jk-ferret 0.11.8.2
Sign up to get free protection for your applications and to get access to all the features.
- data/CHANGELOG +24 -0
- data/MIT-LICENSE +20 -0
- data/README +90 -0
- data/RELEASE_CHANGES +137 -0
- data/RELEASE_NOTES +60 -0
- data/Rakefile +443 -0
- data/TODO +109 -0
- data/TUTORIAL +231 -0
- data/bin/ferret-browser +79 -0
- data/ext/BZLIB_blocksort.c +1094 -0
- data/ext/BZLIB_bzlib.c +1578 -0
- data/ext/BZLIB_compress.c +672 -0
- data/ext/BZLIB_crctable.c +104 -0
- data/ext/BZLIB_decompress.c +626 -0
- data/ext/BZLIB_huffman.c +205 -0
- data/ext/BZLIB_randtable.c +84 -0
- data/ext/STEMMER_api.c +66 -0
- data/ext/STEMMER_libstemmer.c +93 -0
- data/ext/STEMMER_stem_ISO_8859_1_danish.c +337 -0
- data/ext/STEMMER_stem_ISO_8859_1_dutch.c +624 -0
- data/ext/STEMMER_stem_ISO_8859_1_english.c +1117 -0
- data/ext/STEMMER_stem_ISO_8859_1_finnish.c +762 -0
- data/ext/STEMMER_stem_ISO_8859_1_french.c +1246 -0
- data/ext/STEMMER_stem_ISO_8859_1_german.c +503 -0
- data/ext/STEMMER_stem_ISO_8859_1_hungarian.c +1230 -0
- data/ext/STEMMER_stem_ISO_8859_1_italian.c +1065 -0
- data/ext/STEMMER_stem_ISO_8859_1_norwegian.c +297 -0
- data/ext/STEMMER_stem_ISO_8859_1_porter.c +749 -0
- data/ext/STEMMER_stem_ISO_8859_1_portuguese.c +1017 -0
- data/ext/STEMMER_stem_ISO_8859_1_spanish.c +1093 -0
- data/ext/STEMMER_stem_ISO_8859_1_swedish.c +307 -0
- data/ext/STEMMER_stem_ISO_8859_2_romanian.c +998 -0
- data/ext/STEMMER_stem_KOI8_R_russian.c +700 -0
- data/ext/STEMMER_stem_UTF_8_danish.c +339 -0
- data/ext/STEMMER_stem_UTF_8_dutch.c +634 -0
- data/ext/STEMMER_stem_UTF_8_english.c +1125 -0
- data/ext/STEMMER_stem_UTF_8_finnish.c +768 -0
- data/ext/STEMMER_stem_UTF_8_french.c +1256 -0
- data/ext/STEMMER_stem_UTF_8_german.c +509 -0
- data/ext/STEMMER_stem_UTF_8_hungarian.c +1234 -0
- data/ext/STEMMER_stem_UTF_8_italian.c +1073 -0
- data/ext/STEMMER_stem_UTF_8_norwegian.c +299 -0
- data/ext/STEMMER_stem_UTF_8_porter.c +755 -0
- data/ext/STEMMER_stem_UTF_8_portuguese.c +1023 -0
- data/ext/STEMMER_stem_UTF_8_romanian.c +1004 -0
- data/ext/STEMMER_stem_UTF_8_russian.c +694 -0
- data/ext/STEMMER_stem_UTF_8_spanish.c +1097 -0
- data/ext/STEMMER_stem_UTF_8_swedish.c +309 -0
- data/ext/STEMMER_stem_UTF_8_turkish.c +2205 -0
- data/ext/STEMMER_utilities.c +478 -0
- data/ext/analysis.c +1710 -0
- data/ext/analysis.h +266 -0
- data/ext/api.h +26 -0
- data/ext/array.c +125 -0
- data/ext/array.h +62 -0
- data/ext/bitvector.c +96 -0
- data/ext/bitvector.h +594 -0
- data/ext/bzlib.h +282 -0
- data/ext/bzlib_private.h +503 -0
- data/ext/compound_io.c +384 -0
- data/ext/config.h +52 -0
- data/ext/document.c +159 -0
- data/ext/document.h +63 -0
- data/ext/except.c +102 -0
- data/ext/except.h +176 -0
- data/ext/extconf.rb +15 -0
- data/ext/ferret.c +416 -0
- data/ext/ferret.h +94 -0
- data/ext/field_index.c +262 -0
- data/ext/field_index.h +52 -0
- data/ext/filter.c +157 -0
- data/ext/fs_store.c +493 -0
- data/ext/global.c +458 -0
- data/ext/global.h +302 -0
- data/ext/hash.c +524 -0
- data/ext/hash.h +515 -0
- data/ext/hashset.c +192 -0
- data/ext/hashset.h +215 -0
- data/ext/header.h +58 -0
- data/ext/helper.c +63 -0
- data/ext/helper.h +21 -0
- data/ext/index.c +6804 -0
- data/ext/index.h +935 -0
- data/ext/internal.h +1019 -0
- data/ext/lang.c +10 -0
- data/ext/lang.h +68 -0
- data/ext/libstemmer.h +79 -0
- data/ext/mempool.c +88 -0
- data/ext/mempool.h +43 -0
- data/ext/modules.h +190 -0
- data/ext/multimapper.c +351 -0
- data/ext/multimapper.h +60 -0
- data/ext/posh.c +1006 -0
- data/ext/posh.h +973 -0
- data/ext/priorityqueue.c +149 -0
- data/ext/priorityqueue.h +155 -0
- data/ext/q_boolean.c +1621 -0
- data/ext/q_const_score.c +162 -0
- data/ext/q_filtered_query.c +212 -0
- data/ext/q_fuzzy.c +280 -0
- data/ext/q_match_all.c +149 -0
- data/ext/q_multi_term.c +673 -0
- data/ext/q_parser.c +3103 -0
- data/ext/q_phrase.c +1206 -0
- data/ext/q_prefix.c +98 -0
- data/ext/q_range.c +682 -0
- data/ext/q_span.c +2390 -0
- data/ext/q_term.c +337 -0
- data/ext/q_wildcard.c +167 -0
- data/ext/r_analysis.c +2626 -0
- data/ext/r_index.c +3468 -0
- data/ext/r_qparser.c +635 -0
- data/ext/r_search.c +4490 -0
- data/ext/r_store.c +513 -0
- data/ext/r_utils.c +1131 -0
- data/ext/ram_store.c +476 -0
- data/ext/scanner.c +895 -0
- data/ext/scanner.h +36 -0
- data/ext/scanner_mb.c +6701 -0
- data/ext/scanner_utf8.c +4415 -0
- data/ext/search.c +1864 -0
- data/ext/search.h +953 -0
- data/ext/similarity.c +151 -0
- data/ext/similarity.h +89 -0
- data/ext/sort.c +786 -0
- data/ext/stem_ISO_8859_1_danish.h +16 -0
- data/ext/stem_ISO_8859_1_dutch.h +16 -0
- data/ext/stem_ISO_8859_1_english.h +16 -0
- data/ext/stem_ISO_8859_1_finnish.h +16 -0
- data/ext/stem_ISO_8859_1_french.h +16 -0
- data/ext/stem_ISO_8859_1_german.h +16 -0
- data/ext/stem_ISO_8859_1_hungarian.h +16 -0
- data/ext/stem_ISO_8859_1_italian.h +16 -0
- data/ext/stem_ISO_8859_1_norwegian.h +16 -0
- data/ext/stem_ISO_8859_1_porter.h +16 -0
- data/ext/stem_ISO_8859_1_portuguese.h +16 -0
- data/ext/stem_ISO_8859_1_spanish.h +16 -0
- data/ext/stem_ISO_8859_1_swedish.h +16 -0
- data/ext/stem_ISO_8859_2_romanian.h +16 -0
- data/ext/stem_KOI8_R_russian.h +16 -0
- data/ext/stem_UTF_8_danish.h +16 -0
- data/ext/stem_UTF_8_dutch.h +16 -0
- data/ext/stem_UTF_8_english.h +16 -0
- data/ext/stem_UTF_8_finnish.h +16 -0
- data/ext/stem_UTF_8_french.h +16 -0
- data/ext/stem_UTF_8_german.h +16 -0
- data/ext/stem_UTF_8_hungarian.h +16 -0
- data/ext/stem_UTF_8_italian.h +16 -0
- data/ext/stem_UTF_8_norwegian.h +16 -0
- data/ext/stem_UTF_8_porter.h +16 -0
- data/ext/stem_UTF_8_portuguese.h +16 -0
- data/ext/stem_UTF_8_romanian.h +16 -0
- data/ext/stem_UTF_8_russian.h +16 -0
- data/ext/stem_UTF_8_spanish.h +16 -0
- data/ext/stem_UTF_8_swedish.h +16 -0
- data/ext/stem_UTF_8_turkish.h +16 -0
- data/ext/stopwords.c +410 -0
- data/ext/store.c +698 -0
- data/ext/store.h +799 -0
- data/ext/symbol.c +10 -0
- data/ext/symbol.h +23 -0
- data/ext/term_vectors.c +73 -0
- data/ext/threading.h +31 -0
- data/ext/win32.h +62 -0
- data/lib/ferret.rb +30 -0
- data/lib/ferret/browser.rb +246 -0
- data/lib/ferret/browser/s/global.js +192 -0
- data/lib/ferret/browser/s/style.css +148 -0
- data/lib/ferret/browser/views/document/list.rhtml +49 -0
- data/lib/ferret/browser/views/document/show.rhtml +27 -0
- data/lib/ferret/browser/views/error/index.rhtml +7 -0
- data/lib/ferret/browser/views/help/index.rhtml +8 -0
- data/lib/ferret/browser/views/home/index.rhtml +29 -0
- data/lib/ferret/browser/views/layout.rhtml +22 -0
- data/lib/ferret/browser/views/term-vector/index.rhtml +4 -0
- data/lib/ferret/browser/views/term/index.rhtml +199 -0
- data/lib/ferret/browser/views/term/termdocs.rhtml +1 -0
- data/lib/ferret/browser/webrick.rb +14 -0
- data/lib/ferret/document.rb +130 -0
- data/lib/ferret/field_infos.rb +44 -0
- data/lib/ferret/field_symbol.rb +87 -0
- data/lib/ferret/index.rb +973 -0
- data/lib/ferret/number_tools.rb +157 -0
- data/lib/ferret/version.rb +3 -0
- data/setup.rb +1555 -0
- data/test/long_running/largefile/tc_largefile.rb +46 -0
- data/test/test_all.rb +5 -0
- data/test/test_helper.rb +29 -0
- data/test/test_installed.rb +1 -0
- data/test/threading/number_to_spoken.rb +132 -0
- data/test/threading/thread_safety_index_test.rb +88 -0
- data/test/threading/thread_safety_read_write_test.rb +73 -0
- data/test/threading/thread_safety_test.rb +133 -0
- data/test/unit/analysis/tc_analyzer.rb +550 -0
- data/test/unit/analysis/tc_token_stream.rb +653 -0
- data/test/unit/index/tc_index.rb +867 -0
- data/test/unit/index/tc_index_reader.rb +699 -0
- data/test/unit/index/tc_index_writer.rb +447 -0
- data/test/unit/index/th_doc.rb +332 -0
- data/test/unit/query_parser/tc_query_parser.rb +238 -0
- data/test/unit/search/tc_filter.rb +156 -0
- data/test/unit/search/tc_fuzzy_query.rb +147 -0
- data/test/unit/search/tc_index_searcher.rb +67 -0
- data/test/unit/search/tc_multi_searcher.rb +128 -0
- data/test/unit/search/tc_multiple_search_requests.rb +58 -0
- data/test/unit/search/tc_search_and_sort.rb +179 -0
- data/test/unit/search/tc_sort.rb +49 -0
- data/test/unit/search/tc_sort_field.rb +27 -0
- data/test/unit/search/tc_spans.rb +190 -0
- data/test/unit/search/tm_searcher.rb +436 -0
- data/test/unit/store/tc_fs_store.rb +115 -0
- data/test/unit/store/tc_ram_store.rb +35 -0
- data/test/unit/store/tm_store.rb +34 -0
- data/test/unit/store/tm_store_lock.rb +68 -0
- data/test/unit/tc_document.rb +81 -0
- data/test/unit/tc_field_symbol.rb +26 -0
- data/test/unit/ts_analysis.rb +2 -0
- data/test/unit/ts_index.rb +2 -0
- data/test/unit/ts_largefile.rb +4 -0
- data/test/unit/ts_query_parser.rb +2 -0
- data/test/unit/ts_search.rb +2 -0
- data/test/unit/ts_store.rb +2 -0
- data/test/unit/ts_utils.rb +2 -0
- data/test/unit/utils/tc_bit_vector.rb +295 -0
- data/test/unit/utils/tc_number_tools.rb +117 -0
- data/test/unit/utils/tc_priority_queue.rb +106 -0
- data/test/utils/content_generator.rb +226 -0
- metadata +319 -0
data/ext/r_utils.c
ADDED
@@ -0,0 +1,1131 @@
|
|
1
|
+
#include "ferret.h"
|
2
|
+
#include "bitvector.h"
|
3
|
+
#include "multimapper.h"
|
4
|
+
#ifdef FRT_RUBY_VERSION_1_9
|
5
|
+
# include <ruby/st.h>
|
6
|
+
#else
|
7
|
+
# include <st.h>
|
8
|
+
#endif
|
9
|
+
|
10
|
+
/*****************
|
11
|
+
*** BitVector ***
|
12
|
+
*****************/
|
13
|
+
static VALUE cBitVector;
|
14
|
+
|
15
|
+
static void
|
16
|
+
frb_bv_free(void *p)
|
17
|
+
{
|
18
|
+
object_del(p);
|
19
|
+
bv_destroy((BitVector *)p);
|
20
|
+
}
|
21
|
+
|
22
|
+
static VALUE
|
23
|
+
frb_bv_alloc(VALUE klass)
|
24
|
+
{
|
25
|
+
BitVector *bv = bv_new();
|
26
|
+
VALUE rbv = Data_Wrap_Struct(klass, NULL, &frb_bv_free, bv);
|
27
|
+
object_add(bv, rbv);
|
28
|
+
return rbv;
|
29
|
+
}
|
30
|
+
|
31
|
+
#define GET_BV(bv, self) Data_Get_Struct(self, BitVector, bv)
|
32
|
+
|
33
|
+
VALUE
|
34
|
+
frb_get_bv(BitVector *bv)
|
35
|
+
{
|
36
|
+
VALUE rbv;
|
37
|
+
if ((rbv = object_get(bv)) == Qnil) {
|
38
|
+
rbv = Data_Wrap_Struct(cBitVector, NULL, &frb_bv_free, bv);
|
39
|
+
REF(bv);
|
40
|
+
object_add(bv, rbv);
|
41
|
+
}
|
42
|
+
return rbv;
|
43
|
+
}
|
44
|
+
|
45
|
+
/*
|
46
|
+
* call-seq:
|
47
|
+
* BitVector.new() -> new_bit_vector
|
48
|
+
*
|
49
|
+
* Returns a new empty bit vector object
|
50
|
+
*/
|
51
|
+
static VALUE
|
52
|
+
frb_bv_init(VALUE self)
|
53
|
+
{
|
54
|
+
return self;
|
55
|
+
}
|
56
|
+
|
57
|
+
/*
|
58
|
+
* call-seq:
|
59
|
+
* bv[i] = bool -> bool
|
60
|
+
*
|
61
|
+
* Set the bit and _i_ to *val* (+true+ or
|
62
|
+
* +false+).
|
63
|
+
*/
|
64
|
+
VALUE
|
65
|
+
frb_bv_set(VALUE self, VALUE rindex, VALUE rstate)
|
66
|
+
{
|
67
|
+
BitVector *bv;
|
68
|
+
int index = FIX2INT(rindex);
|
69
|
+
GET_BV(bv, self);
|
70
|
+
if (index < 0) {
|
71
|
+
rb_raise(rb_eIndexError, "%d < 0", index);
|
72
|
+
}
|
73
|
+
if (RTEST(rstate)) {
|
74
|
+
bv_set(bv, index);
|
75
|
+
}
|
76
|
+
else {
|
77
|
+
bv_unset(bv, index);
|
78
|
+
}
|
79
|
+
|
80
|
+
return rstate;
|
81
|
+
}
|
82
|
+
|
83
|
+
/*
|
84
|
+
* call-seq:
|
85
|
+
* bv.set(i) -> self
|
86
|
+
*
|
87
|
+
* Set the bit at _i_ to *on* (+true+)
|
88
|
+
*/
|
89
|
+
VALUE
|
90
|
+
frb_bv_set_on(VALUE self, VALUE rindex)
|
91
|
+
{
|
92
|
+
frb_bv_set(self, rindex, Qtrue);
|
93
|
+
return self;
|
94
|
+
}
|
95
|
+
|
96
|
+
/*
|
97
|
+
* call-seq:
|
98
|
+
* bv.unset(i) -> self
|
99
|
+
*
|
100
|
+
* Set the bit at _i_ to *off* (+false+)
|
101
|
+
*/
|
102
|
+
VALUE
|
103
|
+
frb_bv_set_off(VALUE self, VALUE rindex)
|
104
|
+
{
|
105
|
+
frb_bv_set(self, rindex, Qfalse);
|
106
|
+
return self;
|
107
|
+
}
|
108
|
+
|
109
|
+
/*
|
110
|
+
* call-seq:
|
111
|
+
* bv.get(i) -> bool
|
112
|
+
* bv[i] -> bool
|
113
|
+
*
|
114
|
+
* Get the bit value at _i_
|
115
|
+
*/
|
116
|
+
VALUE
|
117
|
+
frb_bv_get(VALUE self, VALUE rindex)
|
118
|
+
{
|
119
|
+
BitVector *bv;
|
120
|
+
int index = FIX2INT(rindex);
|
121
|
+
GET_BV(bv, self);
|
122
|
+
if (index < 0) {
|
123
|
+
rb_raise(rb_eIndexError, "%d < 0", index);
|
124
|
+
}
|
125
|
+
|
126
|
+
return bv_get(bv, index) ? Qtrue : Qfalse;
|
127
|
+
}
|
128
|
+
|
129
|
+
/*
|
130
|
+
* call-seq:
|
131
|
+
* bv.count -> bit_count
|
132
|
+
*
|
133
|
+
* Count the number of bits set in the bit vector. If the bit vector has been
|
134
|
+
* negated using +#not+ then count the number of unset bits
|
135
|
+
* instead.
|
136
|
+
*/
|
137
|
+
VALUE
|
138
|
+
frb_bv_count(VALUE self)
|
139
|
+
{
|
140
|
+
BitVector *bv;
|
141
|
+
GET_BV(bv, self);
|
142
|
+
return INT2FIX(bv->count);
|
143
|
+
}
|
144
|
+
|
145
|
+
/*
|
146
|
+
* call-seq:
|
147
|
+
* bv.clear -> self
|
148
|
+
*
|
149
|
+
* Clears all set bits in the bit vector. Negated bit vectors will still have
|
150
|
+
* all bits set to *off*.
|
151
|
+
*/
|
152
|
+
VALUE
|
153
|
+
frb_bv_clear(VALUE self)
|
154
|
+
{
|
155
|
+
BitVector *bv;
|
156
|
+
GET_BV(bv, self);
|
157
|
+
bv_clear(bv);
|
158
|
+
bv_scan_reset(bv);
|
159
|
+
return self;
|
160
|
+
}
|
161
|
+
|
162
|
+
/*
|
163
|
+
* call-seq:
|
164
|
+
* bv1 == bv2 -> bool
|
165
|
+
* bv1 != bv2 -> bool
|
166
|
+
* bv1.eql(bv2) -> bool
|
167
|
+
*
|
168
|
+
* Compares two bit vectors and returns true if both bit vectors have the same
|
169
|
+
* bits set.
|
170
|
+
*/
|
171
|
+
VALUE
|
172
|
+
frb_bv_eql(VALUE self, VALUE other)
|
173
|
+
{
|
174
|
+
BitVector *bv1, *bv2;
|
175
|
+
GET_BV(bv1, self);
|
176
|
+
GET_BV(bv2, other);
|
177
|
+
return bv_eq(bv1, bv2) ? Qtrue : Qfalse;
|
178
|
+
}
|
179
|
+
|
180
|
+
/*
|
181
|
+
* call-seq:
|
182
|
+
* bv.hash -> int
|
183
|
+
*
|
184
|
+
* Used to store bit vectors in Hashes. Especially useful if you want to
|
185
|
+
* cache them.
|
186
|
+
*/
|
187
|
+
VALUE
|
188
|
+
frb_bv_hash(VALUE self)
|
189
|
+
{
|
190
|
+
BitVector *bv;
|
191
|
+
GET_BV(bv, self);
|
192
|
+
return LONG2NUM(bv_hash(bv));
|
193
|
+
}
|
194
|
+
|
195
|
+
/*
|
196
|
+
* call-seq:
|
197
|
+
* bv1 & bv2 -> anded_bv
|
198
|
+
* bv1.and(bv2) -> anded_bv
|
199
|
+
*
|
200
|
+
* Perform a boolean _and_ operation on +bv1+ and
|
201
|
+
* +bv2+
|
202
|
+
*/
|
203
|
+
VALUE
|
204
|
+
frb_bv_and(VALUE self, VALUE other)
|
205
|
+
{
|
206
|
+
BitVector *bv1, *bv2;
|
207
|
+
GET_BV(bv1, self);
|
208
|
+
GET_BV(bv2, other);
|
209
|
+
return Data_Wrap_Struct(cBitVector, NULL, &bv_destroy, bv_and(bv1, bv2));
|
210
|
+
}
|
211
|
+
|
212
|
+
/*
|
213
|
+
* call-seq:
|
214
|
+
* bv1.and!(bv2) -> self
|
215
|
+
*
|
216
|
+
* Perform a boolean _and_ operation on +bv1+ and
|
217
|
+
* +bv2+ in place on +bv1+
|
218
|
+
*/
|
219
|
+
VALUE
|
220
|
+
frb_bv_and_x(VALUE self, VALUE other)
|
221
|
+
{
|
222
|
+
BitVector *bv1, *bv2;
|
223
|
+
GET_BV(bv1, self);
|
224
|
+
GET_BV(bv2, other);
|
225
|
+
bv_and_x(bv1, bv2);
|
226
|
+
return self;
|
227
|
+
}
|
228
|
+
|
229
|
+
/*
|
230
|
+
* call-seq:
|
231
|
+
* bv1 | bv2 -> ored_bv
|
232
|
+
* bv1.or(bv2) -> ored_bv
|
233
|
+
*
|
234
|
+
* Perform a boolean _or_ operation on +bv1+ and
|
235
|
+
* +bv2+
|
236
|
+
*/
|
237
|
+
VALUE
|
238
|
+
frb_bv_or(VALUE self, VALUE other)
|
239
|
+
{
|
240
|
+
BitVector *bv1, *bv2;
|
241
|
+
GET_BV(bv1, self);
|
242
|
+
GET_BV(bv2, other);
|
243
|
+
return Data_Wrap_Struct(cBitVector, NULL, &bv_destroy, bv_or(bv1, bv2));
|
244
|
+
}
|
245
|
+
|
246
|
+
/*
|
247
|
+
* call-seq:
|
248
|
+
* bv1.or!(bv2) -> self
|
249
|
+
*
|
250
|
+
* Perform a boolean _or_ operation on +bv1+ and
|
251
|
+
* +bv2+ in place on +bv1+
|
252
|
+
*/
|
253
|
+
VALUE
|
254
|
+
frb_bv_or_x(VALUE self, VALUE other)
|
255
|
+
{
|
256
|
+
BitVector *bv1, *bv2;
|
257
|
+
GET_BV(bv1, self);
|
258
|
+
GET_BV(bv2, other);
|
259
|
+
bv_or_x(bv1, bv2);
|
260
|
+
return self;
|
261
|
+
}
|
262
|
+
|
263
|
+
/*
|
264
|
+
* call-seq:
|
265
|
+
* bv1 ^ bv2 -> xored_bv
|
266
|
+
* bv1.xor(bv2) -> xored_bv
|
267
|
+
*
|
268
|
+
* Perform a boolean _xor_ operation on +bv1+ and
|
269
|
+
* +bv2+
|
270
|
+
*/
|
271
|
+
VALUE
|
272
|
+
frb_bv_xor(VALUE self, VALUE other)
|
273
|
+
{
|
274
|
+
BitVector *bv1, *bv2;
|
275
|
+
GET_BV(bv1, self);
|
276
|
+
GET_BV(bv2, other);
|
277
|
+
return Data_Wrap_Struct(cBitVector, NULL, &bv_destroy, bv_xor(bv1, bv2));
|
278
|
+
}
|
279
|
+
|
280
|
+
/*
|
281
|
+
* call-seq:
|
282
|
+
* bv1.xor!(bv2) -> self
|
283
|
+
*
|
284
|
+
* Perform a boolean _xor_ operation on +bv1+ and
|
285
|
+
* +bv2+ in place on +bv1+
|
286
|
+
*/
|
287
|
+
VALUE
|
288
|
+
frb_bv_xor_x(VALUE self, VALUE other)
|
289
|
+
{
|
290
|
+
BitVector *bv1, *bv2;
|
291
|
+
GET_BV(bv1, self);
|
292
|
+
GET_BV(bv2, other);
|
293
|
+
bv_xor_x(bv1, bv2);
|
294
|
+
return self;
|
295
|
+
}
|
296
|
+
|
297
|
+
/*
|
298
|
+
* call-seq:
|
299
|
+
* ~bv -> bv
|
300
|
+
* bv.not -> bv
|
301
|
+
*
|
302
|
+
* Perform a boolean _not_ operation on +bv+
|
303
|
+
* */
|
304
|
+
VALUE
|
305
|
+
frb_bv_not(VALUE self)
|
306
|
+
{
|
307
|
+
BitVector *bv;
|
308
|
+
GET_BV(bv, self);
|
309
|
+
return Data_Wrap_Struct(cBitVector, NULL, &bv_destroy, bv_not(bv));
|
310
|
+
}
|
311
|
+
|
312
|
+
/*
|
313
|
+
* call-seq:
|
314
|
+
* bv.not! -> self
|
315
|
+
*
|
316
|
+
* Perform a boolean _not_ operation on +bv+ in-place
|
317
|
+
*/
|
318
|
+
VALUE
|
319
|
+
frb_bv_not_x(VALUE self)
|
320
|
+
{
|
321
|
+
BitVector *bv;
|
322
|
+
GET_BV(bv, self);
|
323
|
+
bv_not_x(bv);
|
324
|
+
return self;
|
325
|
+
}
|
326
|
+
|
327
|
+
/*
|
328
|
+
* call-seq:
|
329
|
+
* bv.reset_scan -> self
|
330
|
+
*
|
331
|
+
* Resets the BitVector ready for scanning. You should call this method
|
332
|
+
* before calling +#next+ or +#next_unset+. It isn't
|
333
|
+
* necessary for the other scan methods or for the +#each+ method.
|
334
|
+
*/
|
335
|
+
VALUE
|
336
|
+
frb_bv_reset_scan(VALUE self)
|
337
|
+
{
|
338
|
+
BitVector *bv;
|
339
|
+
GET_BV(bv, self);
|
340
|
+
bv_scan_reset(bv);
|
341
|
+
return self;
|
342
|
+
}
|
343
|
+
|
344
|
+
/*
|
345
|
+
* call-seq:
|
346
|
+
* bv.next -> bit_num
|
347
|
+
*
|
348
|
+
* Returns the next set bit in the bit vector scanning from low order to high
|
349
|
+
* order. You should call +#reset_scan+ before calling this method
|
350
|
+
* if you want to scan from the beginning. It is automatically reset when you
|
351
|
+
* first create the bit vector.
|
352
|
+
*/
|
353
|
+
VALUE
|
354
|
+
frb_bv_next(VALUE self)
|
355
|
+
{
|
356
|
+
BitVector *bv;
|
357
|
+
GET_BV(bv, self);
|
358
|
+
return INT2FIX(bv_scan_next(bv));
|
359
|
+
}
|
360
|
+
|
361
|
+
/*
|
362
|
+
* call-seq:
|
363
|
+
* bv.next_unset -> bit_num
|
364
|
+
*
|
365
|
+
* Returns the next unset bit in the bit vector scanning from low order to
|
366
|
+
* high order. This method should only be called on bit vectors which have
|
367
|
+
* been flipped (negated). You should call +#reset_scan+ before
|
368
|
+
* calling this method if you want to scan from the beginning. It is
|
369
|
+
* automatically reset when you first create the bit vector.
|
370
|
+
*/
|
371
|
+
VALUE
|
372
|
+
frb_bv_next_unset(VALUE self)
|
373
|
+
{
|
374
|
+
BitVector *bv;
|
375
|
+
GET_BV(bv, self);
|
376
|
+
return INT2FIX(bv_scan_next_unset(bv));
|
377
|
+
}
|
378
|
+
|
379
|
+
/*
|
380
|
+
* call-seq:
|
381
|
+
* bv.next_from(from) -> bit_num
|
382
|
+
*
|
383
|
+
* Returns the next set bit in the bit vector scanning from low order to
|
384
|
+
* high order and starting at +from+. The scan is inclusive so if
|
385
|
+
* +from+ is equal to 10 and +bv[10]+ is set it will
|
386
|
+
* return the number 10. If the bit vector has been negated than you should
|
387
|
+
* use the +#next_unset_from+ method.
|
388
|
+
*/
|
389
|
+
VALUE
|
390
|
+
frb_bv_next_from(VALUE self, VALUE rfrom)
|
391
|
+
{
|
392
|
+
BitVector *bv;
|
393
|
+
int from = FIX2INT(rfrom);
|
394
|
+
GET_BV(bv, self);
|
395
|
+
if (from < 0) {
|
396
|
+
from = 0;
|
397
|
+
}
|
398
|
+
return INT2FIX(bv_scan_next_from(bv, from));
|
399
|
+
}
|
400
|
+
|
401
|
+
/*
|
402
|
+
* call-seq:
|
403
|
+
* bv.next_unset_from(from) -> bit_num
|
404
|
+
*
|
405
|
+
* Returns the next unset bit in the bit vector scanning from low order to
|
406
|
+
* high order and starting at +from+. The scan is inclusive so if
|
407
|
+
* +from+ is equal to 10 and +bv[10]+ is unset it will
|
408
|
+
* return the number 10. If the bit vector has not been negated than you
|
409
|
+
* should use the +#next_from+ method.
|
410
|
+
*/
|
411
|
+
VALUE
|
412
|
+
frb_bv_next_unset_from(VALUE self, VALUE rfrom)
|
413
|
+
{
|
414
|
+
BitVector *bv;
|
415
|
+
int from = FIX2INT(rfrom);
|
416
|
+
GET_BV(bv, self);
|
417
|
+
if (from < 0) {
|
418
|
+
from = 0;
|
419
|
+
}
|
420
|
+
return INT2FIX(bv_scan_next_unset_from(bv, from));
|
421
|
+
}
|
422
|
+
|
423
|
+
/*
|
424
|
+
* call-seq:
|
425
|
+
* bv.each { |bit_num| }
|
426
|
+
*
|
427
|
+
* Iterate through all the set bits in the bit vector yielding each one in
|
428
|
+
* order
|
429
|
+
*/
|
430
|
+
VALUE
|
431
|
+
frb_bv_each(VALUE self)
|
432
|
+
{
|
433
|
+
BitVector *bv;
|
434
|
+
int bit;
|
435
|
+
GET_BV(bv, self);
|
436
|
+
bv_scan_reset(bv);
|
437
|
+
if (bv->extends_as_ones) {
|
438
|
+
while ((bit = bv_scan_next_unset(bv)) >= 0) {
|
439
|
+
rb_yield(INT2FIX(bit));
|
440
|
+
}
|
441
|
+
}
|
442
|
+
else {
|
443
|
+
while ((bit = bv_scan_next(bv)) >= 0) {
|
444
|
+
rb_yield(INT2FIX(bit));
|
445
|
+
}
|
446
|
+
}
|
447
|
+
return self;
|
448
|
+
}
|
449
|
+
|
450
|
+
/*
|
451
|
+
* call-seq:
|
452
|
+
* bv.to_a
|
453
|
+
*
|
454
|
+
* Iterate through all the set bits in the bit vector adding the index of
|
455
|
+
* each set bit to an array. This is useful if you want to perform array
|
456
|
+
* methods on the bit vector. If you want to convert an array to a bit_vector
|
457
|
+
* simply do this;
|
458
|
+
*
|
459
|
+
* bv = [1, 12, 45, 367, 455].inject(BitVector.new) {|bv, i| bv.set(i)}
|
460
|
+
*/
|
461
|
+
VALUE
|
462
|
+
frb_bv_to_a(VALUE self)
|
463
|
+
{
|
464
|
+
BitVector *bv;
|
465
|
+
int bit;
|
466
|
+
VALUE ary;
|
467
|
+
GET_BV(bv, self);
|
468
|
+
ary = rb_ary_new();
|
469
|
+
bv_scan_reset(bv);
|
470
|
+
if (bv->extends_as_ones) {
|
471
|
+
while ((bit = bv_scan_next_unset(bv)) >= 0) {
|
472
|
+
rb_ary_push(ary, INT2FIX(bit));
|
473
|
+
}
|
474
|
+
}
|
475
|
+
else {
|
476
|
+
while ((bit = bv_scan_next(bv)) >= 0) {
|
477
|
+
rb_ary_push(ary, INT2FIX(bit));
|
478
|
+
}
|
479
|
+
}
|
480
|
+
return ary;
|
481
|
+
}
|
482
|
+
|
483
|
+
static VALUE mUtils;
|
484
|
+
|
485
|
+
/*
|
486
|
+
* Document-class: Ferret::Utils::BitVector
|
487
|
+
*
|
488
|
+
* == Summary
|
489
|
+
*
|
490
|
+
* A BitVector is pretty easy to implement in Ruby using Ruby's BigNum class.
|
491
|
+
* This BitVector however allows you to count the set bits with the
|
492
|
+
* +#count+ method (or unset bits of flipped bit vectors) and also
|
493
|
+
* to quickly scan the set bits.
|
494
|
+
*
|
495
|
+
* == Boolean Operations
|
496
|
+
*
|
497
|
+
* BitVector handles four boolean operations;
|
498
|
+
*
|
499
|
+
* * +&+
|
500
|
+
* * +|+
|
501
|
+
* * +^+
|
502
|
+
* * +~+
|
503
|
+
*
|
504
|
+
* bv1 = BitVector.new
|
505
|
+
* bv2 = BitVector.new
|
506
|
+
* bv3 = BitVector.new
|
507
|
+
*
|
508
|
+
* bv4 = (bv1 & bv2) | ~bv3
|
509
|
+
*
|
510
|
+
* You can also do the operations in-place;
|
511
|
+
*
|
512
|
+
* * +and!+
|
513
|
+
* * +or!+
|
514
|
+
* * +xor!+
|
515
|
+
* * +not!+
|
516
|
+
*
|
517
|
+
* bv4.and!(bv5).not!
|
518
|
+
*
|
519
|
+
* == Set Bit Scanning
|
520
|
+
*
|
521
|
+
* Perhaps the most useful functionality in BitVector is the ability to
|
522
|
+
* quickly scan for set bits. To print all set bits;
|
523
|
+
*
|
524
|
+
* bv.each {|bit| puts bit }
|
525
|
+
*
|
526
|
+
* Alternatively you could use the lower level +next+ or
|
527
|
+
* +next_unset+ methods. Note that the +each+ method will
|
528
|
+
* automatically scan unset bits if the BitVector has been flipped (using
|
529
|
+
* +not+).
|
530
|
+
*/
|
531
|
+
static void
|
532
|
+
Init_BitVector(void)
|
533
|
+
{
|
534
|
+
/* BitVector */
|
535
|
+
cBitVector = rb_define_class_under(mUtils, "BitVector", rb_cObject);
|
536
|
+
rb_define_alloc_func(cBitVector, frb_bv_alloc);
|
537
|
+
|
538
|
+
rb_define_method(cBitVector, "initialize", frb_bv_init, 0);
|
539
|
+
rb_define_method(cBitVector, "set", frb_bv_set_on, 1);
|
540
|
+
rb_define_method(cBitVector, "unset", frb_bv_set_off, 1);
|
541
|
+
rb_define_method(cBitVector, "[]=", frb_bv_set, 2);
|
542
|
+
rb_define_method(cBitVector, "get", frb_bv_get, 1);
|
543
|
+
rb_define_method(cBitVector, "[]", frb_bv_get, 1);
|
544
|
+
rb_define_method(cBitVector, "count", frb_bv_count, 0);
|
545
|
+
rb_define_method(cBitVector, "clear", frb_bv_clear, 0);
|
546
|
+
rb_define_method(cBitVector, "eql?", frb_bv_eql, 1);
|
547
|
+
rb_define_method(cBitVector, "==", frb_bv_eql, 1);
|
548
|
+
rb_define_method(cBitVector, "hash", frb_bv_hash, 0);
|
549
|
+
rb_define_method(cBitVector, "and!", frb_bv_and_x, 1);
|
550
|
+
rb_define_method(cBitVector, "and", frb_bv_and, 1);
|
551
|
+
rb_define_method(cBitVector, "&", frb_bv_and, 1);
|
552
|
+
rb_define_method(cBitVector, "or!", frb_bv_or_x, 1);
|
553
|
+
rb_define_method(cBitVector, "or", frb_bv_or, 1);
|
554
|
+
rb_define_method(cBitVector, "|", frb_bv_or, 1);
|
555
|
+
rb_define_method(cBitVector, "xor!", frb_bv_xor_x, 1);
|
556
|
+
rb_define_method(cBitVector, "xor", frb_bv_xor, 1);
|
557
|
+
rb_define_method(cBitVector, "^", frb_bv_xor, 1);
|
558
|
+
rb_define_method(cBitVector, "not!", frb_bv_not_x, 0);
|
559
|
+
rb_define_method(cBitVector, "not", frb_bv_not, 0);
|
560
|
+
rb_define_method(cBitVector, "~", frb_bv_not, 0);
|
561
|
+
rb_define_method(cBitVector, "reset_scan", frb_bv_reset_scan, 0);
|
562
|
+
rb_define_method(cBitVector, "next", frb_bv_next, 0);
|
563
|
+
rb_define_method(cBitVector, "next_unset", frb_bv_next_unset, 0);
|
564
|
+
rb_define_method(cBitVector, "next_from", frb_bv_next_from, 1);
|
565
|
+
rb_define_method(cBitVector, "next_unset_from", frb_bv_next_unset_from, 1);
|
566
|
+
rb_define_method(cBitVector, "each", frb_bv_each, 0);
|
567
|
+
rb_define_method(cBitVector, "to_a", frb_bv_to_a, 0);
|
568
|
+
}
|
569
|
+
|
570
|
+
/*******************
|
571
|
+
*** MultiMapper ***
|
572
|
+
*******************/
|
573
|
+
static VALUE cMultiMapper;
|
574
|
+
|
575
|
+
static void
|
576
|
+
frb_mulmap_free(void *p)
|
577
|
+
{
|
578
|
+
object_del(p);
|
579
|
+
mulmap_destroy((MultiMapper *)p);
|
580
|
+
}
|
581
|
+
|
582
|
+
static VALUE
|
583
|
+
frb_mulmap_alloc(VALUE klass)
|
584
|
+
{
|
585
|
+
MultiMapper *mulmap = mulmap_new();
|
586
|
+
VALUE rmulmap = Data_Wrap_Struct(klass, NULL, &frb_mulmap_free, mulmap);
|
587
|
+
object_add(mulmap, rmulmap);
|
588
|
+
return rmulmap;
|
589
|
+
}
|
590
|
+
|
591
|
+
/* XXX: Duplication from frb_add_mapping_i in r_analysis.c */
|
592
|
+
static INLINE void frb_mulmap_add_mapping_i(MultiMapper *mulmap, VALUE from,
|
593
|
+
const char *to)
|
594
|
+
{
|
595
|
+
switch (TYPE(from)) {
|
596
|
+
case T_STRING:
|
597
|
+
mulmap_add_mapping(mulmap, rs2s(from), to);
|
598
|
+
break;
|
599
|
+
case T_SYMBOL:
|
600
|
+
mulmap_add_mapping(mulmap, rb_id2name(SYM2ID(from)), to);
|
601
|
+
break;
|
602
|
+
default:
|
603
|
+
rb_raise(rb_eArgError,
|
604
|
+
"cannot map from %s with MappingFilter",
|
605
|
+
rs2s(rb_obj_as_string(from)));
|
606
|
+
break;
|
607
|
+
}
|
608
|
+
}
|
609
|
+
|
610
|
+
/* XXX: Duplication from frb_add_mappings_i in r_analysis.c */
|
611
|
+
static int frb_mulmap_add_mappings_i(VALUE key, VALUE value, VALUE arg)
|
612
|
+
{
|
613
|
+
if (key == Qundef) {
|
614
|
+
return ST_CONTINUE;
|
615
|
+
} else {
|
616
|
+
MultiMapper *mulmap = (MultiMapper *)arg;
|
617
|
+
const char *to;
|
618
|
+
switch (TYPE(value)) {
|
619
|
+
case T_STRING:
|
620
|
+
to = rs2s(value);
|
621
|
+
break;
|
622
|
+
case T_SYMBOL:
|
623
|
+
to = rb_id2name(SYM2ID(value));
|
624
|
+
break;
|
625
|
+
default:
|
626
|
+
rb_raise(rb_eArgError,
|
627
|
+
"cannot map to %s with MultiMapper",
|
628
|
+
rs2s(rb_obj_as_string(key)));
|
629
|
+
break;
|
630
|
+
}
|
631
|
+
if (TYPE(key) == T_ARRAY) {
|
632
|
+
int i;
|
633
|
+
for (i = RARRAY_LEN(key) - 1; i >= 0; i--) {
|
634
|
+
frb_mulmap_add_mapping_i(mulmap, RARRAY_PTR(key)[i], to);
|
635
|
+
}
|
636
|
+
}
|
637
|
+
else {
|
638
|
+
frb_mulmap_add_mapping_i(mulmap, key, to);
|
639
|
+
}
|
640
|
+
}
|
641
|
+
return ST_CONTINUE;
|
642
|
+
}
|
643
|
+
|
644
|
+
/*
|
645
|
+
* call-seq:
|
646
|
+
* MultiMapper.new() -> new_multi_mapper
|
647
|
+
*
|
648
|
+
* Returns a new multi-mapper object and compiles it for optimization.
|
649
|
+
*
|
650
|
+
* Note that MultiMapper is immutable.
|
651
|
+
*/
|
652
|
+
static VALUE
|
653
|
+
frb_mulmap_init(VALUE self, VALUE rmappings)
|
654
|
+
{
|
655
|
+
MultiMapper *mulmap = DATA_PTR(self);
|
656
|
+
rb_hash_foreach(rmappings, frb_mulmap_add_mappings_i, (VALUE)mulmap);
|
657
|
+
mulmap_compile(mulmap);
|
658
|
+
|
659
|
+
return self;
|
660
|
+
}
|
661
|
+
|
662
|
+
/*
|
663
|
+
* call-seq:
|
664
|
+
* multi_mapper.map(string) -> mapped_string
|
665
|
+
*
|
666
|
+
* Performs all the mappings on the string.
|
667
|
+
*/
|
668
|
+
VALUE
|
669
|
+
frb_mulmap_map(VALUE self, VALUE rstring)
|
670
|
+
{
|
671
|
+
MultiMapper *mulmap = DATA_PTR(self);
|
672
|
+
char *string = rs2s(rb_obj_as_string(rstring));
|
673
|
+
char *mapped_string = mulmap_dynamic_map(mulmap, string);
|
674
|
+
VALUE rmapped_string = rb_str_new2(mapped_string);
|
675
|
+
free(mapped_string);
|
676
|
+
return rmapped_string;
|
677
|
+
}
|
678
|
+
|
679
|
+
/*
|
680
|
+
* Document-class: Ferret::Utils::MultiMapper
|
681
|
+
*
|
682
|
+
* == Summary
|
683
|
+
*
|
684
|
+
* A MultiMapper performs a list of mappings from one string to another. You
|
685
|
+
* could of course just use gsub to do this but when you are just mapping
|
686
|
+
* strings, this is much faster.
|
687
|
+
*
|
688
|
+
* Note that MultiMapper is immutable.
|
689
|
+
*
|
690
|
+
* == Example
|
691
|
+
*
|
692
|
+
* mapping = {
|
693
|
+
* ['à','á','â','ã','ä','å','ā','ă'] => 'a',
|
694
|
+
* 'æ' => 'ae',
|
695
|
+
* ['ď','đ'] => 'd',
|
696
|
+
* ['ç','ć','č','ĉ','ċ'] => 'c',
|
697
|
+
* ['è','é','ê','ë','ē','ę','ě','ĕ','ė',] => 'e',
|
698
|
+
* ['ƒ'] => 'f',
|
699
|
+
* ['ĝ','ğ','ġ','ģ'] => 'g',
|
700
|
+
* ['ĥ','ħ'] => 'h',
|
701
|
+
* ['ì','ì','í','î','ï','ī','ĩ','ĭ'] => 'i',
|
702
|
+
* ['į','ı','ij','ĵ'] => 'j',
|
703
|
+
* ['ķ','ĸ'] => 'k',
|
704
|
+
* ['ł','ľ','ĺ','ļ','ŀ'] => 'l',
|
705
|
+
* ['ñ','ń','ň','ņ','ʼn','ŋ'] => 'n',
|
706
|
+
* ['ò','ó','ô','õ','ö','ø','ō','ő','ŏ','ŏ'] => 'o',
|
707
|
+
* ['œ'] => 'oek',
|
708
|
+
* ['ą'] => 'q',
|
709
|
+
* ['ŕ','ř','ŗ'] => 'r',
|
710
|
+
* ['ś','š','ş','ŝ','ș'] => 's',
|
711
|
+
* ['ť','ţ','ŧ','ț'] => 't',
|
712
|
+
* ['ù','ú','û','ü','ū','ů','ű','ŭ','ũ','ų'] => 'u',
|
713
|
+
* ['ŵ'] => 'w',
|
714
|
+
* ['ý','ÿ','ŷ'] => 'y',
|
715
|
+
* ['ž','ż','ź'] => 'z'
|
716
|
+
* mapper = MultiMapper.new(mapping)
|
717
|
+
* mapped_string = mapper.map(string)
|
718
|
+
*/
|
719
|
+
static void
|
720
|
+
Init_MultiMapper(void)
|
721
|
+
{
|
722
|
+
/* MultiMapper */
|
723
|
+
cMultiMapper = rb_define_class_under(mUtils, "MultiMapper", rb_cObject);
|
724
|
+
rb_define_alloc_func(cMultiMapper, frb_mulmap_alloc);
|
725
|
+
|
726
|
+
rb_define_method(cMultiMapper, "initialize", frb_mulmap_init, 1);
|
727
|
+
rb_define_method(cMultiMapper, "map", frb_mulmap_map, 1);
|
728
|
+
}
|
729
|
+
|
730
|
+
/*********************
|
731
|
+
*** PriorityQueue ***
|
732
|
+
*********************/
|
733
|
+
typedef struct PriQ
|
734
|
+
{
|
735
|
+
int size;
|
736
|
+
int capa;
|
737
|
+
int mem_capa;
|
738
|
+
VALUE *heap;
|
739
|
+
VALUE proc;
|
740
|
+
} PriQ;
|
741
|
+
|
742
|
+
#define PQ_START_CAPA 32
|
743
|
+
|
744
|
+
static bool frb_pq_lt(VALUE proc, VALUE v1, VALUE v2)
|
745
|
+
{
|
746
|
+
if (proc == Qnil) {
|
747
|
+
return RTEST(rb_funcall(v1, id_lt, 1, v2));
|
748
|
+
}
|
749
|
+
else {
|
750
|
+
return RTEST(rb_funcall(proc, id_call, 2, v1, v2));
|
751
|
+
}
|
752
|
+
}
|
753
|
+
|
754
|
+
static void pq_up(PriQ *pq)
|
755
|
+
{
|
756
|
+
VALUE *heap = pq->heap;
|
757
|
+
VALUE node;
|
758
|
+
int i = pq->size;
|
759
|
+
int j = i >> 1;
|
760
|
+
|
761
|
+
node = heap[i];
|
762
|
+
|
763
|
+
while ((j > 0) && frb_pq_lt(pq->proc, node, heap[j])) {
|
764
|
+
heap[i] = heap[j];
|
765
|
+
i = j;
|
766
|
+
j = j >> 1;
|
767
|
+
}
|
768
|
+
heap[i] = node;
|
769
|
+
}
|
770
|
+
|
771
|
+
static void pq_down(PriQ *pq)
|
772
|
+
{
|
773
|
+
register int i = 1;
|
774
|
+
register int j = 2; /* i << 1; */
|
775
|
+
register int k = 3; /* j + 1; */
|
776
|
+
register int size = pq->size;
|
777
|
+
VALUE *heap = pq->heap;
|
778
|
+
VALUE node = heap[i]; /* save top node */
|
779
|
+
|
780
|
+
if ((k <= size) && (frb_pq_lt(pq->proc, heap[k], heap[j]))) {
|
781
|
+
j = k;
|
782
|
+
}
|
783
|
+
|
784
|
+
while ((j <= size) && frb_pq_lt(pq->proc, heap[j], node)) {
|
785
|
+
heap[i] = heap[j]; /* shift up child */
|
786
|
+
i = j;
|
787
|
+
j = i << 1;
|
788
|
+
k = j + 1;
|
789
|
+
if ((k <= size) && frb_pq_lt(pq->proc, heap[k], heap[j])) {
|
790
|
+
j = k;
|
791
|
+
}
|
792
|
+
}
|
793
|
+
heap[i] = node;
|
794
|
+
}
|
795
|
+
|
796
|
+
static void pq_push(PriQ *pq, VALUE elem)
|
797
|
+
{
|
798
|
+
pq->size++;
|
799
|
+
if (pq->size >= pq->mem_capa) {
|
800
|
+
pq->mem_capa <<= 1;
|
801
|
+
REALLOC_N(pq->heap, VALUE, pq->mem_capa);
|
802
|
+
}
|
803
|
+
pq->heap[pq->size] = elem;
|
804
|
+
pq_up(pq);
|
805
|
+
}
|
806
|
+
|
807
|
+
static VALUE cPriorityQueue;
|
808
|
+
|
809
|
+
static void
|
810
|
+
frb_pq_mark(void *p)
|
811
|
+
{
|
812
|
+
PriQ *pq = (PriQ *)p;
|
813
|
+
int i;
|
814
|
+
for (i = pq->size; i > 0; i--) {
|
815
|
+
rb_gc_mark_maybe(pq->heap[i]);
|
816
|
+
}
|
817
|
+
}
|
818
|
+
|
819
|
+
static void frb_pq_free(PriQ *pq)
|
820
|
+
{
|
821
|
+
free(pq->heap);
|
822
|
+
free(pq);
|
823
|
+
}
|
824
|
+
|
825
|
+
static VALUE
|
826
|
+
frb_pq_alloc(VALUE klass)
|
827
|
+
{
|
828
|
+
PriQ *pq = ALLOC_AND_ZERO(PriQ);
|
829
|
+
pq->capa = PQ_START_CAPA;
|
830
|
+
pq->mem_capa = PQ_START_CAPA;
|
831
|
+
pq->heap = ALLOC_N(VALUE, PQ_START_CAPA);
|
832
|
+
pq->proc = Qnil;
|
833
|
+
return Data_Wrap_Struct(klass, &frb_pq_mark, &frb_pq_free, pq);
|
834
|
+
}
|
835
|
+
|
836
|
+
#define GET_PQ(pq, self) Data_Get_Struct(self, PriQ, pq)
|
837
|
+
/*
|
838
|
+
* call-seq:
|
839
|
+
* PriorityQueue.new(capacity = 32) -> new_pq
|
840
|
+
* PriorityQueue.new({:capacity => 32,
|
841
|
+
* :less_than_proc => lambda{|a, b| a < b}) -> new_pq
|
842
|
+
* PriorityQueue.new({:capacity => 32}) {|a, b| a < b} -> new_pq
|
843
|
+
*
|
844
|
+
* Returns a new empty priority queue object with an optional capacity.
|
845
|
+
* Once the capacity is filled, the lowest valued elements will be
|
846
|
+
* automatically popped off the top of the queue as more elements are
|
847
|
+
* inserted into the queue.
|
848
|
+
*/
|
849
|
+
static VALUE
|
850
|
+
frb_pq_init(int argc, VALUE *argv, VALUE self)
|
851
|
+
{
|
852
|
+
if (argc >= 1) {
|
853
|
+
PriQ *pq;
|
854
|
+
VALUE options = argv[0];
|
855
|
+
VALUE param;
|
856
|
+
int capa = PQ_START_CAPA;
|
857
|
+
GET_PQ(pq, self);
|
858
|
+
switch (TYPE(options)) {
|
859
|
+
case T_FIXNUM:
|
860
|
+
capa = FIX2INT(options);
|
861
|
+
break;
|
862
|
+
case T_HASH:
|
863
|
+
if (!NIL_P(param = rb_hash_aref(options,
|
864
|
+
ID2SYM(id_capacity)))) {
|
865
|
+
capa = FIX2INT(param);
|
866
|
+
}
|
867
|
+
if (!NIL_P(param = rb_hash_aref(options,
|
868
|
+
ID2SYM(id_less_than)))) {
|
869
|
+
pq->proc = param;
|
870
|
+
}
|
871
|
+
break;
|
872
|
+
default:
|
873
|
+
rb_raise(rb_eArgError,
|
874
|
+
"PriorityQueue#initialize only takes a Hash or "
|
875
|
+
"an integer");
|
876
|
+
|
877
|
+
break;
|
878
|
+
}
|
879
|
+
if (capa < 0) {
|
880
|
+
rb_raise(rb_eIndexError,
|
881
|
+
"PriorityQueue must have a capacity > 0. %d < 0",
|
882
|
+
capa);
|
883
|
+
}
|
884
|
+
pq->capa = capa;
|
885
|
+
if (rb_block_given_p()) {
|
886
|
+
pq->proc = rb_block_proc();
|
887
|
+
}
|
888
|
+
if (argc > 1) {
|
889
|
+
rb_raise(rb_eArgError,
|
890
|
+
"PriorityQueue#initialize only takes one parameter");
|
891
|
+
}
|
892
|
+
}
|
893
|
+
|
894
|
+
return self;
|
895
|
+
}
|
896
|
+
|
897
|
+
/*
|
898
|
+
* call-seq:
|
899
|
+
* pq.clone -> pq_clone
|
900
|
+
*
|
901
|
+
* Returns a shallow clone of the priority queue. That is only the priority
|
902
|
+
* queue is cloned, its contents are not cloned.
|
903
|
+
*/
|
904
|
+
static VALUE
|
905
|
+
frb_pq_clone(VALUE self)
|
906
|
+
{
|
907
|
+
PriQ *pq, *new_pq = ALLOC(PriQ);
|
908
|
+
GET_PQ(pq, self);
|
909
|
+
memcpy(new_pq, pq, sizeof(PriQ));
|
910
|
+
new_pq->heap = ALLOC_N(VALUE, new_pq->mem_capa);
|
911
|
+
memcpy(new_pq->heap, pq->heap, sizeof(VALUE) * (new_pq->size + 1));
|
912
|
+
|
913
|
+
return Data_Wrap_Struct(cPriorityQueue, &frb_pq_mark, &frb_pq_free, new_pq);
|
914
|
+
}
|
915
|
+
|
916
|
+
/*
|
917
|
+
* call-seq:
|
918
|
+
* pq.clear -> self
|
919
|
+
*
|
920
|
+
* Clears all elements from the priority queue. The size will be reset to 0.
|
921
|
+
*/
|
922
|
+
static VALUE
|
923
|
+
frb_pq_clear(VALUE self)
|
924
|
+
{
|
925
|
+
PriQ *pq;
|
926
|
+
GET_PQ(pq, self);
|
927
|
+
pq->size = 0;
|
928
|
+
return self;
|
929
|
+
}
|
930
|
+
|
931
|
+
/*
|
932
|
+
* call-seq:
|
933
|
+
* pq.insert(elem) -> self
|
934
|
+
* pq << elem -> self
|
935
|
+
*
|
936
|
+
* Insert an element into a queue. It will be inserted into the correct
|
937
|
+
* position in the queue according to its priority.
|
938
|
+
*/
|
939
|
+
static VALUE
|
940
|
+
frb_pq_insert(VALUE self, VALUE elem)
|
941
|
+
{
|
942
|
+
PriQ *pq;
|
943
|
+
GET_PQ(pq, self);
|
944
|
+
if (pq->size < pq->capa) {
|
945
|
+
pq_push(pq, elem);
|
946
|
+
}
|
947
|
+
else if (pq->size > 0 && frb_pq_lt(pq->proc, pq->heap[1], elem)) {
|
948
|
+
pq->heap[1] = elem;
|
949
|
+
pq_down(pq);
|
950
|
+
}
|
951
|
+
/* else ignore the element */
|
952
|
+
return self;
|
953
|
+
}
|
954
|
+
|
955
|
+
/*
|
956
|
+
* call-seq:
|
957
|
+
* pq.adjust -> self
|
958
|
+
*
|
959
|
+
* Sometimes you modify the top element in the priority queue so that its
|
960
|
+
* priority changes. When you do this you need to reorder the queue and you
|
961
|
+
* do this by calling the adjust method.
|
962
|
+
*/
|
963
|
+
static VALUE
|
964
|
+
frb_pq_adjust(VALUE self)
|
965
|
+
{
|
966
|
+
PriQ *pq;
|
967
|
+
GET_PQ(pq, self);
|
968
|
+
pq_down(pq);
|
969
|
+
return self;
|
970
|
+
}
|
971
|
+
|
972
|
+
/*
|
973
|
+
* call-seq:
|
974
|
+
* pq.top -> elem
|
975
|
+
*
|
976
|
+
* Returns the top element in the queue but does not remove it from the
|
977
|
+
* queue.
|
978
|
+
*/
|
979
|
+
static VALUE
|
980
|
+
frb_pq_top(VALUE self)
|
981
|
+
{
|
982
|
+
PriQ *pq;
|
983
|
+
GET_PQ(pq, self);
|
984
|
+
return (pq->size > 0) ? pq->heap[1] : Qnil;
|
985
|
+
}
|
986
|
+
|
987
|
+
/*
|
988
|
+
* call-seq:
|
989
|
+
* pq.pop -> elem
|
990
|
+
*
|
991
|
+
* Returns the top element in the queue removing it from the queue.
|
992
|
+
*/
|
993
|
+
static VALUE
|
994
|
+
frb_pq_pop(VALUE self)
|
995
|
+
{
|
996
|
+
PriQ *pq;
|
997
|
+
GET_PQ(pq, self);
|
998
|
+
if (pq->size > 0) {
|
999
|
+
VALUE result = pq->heap[1]; /* save first value */
|
1000
|
+
pq->heap[1] = pq->heap[pq->size]; /* move last to first */
|
1001
|
+
pq->heap[pq->size] = Qnil;
|
1002
|
+
pq->size--;
|
1003
|
+
pq_down(pq); /* adjust heap */
|
1004
|
+
return result;
|
1005
|
+
}
|
1006
|
+
else {
|
1007
|
+
return Qnil;
|
1008
|
+
}
|
1009
|
+
}
|
1010
|
+
|
1011
|
+
/*
|
1012
|
+
* call-seq:
|
1013
|
+
* pq.size -> integer
|
1014
|
+
*
|
1015
|
+
* Returns the size of the queue, ie. the number of elements currently stored
|
1016
|
+
* in the queue. The _size_ of a PriorityQueue can never be greater than
|
1017
|
+
* its _capacity_
|
1018
|
+
*/
|
1019
|
+
static VALUE
|
1020
|
+
frb_pq_size(VALUE self)
|
1021
|
+
{
|
1022
|
+
PriQ *pq;
|
1023
|
+
GET_PQ(pq, self);
|
1024
|
+
return INT2FIX(pq->size);
|
1025
|
+
}
|
1026
|
+
|
1027
|
+
/*
|
1028
|
+
* call-seq:
|
1029
|
+
* pq.capacity -> integer
|
1030
|
+
*
|
1031
|
+
* Returns the capacity of the queue, ie. the number of elements that can be
|
1032
|
+
* stored in a Priority queue before they start to drop off the end. The
|
1033
|
+
* _size_ of a PriorityQueue can never be greater than its
|
1034
|
+
* _capacity_
|
1035
|
+
*/
|
1036
|
+
static VALUE
|
1037
|
+
frb_pq_capa(VALUE self)
|
1038
|
+
{
|
1039
|
+
PriQ *pq;
|
1040
|
+
GET_PQ(pq, self);
|
1041
|
+
return INT2FIX(pq->capa);
|
1042
|
+
}
|
1043
|
+
|
1044
|
+
/*
|
1045
|
+
* Document-class: Ferret::Utils::PriorityQueue
|
1046
|
+
*
|
1047
|
+
* == Summary
|
1048
|
+
*
|
1049
|
+
* A PriorityQueue is a very useful data structure and one that needs a fast
|
1050
|
+
* implementation. Hence this priority queue is implemented in C. It is
|
1051
|
+
* pretty easy to use; basically you just insert elements into the queue and
|
1052
|
+
* pop them off.
|
1053
|
+
*
|
1054
|
+
* The elements are sorted with the lowest valued elements on the top of
|
1055
|
+
* the heap, ie the first to be popped off. Elements are ordered using the
|
1056
|
+
* less_than '<' method. To change the order of the queue you can either
|
1057
|
+
* reimplement the '<' method pass a block when you initialize the queue.
|
1058
|
+
*
|
1059
|
+
* You can also set the capacity of the PriorityQueue. Once you hit the
|
1060
|
+
* capacity, the lowest values elements are automatically popped of the top
|
1061
|
+
* of the queue as more elements are added.
|
1062
|
+
*
|
1063
|
+
* == Example
|
1064
|
+
*
|
1065
|
+
* Here is a toy example that sorts strings by their length and has a capacity
|
1066
|
+
* of 5;
|
1067
|
+
*
|
1068
|
+
* q = PriorityQueue.new(5) {|a, b| a.size < b.size}
|
1069
|
+
* q << "x"
|
1070
|
+
* q << "xxxxx"
|
1071
|
+
* q << "xxx"
|
1072
|
+
* q << "xxxx"
|
1073
|
+
* q << "xxxxxx"
|
1074
|
+
* q << "xx" # hit capacity so "x" will be popped off the top
|
1075
|
+
*
|
1076
|
+
* puts q.size #=> 5
|
1077
|
+
* word = q.pop #=> "xx"
|
1078
|
+
* q.top << "yyyy" # "xxxyyyy" will still be at the top of the queue
|
1079
|
+
* q.adjust # move "xxxyyyy" to its correct location in queue
|
1080
|
+
* word = q.pop #=> "xxxx"
|
1081
|
+
* word = q.pop #=> "xxxxx"
|
1082
|
+
* word = q.pop #=> "xxxxxx"
|
1083
|
+
* word = q.pop #=> "xxxyyyy"
|
1084
|
+
* word = q.pop #=> nil
|
1085
|
+
*/
|
1086
|
+
static void
|
1087
|
+
Init_PriorityQueue(void)
|
1088
|
+
{
|
1089
|
+
/* PriorityQueue */
|
1090
|
+
cPriorityQueue = rb_define_class_under(mUtils, "PriorityQueue", rb_cObject);
|
1091
|
+
rb_define_alloc_func(cPriorityQueue, frb_pq_alloc);
|
1092
|
+
|
1093
|
+
rb_define_method(cPriorityQueue, "initialize", frb_pq_init, -1);
|
1094
|
+
rb_define_method(cPriorityQueue, "clone", frb_pq_clone, 0);
|
1095
|
+
rb_define_method(cPriorityQueue, "clear", frb_pq_clear, 0);
|
1096
|
+
rb_define_method(cPriorityQueue, "insert", frb_pq_insert, 1);
|
1097
|
+
rb_define_method(cPriorityQueue, "<<", frb_pq_insert, 1);
|
1098
|
+
rb_define_method(cPriorityQueue, "top", frb_pq_top, 0);
|
1099
|
+
rb_define_method(cPriorityQueue, "pop", frb_pq_pop, 0);
|
1100
|
+
rb_define_method(cPriorityQueue, "size", frb_pq_size, 0);
|
1101
|
+
rb_define_method(cPriorityQueue, "capacity", frb_pq_capa, 0);
|
1102
|
+
rb_define_method(cPriorityQueue, "adjust", frb_pq_adjust, 0);
|
1103
|
+
}
|
1104
|
+
|
1105
|
+
/* rdoc hack
|
1106
|
+
extern VALUE mFerret = rb_define_module("Ferret");
|
1107
|
+
*/
|
1108
|
+
|
1109
|
+
/*
|
1110
|
+
* Document-module: Ferret::Utils
|
1111
|
+
*
|
1112
|
+
* The Utils module contains a number of helper classes and modules that are
|
1113
|
+
* useful when indexing with Ferret. They are;
|
1114
|
+
*
|
1115
|
+
* * BitVector
|
1116
|
+
* * MultiMapper
|
1117
|
+
* * PriorityQueue
|
1118
|
+
* * => more to come
|
1119
|
+
*
|
1120
|
+
* These helper classes could also be quite useful outside of Ferret and may
|
1121
|
+
* one day find themselves in their own separate library.
|
1122
|
+
*/
|
1123
|
+
void
|
1124
|
+
Init_Utils(void)
|
1125
|
+
{
|
1126
|
+
mUtils = rb_define_module_under(mFerret, "Utils");
|
1127
|
+
|
1128
|
+
Init_BitVector();
|
1129
|
+
Init_MultiMapper();
|
1130
|
+
Init_PriorityQueue();
|
1131
|
+
}
|