TermEnum *te = te_create();
|
471
|
-
te->next = &mte_next;
|
472
|
-
te->clone = &mte_clone;
|
473
|
-
te->close = &mte_close;
|
474
|
-
|
475
|
-
te->data = mte;
|
476
|
-
|
477
|
-
mte->smi_queue = pq_create(rcnt, (lt_ft)&smi_lt);
|
478
|
-
mte->smi_queue->free_elem = (free_ft)&smi_destroy;
|
479
|
-
|
480
|
-
for (i = 0; i < rcnt; i++) {
|
481
|
-
SegmentMergeInfo *smi;
|
482
|
-
reader = readers[i];
|
483
|
-
|
484
|
-
if (t != NULL) {
|
485
|
-
sub_te = reader->terms_from(reader, t);
|
486
|
-
} else {
|
487
|
-
sub_te = reader->terms(reader);
|
488
|
-
}
|
489
|
-
|
490
|
-
smi = smi_create(starts[i], sub_te, reader);
|
491
|
-
if (((t == NULL) && smi_next(smi)) ||
|
492
|
-
(sub_te->tb_curr->field != (char *)EMPTY_STRING)) {
|
493
|
-
pq_push(mte->smi_queue, smi); /* initialize queue */
|
494
|
-
} else {
|
495
|
-
smi_destroy(smi);
|
496
|
-
}
|
497
|
-
}
|
498
|
-
|
499
|
-
if ((t != NULL) && (mte->smi_queue->count > 0)) {
|
500
|
-
mte_next(te);
|
501
|
-
}
|
502
|
-
|
503
|
-
return te;
|
504
|
-
}
|
505
|
-
|
506
|
-
/****************************************************************************
|
507
|
-
*
|
508
|
-
* TermInfosWriter
|
509
|
-
*
|
510
|
-
****************************************************************************/
|
511
|
-
|
512
|
-
const Term EmptyTerm = {"", ""};
|
513
|
-
|
514
|
-
TermInfosWriter *tiw_open_internal(Store *store,
|
515
|
-
char *segment,
|
516
|
-
FieldInfos *fis,
|
517
|
-
int interval,
|
518
|
-
int is_index)
|
519
|
-
{
|
520
|
-
char fname[SEGMENT_NAME_MAX_LENGTH];
|
521
|
-
TermInfosWriter *tiw = ALLOC(TermInfosWriter);
|
522
|
-
OutStream *os;
|
523
|
-
|
524
|
-
tiw->index_interval = interval;
|
525
|
-
tiw->skip_interval = 16;
|
526
|
-
tiw->last_index_pointer = 0;
|
527
|
-
tiw->last_term = (Term *)&EmptyTerm;
|
528
|
-
tiw->last_term_info = ti_create(0,0,0,0);
|
529
|
-
tiw->size = 0;
|
530
|
-
tiw->is_index = is_index;
|
531
|
-
tiw->fis = fis;
|
532
|
-
tiw->curr_field = NULL;
|
533
|
-
tiw->curr_field_num = -1;
|
534
|
-
|
535
|
-
strcpy(fname, segment);
|
536
|
-
strcat(fname, (is_index ? ".tii" : ".tis"));
|
537
|
-
os = tiw->os = store->create_output(store, fname);
|
538
|
-
os_write_int(os, TERM_INFO_FORMAT); /* write format */
|
539
|
-
os_write_long(os, 0); /* leave space for size */
|
540
|
-
os_write_int(os, tiw->index_interval); /* write index_interval */
|
541
|
-
os_write_int(os, tiw->skip_interval); /* write skip_interval */
|
542
|
-
if (!is_index) {
|
543
|
-
tiw->other = tiw_open_internal(store, segment, fis, interval, true);
|
544
|
-
tiw->other->other = tiw;
|
545
|
-
}
|
546
|
-
return tiw;
|
547
|
-
}
|
548
|
-
|
549
|
-
TermInfosWriter *tiw_open(Store *store, char *segment, FieldInfos *fis, int interval)
|
550
|
-
{
|
551
|
-
return tiw_open_internal(store, segment, fis, interval, false);
|
552
|
-
}
|
553
|
-
|
554
|
-
void tiw_write_term(TermInfosWriter *tiw, OutStream *os, Term *t)
|
555
|
-
{
|
556
|
-
//printf("%s, %s\n", tiw->last_term->text, t->text);
|
557
|
-
int start = hlp_string_diff(tiw->last_term->text, t->text);
|
558
|
-
int length = (int)strlen(t->text) - start;
|
559
|
-
|
560
|
-
os_write_vint(os, start); /* write shared prefix length */
|
561
|
-
os_write_vint(os, length); /* write delta length */
|
562
|
-
os_write_chars(os, t->text, start, length); /* write delta chars */
|
563
|
-
if (tiw->curr_field != t->field) {
|
564
|
-
tiw->curr_field = t->field;
|
565
|
-
tiw->curr_field_num = fis_get_number(tiw->fis, t->field);
|
566
|
-
}
|
567
|
-
os_write_vint(os, tiw->curr_field_num);
|
568
|
-
tiw->last_term = t;
|
569
|
-
}
|
570
|
-
|
571
|
-
void tiw_add(TermInfosWriter *tiw, Term *t, TermInfo *ti)
|
572
|
-
{
|
573
|
-
#ifdef DEBUG
|
574
|
-
if (tiw->is_index && term_cmp(tiw->last_term, t) > 0) {
|
575
|
-
RAISE(STATE_ERROR, TERM_ORDER_ERROR_MSG);
|
576
|
-
}
|
577
|
-
if (ti->freq_pointer < tiw->last_term_info->freq_pointer) {
|
578
|
-
RAISE(STATE_ERROR, FP_ORDER_ERROR_MSG);
|
579
|
-
}
|
580
|
-
if (ti->prox_pointer < tiw->last_term_info->prox_pointer) {
|
581
|
-
RAISE(STATE_ERROR, PP_ORDER_ERROR_MSG);
|
582
|
-
}
|
583
|
-
#endif
|
584
|
-
|
585
|
-
if (!tiw->is_index && (tiw->size % tiw->index_interval) == 0) {
|
586
|
-
/* add an index term */
|
587
|
-
tiw_add(tiw->other, tiw->last_term, tiw->last_term_info);
|
588
|
-
}
|
589
|
-
|
590
|
-
tiw_write_term(tiw, tiw->os, t); /* write term */
|
591
|
-
os_write_vint(tiw->os, ti->doc_freq); /* write doc freq */
|
592
|
-
os_write_vint(tiw->os, ti->freq_pointer - tiw->last_term_info->freq_pointer);
|
593
|
-
os_write_vint(tiw->os, ti->prox_pointer - tiw->last_term_info->prox_pointer);
|
594
|
-
|
595
|
-
if (ti->doc_freq >= tiw->skip_interval) {
|
596
|
-
os_write_vint(tiw->os, ti->skip_offset);
|
597
|
-
}
|
598
|
-
|
599
|
-
if (tiw->is_index) {
|
600
|
-
OutStream *other_os = tiw->other->os;
|
601
|
-
int other_pos = os_pos(other_os);
|
602
|
-
os_write_vint(tiw->os, other_pos - tiw->last_index_pointer);
|
603
|
-
tiw->last_index_pointer = other_pos; /* write pointer */
|
604
|
-
}
|
605
|
-
|
606
|
-
ti_cpy(tiw->last_term_info, ti);
|
607
|
-
tiw->size++;
|
608
|
-
}
|
609
|
-
|
610
|
-
void tiw_close(TermInfosWriter *tiw)
|
611
|
-
{
|
612
|
-
OutStream *os = tiw->os;
|
613
|
-
os_seek(os, 4); /* write @size after format */
|
614
|
-
os_write_long(os, tiw->size);
|
615
|
-
os_close(os);
|
616
|
-
|
617
|
-
if (!tiw->is_index)
|
618
|
-
tiw_close(tiw->other);
|
619
|
-
|
620
|
-
ti_destroy(tiw->last_term_info);
|
621
|
-
free(tiw);
|
622
|
-
}
|
623
|
-
|
624
|
-
/****************************************************************************
|
625
|
-
*
|
626
|
-
* TermInfosReader
|
627
|
-
*
|
628
|
-
****************************************************************************/
|
629
|
-
|
630
|
-
void tir_close(TermInfosReader *tir)
|
631
|
-
{
|
632
|
-
int i;
|
633
|
-
if (tir->index_terms != NULL) {
|
634
|
-
for (i = 0; i < tir->index_size; i++) {
|
635
|
-
term_destroy(tir->index_terms[i]);
|
636
|
-
ti_destroy(tir->index_term_infos[i]);
|
637
|
-
}
|
638
|
-
free(tir->index_terms);
|
639
|
-
free(tir->index_term_infos);
|
640
|
-
free(tir->index_pointers);
|
641
|
-
}
|
642
|
-
if (tir->orig_te) tir->orig_te->close(tir->orig_te);
|
643
|
-
thread_key_delete(tir->thread_te);
|
644
|
-
ary_destroy(tir->te_bucket);
|
645
|
-
if (tir->index_te) tir->index_te->close(tir->index_te);
|
646
|
-
mutex_destroy(&tir->mutex);
|
647
|
-
free(tir);
|
648
|
-
}
|
649
|
-
|
650
|
-
TermInfosReader *tir_open(Store *store, char *segment, FieldInfos *fis)
|
651
|
-
{
|
652
|
-
SegmentTermEnum *ste;
|
653
|
-
TermInfosReader *tir = ALLOC(TermInfosReader);
|
654
|
-
char fname[SEGMENT_NAME_MAX_LENGTH];
|
655
|
-
InStream *is;
|
656
|
-
|
657
|
-
mutex_init(&tir->mutex, NULL);
|
658
|
-
strcpy(fname, segment);
|
659
|
-
strcpy(fname + strlen(segment), ".tis");
|
660
|
-
is = store->open_input(store, fname);
|
661
|
-
tir->orig_te = ste_create(is, fis, false);
|
662
|
-
thread_key_create(&tir->thread_te, NULL);
|
663
|
-
tir->te_bucket = ary_create(1, (free_ft)tir->orig_te->close);
|
664
|
-
|
665
|
-
ste = tir->orig_te->data;
|
666
|
-
tir->size = ste->size;
|
667
|
-
tir->skip_interval = ste->skip_interval;
|
668
|
-
|
669
|
-
strcpy(fname + strlen(segment), ".tii");
|
670
|
-
is = store->open_input(store, fname);
|
671
|
-
tir->index_te = ste_create(is, fis, true);
|
672
|
-
tir->index_terms = NULL;
|
673
|
-
tir->index_term_infos = NULL;
|
674
|
-
tir->index_pointers = NULL;
|
675
|
-
return tir;
|
676
|
-
}
|
677
|
-
|
678
|
-
void tir_ensure_index_is_read(TermInfosReader *tir)
|
679
|
-
{
|
680
|
-
mutex_lock(&tir->mutex);
|
681
|
-
if (tir->index_terms == NULL) {
|
682
|
-
TermEnum *index_te;
|
683
|
-
SegmentTermEnum *ste;
|
684
|
-
int i = 0;
|
685
|
-
int index_size = ((SegmentTermEnum *)tir->index_te->data)->size;
|
686
|
-
tir->index_size = index_size;
|
687
|
-
|
688
|
-
tir->index_terms = ALLOC_N(Term *, index_size);
|
689
|
-
tir->index_term_infos = ALLOC_N(TermInfo *, index_size);
|
690
|
-
tir->index_pointers = ALLOC_N(int, index_size);
|
691
|
-
|
692
|
-
index_te = tir->index_te;
|
693
|
-
ste = index_te->data;
|
694
|
-
|
695
|
-
TRY
|
696
|
-
while (ste_next(index_te) != NULL) {
|
697
|
-
tir->index_terms[i] = te_get_term(index_te);
|
698
|
-
tir->index_term_infos[i] = te_get_ti(index_te);
|
699
|
-
tir->index_pointers[i] = ste->index_pointer;
|
700
|
-
i++;
|
701
|
-
}
|
702
|
-
XFINALLY
|
703
|
-
index_te->close(index_te);
|
704
|
-
tir->index_te = NULL;
|
705
|
-
XENDTRY
|
706
|
-
}
|
707
|
-
mutex_unlock(&tir->mutex);
|
708
|
-
}
|
709
|
-
|
710
|
-
static inline TermEnum *tir_enum(TermInfosReader *tir)
|
711
|
-
{
|
712
|
-
TermEnum *te;
|
713
|
-
if ((te = thread_getspecific(tir->thread_te)) == NULL) {
|
714
|
-
te = tir->orig_te->clone(tir->orig_te);
|
715
|
-
ary_append(tir->te_bucket, te);
|
716
|
-
thread_setspecific(tir->thread_te, te);
|
717
|
-
}
|
718
|
-
return te;
|
719
|
-
}
|
720
|
-
|
721
|
-
void tir_seek_enum(TermInfosReader *tir, int ind_offset)
|
722
|
-
{
|
723
|
-
TermEnum *te = tir_enum(tir);
|
724
|
-
SegmentTermEnum *ste = (SegmentTermEnum *)te->data;
|
725
|
-
ste_seek(te, tir->index_pointers[ind_offset],
|
726
|
-
(ind_offset * ste->index_interval) - 1,
|
727
|
-
tir->index_terms[ind_offset],
|
728
|
-
tir->index_term_infos[ind_offset]);
|
729
|
-
}
|
730
|
-
|
731
|
-
int tir_get_index_offset(TermInfosReader *tir, Term *t)
|
732
|
-
{
|
733
|
-
int lo = 0; /* binary search tir->index_terms[] */
|
734
|
-
int hi = tir->index_size - 1;
|
735
|
-
int mid, delta;
|
736
|
-
Term **index_terms = tir->index_terms;
|
737
|
-
|
738
|
-
while (hi >= lo) {
|
739
|
-
mid = (lo + hi) >> 1;
|
740
|
-
delta = term_cmp(t, index_terms[mid]);
|
741
|
-
if (delta < 0) {
|
742
|
-
hi = mid - 1;
|
743
|
-
} else if (delta > 0) {
|
744
|
-
lo = mid + 1;
|
745
|
-
} else {
|
746
|
-
return mid;
|
747
|
-
}
|
748
|
-
}
|
749
|
-
return hi;
|
750
|
-
}
|
751
|
-
|
752
|
-
TermInfo *tir_get_ti(TermInfosReader *tir, Term *t)
|
753
|
-
{
|
754
|
-
TermEnum *te;
|
755
|
-
SegmentTermEnum *ste;
|
756
|
-
if (tir->size == 0) {
|
757
|
-
return NULL;
|
758
|
-
}
|
759
|
-
|
760
|
-
tir_ensure_index_is_read(tir);
|
761
|
-
|
762
|
-
/* optimize sequential access: first try scanning cached enum w/o seeking */
|
763
|
-
te = tir_enum(tir);
|
764
|
-
ste = (SegmentTermEnum *)te->data;
|
765
|
-
if (ste->pos < ste->size && tb_term_cmp(te->tb_curr, t) <= 0) {
|
766
|
-
SegmentTermEnum *ste = (SegmentTermEnum *)te->data;
|
767
|
-
int enum_offset = (int)(ste->pos / ste->index_interval) + 1;
|
768
|
-
if (tir->index_size == enum_offset ||
|
769
|
-
term_cmp(t, tir->index_terms[enum_offset]) < 0) { /* but before end of block */
|
770
|
-
return ste_scan_for_term_info(te, t); /* no need to seek */
|
771
|
-
}
|
772
|
-
}
|
773
|
-
|
774
|
-
/* random-access: must seek */
|
775
|
-
tir_seek_enum(tir, tir_get_index_offset(tir, t));
|
776
|
-
return ste_scan_for_term_info(te, t);
|
777
|
-
}
|
778
|
-
|
779
|
-
Term *tir_get_term(TermInfosReader *tir, int pos)
|
780
|
-
{
|
781
|
-
if (tir->size == 0) {
|
782
|
-
return NULL;
|
783
|
-
} else {
|
784
|
-
TermEnum *te = tir_enum(tir);
|
785
|
-
SegmentTermEnum *ste = (SegmentTermEnum *)te->data;
|
786
|
-
if (pos >= ste->pos &&
|
787
|
-
pos < (ste->pos + ste->index_interval)) {
|
788
|
-
return ste_scan_for_term(te, pos); /* can avoid seek */
|
789
|
-
}
|
790
|
-
|
791
|
-
tir_seek_enum(tir, (int)(pos / ste->index_interval)); /* must seek */
|
792
|
-
return ste_scan_for_term(te, pos);
|
793
|
-
}
|
794
|
-
}
|
795
|
-
|
796
|
-
int tir_get_term_pos(TermInfosReader *tir, Term *t)
|
797
|
-
{
|
798
|
-
if (tir->size == 0) {
|
799
|
-
return -1;
|
800
|
-
} else {
|
801
|
-
TermEnum *te;
|
802
|
-
int ind_offset;
|
803
|
-
|
804
|
-
tir_ensure_index_is_read(tir);
|
805
|
-
|
806
|
-
ind_offset = tir_get_index_offset(tir, t);
|
807
|
-
tir_seek_enum(tir, ind_offset);
|
808
|
-
|
809
|
-
te = tir_enum(tir);
|
810
|
-
while ((tb_term_cmp(te->tb_curr, t) < 0) && (ste_next(te) != NULL)) {
|
811
|
-
}
|
812
|
-
|
813
|
-
if (tb_term_cmp(te->tb_curr, t) == 0) {
|
814
|
-
return ((SegmentTermEnum *)te->data)->pos;
|
815
|
-
} else {
|
816
|
-
return -1;
|
817
|
-
}
|
818
|
-
}
|
819
|
-
}
|
820
|
-
|