1 /* The MIT License
2
3 Copyright (c) 2008, by Attractive Chaos <attractivechaos@aol.co.uk>
4
5 Permission is hereby granted, free of charge, to any person obtaining
6 a copy of this software and associated documentation files (the
7 "Software"), to deal in the Software without restriction, including
8 without limitation the rights to use, copy, modify, merge, publish,
9 distribute, sublicense, and/or sell copies of the Software, and to
10 permit persons to whom the Software is furnished to do so, subject to
11 the following conditions:
12
13 The above copyright notice and this permission notice shall be
14 included in all copies or substantial portions of the Software.
15
16 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
17 EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
18 MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
19 NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
20 BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
21 ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
22 CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
23 SOFTWARE.
24 */
25
26 /*
27 An example:
28
29 #include "khash.h"
30 KHASH_MAP_INIT_INT(32, char)
31 int main() {
32 int ret, is_missing;
33 khiter_t k;
34 khash_t(32) *h = kh_init(32);
35 k = kh_put(32, h, 5, &ret);
36 if (!ret) kh_del(32, h, k);
37 kh_value(h, k) = 10;
38 k = kh_get(32, h, 10);
39 is_missing = (k == kh_end(h));
40 k = kh_get(32, h, 5);
41 kh_del(32, h, k);
42 for (k = kh_begin(h); k != kh_end(h); ++k)
43 if (kh_exist(h, k)) kh_value(h, k) = 1;
44 kh_destroy(32, h);
45 return 0;
46 }
47 */
48
49 /*
50 2008-09-19 (0.2.3):
51
52 * Corrected the example
53 * Improved interfaces
54
55 2008-09-11 (0.2.2):
56
57 * Improved speed a little in kh_put()
58
59 2008-09-10 (0.2.1):
60
61 * Added kh_clear()
62 * Fixed a compiling error
63
64 2008-09-02 (0.2.0):
65
66 * Changed to token concatenation which increases flexibility.
67
68 2008-08-31 (0.1.2):
69
70 * Fixed a bug in kh_get(), which has not been tested previously.
71
72 2008-08-31 (0.1.1):
73
74 * Added destructor
75 */
76
77
78 #ifndef __AC_KHASH_H
79 #define __AC_KHASH_H
80
81 #define AC_VERSION_KHASH_H "0.2.2"
82
83 #include <stdint.h>
84 #include <stdlib.h>
85 #include <string.h>
86
87 typedef uint32_t khint_t;
88 typedef khint_t khiter_t;
89
90 #define __ac_HASH_PRIME_SIZE 32
91 static const uint32_t __ac_prime_list[__ac_HASH_PRIME_SIZE] =
92 {
93 0ul, 3ul, 11ul, 23ul, 53ul,
94 97ul, 193ul, 389ul, 769ul, 1543ul,
95 3079ul, 6151ul, 12289ul, 24593ul, 49157ul,
96 98317ul, 196613ul, 393241ul, 786433ul, 1572869ul,
97 3145739ul, 6291469ul, 12582917ul, 25165843ul, 50331653ul,
98 100663319ul, 201326611ul, 402653189ul, 805306457ul, 1610612741ul,
99 3221225473ul, 4294967291ul
100 };
101
102 #define __ac_isempty(flag, i) ((flag[i>>4]>>((i&0xfU)<<1))&2)
103 #define __ac_isdel(flag, i) ((flag[i>>4]>>((i&0xfU)<<1))&1)
104 #define __ac_iseither(flag, i) ((flag[i>>4]>>((i&0xfU)<<1))&3)
105 #define __ac_set_isdel_false(flag, i) (flag[i>>4]&=~(1ul<<((i&0xfU)<<1)))
106 #define __ac_set_isempty_false(flag, i) (flag[i>>4]&=~(2ul<<((i&0xfU)<<1)))
107 #define __ac_set_isboth_false(flag, i) (flag[i>>4]&=~(3ul<<((i&0xfU)<<1)))
108 #define __ac_set_isdel_true(flag, i) (flag[i>>4]|=1ul<<((i&0xfU)<<1))
109
110 static const double __ac_HASH_UPPER = 0.77;
111
112 #define KHASH_INIT(name, khkey_t, khval_t, kh_is_map, __hash_func, __hash_equal) \
113 typedef struct { \
114 khint_t n_buckets, size, n_occupied, upper_bound; \
115 uint32_t *flags; \
116 khkey_t *keys; \
117 khval_t *vals; \
118 } kh_##name##_t; \
119 static inline kh_##name##_t *kh_init_##name() { \
120 return (kh_##name##_t*)calloc(1,sizeof(kh_##name##_t)); \
121 } \
122 static inline void kh_destroy_##name(kh_##name##_t *h) \
123 { \
124 if (h) { \
125 free(h->keys); \
126 free(h->flags); \
127 free(h->vals); \
128 /*free(h);*/ \
129 } \
130 } \
131 static inline void kh_clear_##name(kh_##name##_t *h) \
132 { \
133 if (h && h->flags) { \
134 memset(h->flags, 0xaa, ((h->n_buckets>>4) + 1) * sizeof(uint32_t)); \
135 h->size = h->n_occupied = 0; \
136 } \
137 } \
138 static inline khint_t kh_get_##name(kh_##name##_t *h, khkey_t key) \
139 { \
140 if (h->n_buckets) { \
141 khint_t inc, k, i, last; \
142 k = __hash_func(key); i = k % h->n_buckets; \
143 inc = 1 + k % (h->n_buckets - 1); last = i; \
144 while (!__ac_isempty(h->flags, i) && (__ac_isdel(h->flags, i) || !__hash_equal(h->keys[i], key))) { \
145 if (i + inc >= h->n_buckets) i = i + inc - h->n_buckets; \
146 else i += inc; \
147 if (i == last) return h->n_buckets; \
148 } \
149 return __ac_iseither(h->flags, i)? h->n_buckets : i; \
150 } else return 0; \
151 } \
152 static inline void kh_resize_##name(kh_##name##_t *h, khint_t new_n_buckets) \
153 { \
154 uint32_t *new_flags = 0; \
155 khint_t j = 1; \
156 { \
157 khint_t t = __ac_HASH_PRIME_SIZE - 1; \
158 while (__ac_prime_list[t] > new_n_buckets) --t; \
159 new_n_buckets = __ac_prime_list[t+1]; \
160 if (h->size >= (khint_t)(new_n_buckets * __ac_HASH_UPPER + 0.5)) j = 0; \
161 else { \
162 new_flags = (uint32_t*)malloc(((new_n_buckets>>4) + 1) * sizeof(uint32_t)); \
163 memset(new_flags, 0xaa, ((new_n_buckets>>4) + 1) * sizeof(uint32_t)); \
164 if (h->n_buckets < new_n_buckets) { \
165 h->keys = (khkey_t*)realloc(h->keys, new_n_buckets * sizeof(khkey_t)); \
166 if (kh_is_map) \
167 h->vals = (khval_t*)realloc(h->vals, new_n_buckets * sizeof(khval_t)); \
168 } \
169 } \
170 } \
171 if (j) { \
172 for (j = 0; j != h->n_buckets; ++j) { \
173 if (__ac_iseither(h->flags, j) == 0) { \
174 khkey_t key = h->keys[j]; \
175 khval_t val; \
176 if (kh_is_map) val = h->vals[j]; \
177 __ac_set_isdel_true(h->flags, j); \
178 while (1) { \
179 khint_t inc, k, i; \
180 k = __hash_func(key); \
181 i = k % new_n_buckets; \
182 inc = 1 + k % (new_n_buckets - 1); \
183 while (!__ac_isempty(new_flags, i)) { \
184 if (i + inc >= new_n_buckets) i = i + inc - new_n_buckets; \
185 else i += inc; \
186 } \
187 __ac_set_isempty_false(new_flags, i); \
188 if (i < h->n_buckets && __ac_iseither(h->flags, i) == 0) { \
189 { khkey_t tmp = h->keys[i]; h->keys[i] = key; key = tmp; } \
190 if (kh_is_map) { khval_t tmp = h->vals[i]; h->vals[i] = val; val = tmp; } \
191 __ac_set_isdel_true(h->flags, i); \
192 } else { \
193 h->keys[i] = key; \
194 if (kh_is_map) h->vals[i] = val; \
195 break; \
196 } \
197 } \
198 } \
199 } \
200 if (h->n_buckets > new_n_buckets) { \
201 h->keys = (khkey_t*)realloc(h->keys, new_n_buckets * sizeof(khkey_t)); \
202 if (kh_is_map) \
203 h->vals = (khval_t*)realloc(h->vals, new_n_buckets * sizeof(khval_t)); \
204 } \
205 free(h->flags); \
206 h->flags = new_flags; \
207 h->n_buckets = new_n_buckets; \
208 h->n_occupied = h->size; \
209 h->upper_bound = (khint_t)(h->n_buckets * __ac_HASH_UPPER + 0.5); \
210 } \
211 } \
212 static inline khint_t kh_put_##name(kh_##name##_t *h, khkey_t key, int *ret) \
213 { \
214 khint_t x; \
215 if (h->n_occupied >= h->upper_bound) { \
216 if (h->n_buckets > (h->size<<1)) kh_resize_##name(h, h->n_buckets - 1); \
217 else kh_resize_##name(h, h->n_buckets + 1); \
218 } \
219 { \
220 khint_t inc, k, i, site, last; \
221 x = site = h->n_buckets; k = __hash_func(key); i = k % h->n_buckets; \
222 if (__ac_isempty(h->flags, i)) x = i; \
223 else { \
224 inc = 1 + k % (h->n_buckets - 1); last = i; \
225 while (!__ac_isempty(h->flags, i) && (__ac_isdel(h->flags, i) || !__hash_equal(h->keys[i], key))) { \
226 if (__ac_isdel(h->flags, i)) site = i; \
227 if (i + inc >= h->n_buckets) i = i + inc - h->n_buckets; \
228 else i += inc; \
229 if (i == last) { x = site; break; } \
230 } \
231 if (x == h->n_buckets) { \
232 if (__ac_isempty(h->flags, i) && site != h->n_buckets) x = site; \
233 else x = i; \
234 } \
235 } \
236 } \
237 if (__ac_isempty(h->flags, x)) { \
238 h->keys[x] = key; \
239 __ac_set_isboth_false(h->flags, x); \
240 ++h->size; ++h->n_occupied; \
241 *ret = 1; \
242 } else if (__ac_isdel(h->flags, x)) { \
243 h->keys[x] = key; \
244 __ac_set_isboth_false(h->flags, x); \
245 ++h->size; \
246 *ret = 2; \
247 } else *ret = 0; \
248 return x; \
249 } \
250 static inline void kh_del_##name(kh_##name##_t *h, khint_t x) \
251 { \
252 if (x != h->n_buckets && !__ac_iseither(h->flags, x)) { \
253 __ac_set_isdel_true(h->flags, x); \
254 --h->size; \
255 } \
256 }
257
258 /* --- BEGIN OF HASH FUNCTIONS --- */
259
260 #define kh_int_hash_func(key) (uint32_t)(key)
261 #define kh_int_hash_equal(a, b) (a == b)
262 #define kh_int64_hash_func(key) (uint32_t)((key)>>33^(key)^(key)<<11)
263 #define kh_int64_hash_equal(a, b) (a == b)
__ac_X31_hash_string(const char * s)264 static inline khint_t __ac_X31_hash_string(const char *s)
265 {
266 khint_t h = *s;
267 if (h) for (++s ; *s; ++s) h = (h << 5) - h + *s;
268 return h;
269 }
270 #define kh_str_hash_func(key) __ac_X31_hash_string(key)
271 #define kh_str_hash_equal(a, b) (strcmp(a, b) == 0)
272
273 /* --- END OF HASH FUNCTIONS --- */
274
275 /* Other necessary macros... */
276
277 #define khash_t(name) kh_##name##_t
278
279 #define kh_init(name) kh_init_##name()
280 #define kh_destroy(name, h) kh_destroy_##name(h)
281 #define kh_clear(name, h) kh_clear_##name(h)
282 #define kh_resize(name, h, s) kh_resize_##name(h, s)
283 #define kh_put(name, h, k, r) kh_put_##name(h, k, r)
284 #define kh_get(name, h, k) kh_get_##name(h, k)
285 #define kh_del(name, h, k) kh_del_##name(h, k)
286
287 #define kh_exist(h, x) (!__ac_iseither((h)->flags, (x)))
288 #define kh_key(h, x) ((h)->keys[x])
289 #define kh_val(h, x) ((h)->vals[x])
290 #define kh_value(h, x) ((h)->vals[x])
291 #define kh_begin(h) (khint_t)(0)
292 #define kh_end(h) ((h)->n_buckets)
293 #define kh_size(h) ((h)->size)
294 #define kh_n_buckets(h) ((h)->n_buckets)
295
296 /* More conenient interfaces */
297
298 #define KHASH_SET_INIT_INT(name) \
299 KHASH_INIT(name, uint32_t, char, 0, kh_int_hash_func, kh_int_hash_equal)
300
301 #define KHASH_MAP_INIT_INT(name, khval_t) \
302 KHASH_INIT(name, uint32_t, khval_t, 1, kh_int_hash_func, kh_int_hash_equal)
303
304 #define KHASH_SET_INIT_INT64(name) \
305 KHASH_INIT(name, uint64_t, char, 0, kh_int64_hash_func, kh_int64_hash_equal)
306
307 #define KHASH_MAP_INIT_INT64(name, khval_t) \
308 KHASH_INIT(name, uint64_t, khval_t, 1, kh_int64_hash_func, kh_int64_hash_equal)
309
310 typedef const char *kh_cstr_t;
311 #define KHASH_SET_INIT_STR(name) \
312 KHASH_INIT(name, kh_cstr_t, char, 0, kh_str_hash_func, kh_str_hash_equal)
313
314 #define KHASH_MAP_INIT_STR(name, khval_t) \
315 KHASH_INIT(name, kh_cstr_t, khval_t, 1, kh_str_hash_func, kh_str_hash_equal)
316
317 #endif /* __AC_KHASH_H */
318