annotate contrib/python-zstandard/zstd/dictBuilder/cover.c @ 35320:d901a88891fe

rebase: rerun a rebase on-disk if IMM merge conflicts arise Differential Revision: https://phab.mercurial-scm.org/D1249
author Phil Cohen <phillco@fb.com>
date Fri, 08 Dec 2017 12:14:38 -0800
parents c32454d69b85
children b1fb341d8a61
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
30895
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
2 * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
3 * All rights reserved.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
4 *
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
5 * This source code is licensed under the BSD-style license found in the
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
6 * LICENSE file in the root directory of this source tree. An additional grant
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
7 * of patent rights can be found in the PATENTS file in the same directory.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
8 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
9
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
10 /*-*************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
11 * Dependencies
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
12 ***************************************/
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
13 #include <stdio.h> /* fprintf */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
14 #include <stdlib.h> /* malloc, free, qsort */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
15 #include <string.h> /* memset */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
16 #include <time.h> /* clock */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
17
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
18 #include "mem.h" /* read */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
19 #include "pool.h"
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
20 #include "threading.h"
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
21 #include "zstd_internal.h" /* includes zstd.h */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
22 #ifndef ZDICT_STATIC_LINKING_ONLY
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
23 #define ZDICT_STATIC_LINKING_ONLY
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
24 #endif
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
25 #include "zdict.h"
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
26
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
27 /*-*************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
28 * Constants
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
29 ***************************************/
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
30 #define COVER_MAX_SAMPLES_SIZE (sizeof(size_t) == 8 ? ((U32)-1) : ((U32)1 GB))
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
31
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
32 /*-*************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
33 * Console display
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
34 ***************************************/
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
35 static int g_displayLevel = 2;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
36 #define DISPLAY(...) \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
37 { \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
38 fprintf(stderr, __VA_ARGS__); \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
39 fflush(stderr); \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
40 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
41 #define LOCALDISPLAYLEVEL(displayLevel, l, ...) \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
42 if (displayLevel >= l) { \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
43 DISPLAY(__VA_ARGS__); \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
44 } /* 0 : no display; 1: errors; 2: default; 3: details; 4: debug */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
45 #define DISPLAYLEVEL(l, ...) LOCALDISPLAYLEVEL(g_displayLevel, l, __VA_ARGS__)
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
46
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
47 #define LOCALDISPLAYUPDATE(displayLevel, l, ...) \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
48 if (displayLevel >= l) { \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
49 if ((clock() - g_time > refreshRate) || (displayLevel >= 4)) { \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
50 g_time = clock(); \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
51 DISPLAY(__VA_ARGS__); \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
52 if (displayLevel >= 4) \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
53 fflush(stdout); \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
54 } \
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
55 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
56 #define DISPLAYUPDATE(l, ...) LOCALDISPLAYUPDATE(g_displayLevel, l, __VA_ARGS__)
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
57 static const clock_t refreshRate = CLOCKS_PER_SEC * 15 / 100;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
58 static clock_t g_time = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
59
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
60 /*-*************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
61 * Hash table
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
62 ***************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
63 * A small specialized hash map for storing activeDmers.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
64 * The map does not resize, so if it becomes full it will loop forever.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
65 * Thus, the map must be large enough to store every value.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
66 * The map implements linear probing and keeps its load less than 0.5.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
67 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
68
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
69 #define MAP_EMPTY_VALUE ((U32)-1)
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
70 typedef struct COVER_map_pair_t_s {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
71 U32 key;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
72 U32 value;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
73 } COVER_map_pair_t;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
74
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
75 typedef struct COVER_map_s {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
76 COVER_map_pair_t *data;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
77 U32 sizeLog;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
78 U32 size;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
79 U32 sizeMask;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
80 } COVER_map_t;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
81
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
82 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
83 * Clear the map.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
84 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
85 static void COVER_map_clear(COVER_map_t *map) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
86 memset(map->data, MAP_EMPTY_VALUE, map->size * sizeof(COVER_map_pair_t));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
87 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
88
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
89 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
90 * Initializes a map of the given size.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
91 * Returns 1 on success and 0 on failure.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
92 * The map must be destroyed with COVER_map_destroy().
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
93 * The map is only guaranteed to be large enough to hold size elements.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
94 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
95 static int COVER_map_init(COVER_map_t *map, U32 size) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
96 map->sizeLog = ZSTD_highbit32(size) + 2;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
97 map->size = (U32)1 << map->sizeLog;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
98 map->sizeMask = map->size - 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
99 map->data = (COVER_map_pair_t *)malloc(map->size * sizeof(COVER_map_pair_t));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
100 if (!map->data) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
101 map->sizeLog = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
102 map->size = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
103 return 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
104 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
105 COVER_map_clear(map);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
106 return 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
107 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
108
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
109 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
110 * Internal hash function
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
111 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
112 static const U32 prime4bytes = 2654435761U;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
113 static U32 COVER_map_hash(COVER_map_t *map, U32 key) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
114 return (key * prime4bytes) >> (32 - map->sizeLog);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
115 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
116
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
117 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
118 * Helper function that returns the index that a key should be placed into.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
119 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
120 static U32 COVER_map_index(COVER_map_t *map, U32 key) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
121 const U32 hash = COVER_map_hash(map, key);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
122 U32 i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
123 for (i = hash;; i = (i + 1) & map->sizeMask) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
124 COVER_map_pair_t *pos = &map->data[i];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
125 if (pos->value == MAP_EMPTY_VALUE) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
126 return i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
127 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
128 if (pos->key == key) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
129 return i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
130 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
131 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
132 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
133
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
134 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
135 * Returns the pointer to the value for key.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
136 * If key is not in the map, it is inserted and the value is set to 0.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
137 * The map must not be full.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
138 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
139 static U32 *COVER_map_at(COVER_map_t *map, U32 key) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
140 COVER_map_pair_t *pos = &map->data[COVER_map_index(map, key)];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
141 if (pos->value == MAP_EMPTY_VALUE) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
142 pos->key = key;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
143 pos->value = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
144 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
145 return &pos->value;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
146 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
147
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
148 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
149 * Deletes key from the map if present.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
150 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
151 static void COVER_map_remove(COVER_map_t *map, U32 key) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
152 U32 i = COVER_map_index(map, key);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
153 COVER_map_pair_t *del = &map->data[i];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
154 U32 shift = 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
155 if (del->value == MAP_EMPTY_VALUE) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
156 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
157 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
158 for (i = (i + 1) & map->sizeMask;; i = (i + 1) & map->sizeMask) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
159 COVER_map_pair_t *const pos = &map->data[i];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
160 /* If the position is empty we are done */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
161 if (pos->value == MAP_EMPTY_VALUE) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
162 del->value = MAP_EMPTY_VALUE;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
163 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
164 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
165 /* If pos can be moved to del do so */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
166 if (((i - COVER_map_hash(map, pos->key)) & map->sizeMask) >= shift) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
167 del->key = pos->key;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
168 del->value = pos->value;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
169 del = pos;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
170 shift = 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
171 } else {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
172 ++shift;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
173 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
174 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
175 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
176
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
177 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
178 * Destroyes a map that is inited with COVER_map_init().
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
179 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
180 static void COVER_map_destroy(COVER_map_t *map) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
181 if (map->data) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
182 free(map->data);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
183 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
184 map->data = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
185 map->size = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
186 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
187
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
188 /*-*************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
189 * Context
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
190 ***************************************/
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
191
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
192 typedef struct {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
193 const BYTE *samples;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
194 size_t *offsets;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
195 const size_t *samplesSizes;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
196 size_t nbSamples;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
197 U32 *suffix;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
198 size_t suffixSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
199 U32 *freqs;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
200 U32 *dmerAt;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
201 unsigned d;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
202 } COVER_ctx_t;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
203
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
204 /* We need a global context for qsort... */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
205 static COVER_ctx_t *g_ctx = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
206
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
207 /*-*************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
208 * Helper functions
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
209 ***************************************/
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
210
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
211 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
212 * Returns the sum of the sample sizes.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
213 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
214 static size_t COVER_sum(const size_t *samplesSizes, unsigned nbSamples) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
215 size_t sum = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
216 size_t i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
217 for (i = 0; i < nbSamples; ++i) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
218 sum += samplesSizes[i];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
219 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
220 return sum;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
221 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
222
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
223 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
224 * Returns -1 if the dmer at lp is less than the dmer at rp.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
225 * Return 0 if the dmers at lp and rp are equal.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
226 * Returns 1 if the dmer at lp is greater than the dmer at rp.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
227 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
228 static int COVER_cmp(COVER_ctx_t *ctx, const void *lp, const void *rp) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
229 const U32 lhs = *(const U32 *)lp;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
230 const U32 rhs = *(const U32 *)rp;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
231 return memcmp(ctx->samples + lhs, ctx->samples + rhs, ctx->d);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
232 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
233
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
234 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
235 * Same as COVER_cmp() except ties are broken by pointer value
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
236 * NOTE: g_ctx must be set to call this function. A global is required because
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
237 * qsort doesn't take an opaque pointer.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
238 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
239 static int COVER_strict_cmp(const void *lp, const void *rp) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
240 int result = COVER_cmp(g_ctx, lp, rp);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
241 if (result == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
242 result = lp < rp ? -1 : 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
243 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
244 return result;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
245 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
246
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
247 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
248 * Returns the first pointer in [first, last) whose element does not compare
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
249 * less than value. If no such element exists it returns last.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
250 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
251 static const size_t *COVER_lower_bound(const size_t *first, const size_t *last,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
252 size_t value) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
253 size_t count = last - first;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
254 while (count != 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
255 size_t step = count / 2;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
256 const size_t *ptr = first;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
257 ptr += step;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
258 if (*ptr < value) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
259 first = ++ptr;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
260 count -= step + 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
261 } else {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
262 count = step;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
263 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
264 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
265 return first;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
266 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
267
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
268 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
269 * Generic groupBy function.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
270 * Groups an array sorted by cmp into groups with equivalent values.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
271 * Calls grp for each group.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
272 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
273 static void
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
274 COVER_groupBy(const void *data, size_t count, size_t size, COVER_ctx_t *ctx,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
275 int (*cmp)(COVER_ctx_t *, const void *, const void *),
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
276 void (*grp)(COVER_ctx_t *, const void *, const void *)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
277 const BYTE *ptr = (const BYTE *)data;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
278 size_t num = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
279 while (num < count) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
280 const BYTE *grpEnd = ptr + size;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
281 ++num;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
282 while (num < count && cmp(ctx, ptr, grpEnd) == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
283 grpEnd += size;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
284 ++num;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
285 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
286 grp(ctx, ptr, grpEnd);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
287 ptr = grpEnd;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
288 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
289 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
290
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
291 /*-*************************************
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
292 * Cover functions
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
293 ***************************************/
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
294
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
295 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
296 * Called on each group of positions with the same dmer.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
297 * Counts the frequency of each dmer and saves it in the suffix array.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
298 * Fills `ctx->dmerAt`.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
299 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
300 static void COVER_group(COVER_ctx_t *ctx, const void *group,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
301 const void *groupEnd) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
302 /* The group consists of all the positions with the same first d bytes. */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
303 const U32 *grpPtr = (const U32 *)group;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
304 const U32 *grpEnd = (const U32 *)groupEnd;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
305 /* The dmerId is how we will reference this dmer.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
306 * This allows us to map the whole dmer space to a much smaller space, the
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
307 * size of the suffix array.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
308 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
309 const U32 dmerId = (U32)(grpPtr - ctx->suffix);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
310 /* Count the number of samples this dmer shows up in */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
311 U32 freq = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
312 /* Details */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
313 const size_t *curOffsetPtr = ctx->offsets;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
314 const size_t *offsetsEnd = ctx->offsets + ctx->nbSamples;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
315 /* Once *grpPtr >= curSampleEnd this occurrence of the dmer is in a
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
316 * different sample than the last.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
317 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
318 size_t curSampleEnd = ctx->offsets[0];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
319 for (; grpPtr != grpEnd; ++grpPtr) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
320 /* Save the dmerId for this position so we can get back to it. */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
321 ctx->dmerAt[*grpPtr] = dmerId;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
322 /* Dictionaries only help for the first reference to the dmer.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
323 * After that zstd can reference the match from the previous reference.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
324 * So only count each dmer once for each sample it is in.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
325 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
326 if (*grpPtr < curSampleEnd) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
327 continue;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
328 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
329 freq += 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
330 /* Binary search to find the end of the sample *grpPtr is in.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
331 * In the common case that grpPtr + 1 == grpEnd we can skip the binary
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
332 * search because the loop is over.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
333 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
334 if (grpPtr + 1 != grpEnd) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
335 const size_t *sampleEndPtr =
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
336 COVER_lower_bound(curOffsetPtr, offsetsEnd, *grpPtr);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
337 curSampleEnd = *sampleEndPtr;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
338 curOffsetPtr = sampleEndPtr + 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
339 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
340 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
341 /* At this point we are never going to look at this segment of the suffix
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
342 * array again. We take advantage of this fact to save memory.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
343 * We store the frequency of the dmer in the first position of the group,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
344 * which is dmerId.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
345 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
346 ctx->suffix[dmerId] = freq;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
347 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
348
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
349 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
350 * A segment is a range in the source as well as the score of the segment.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
351 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
352 typedef struct {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
353 U32 begin;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
354 U32 end;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
355 double score;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
356 } COVER_segment_t;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
357
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
358 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
359 * Selects the best segment in an epoch.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
360 * Segments of are scored according to the function:
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
361 *
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
362 * Let F(d) be the frequency of dmer d.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
363 * Let S_i be the dmer at position i of segment S which has length k.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
364 *
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
365 * Score(S) = F(S_1) + F(S_2) + ... + F(S_{k-d+1})
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
366 *
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
367 * Once the dmer d is in the dictionay we set F(d) = 0.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
368 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
369 static COVER_segment_t COVER_selectSegment(const COVER_ctx_t *ctx, U32 *freqs,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
370 COVER_map_t *activeDmers, U32 begin,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
371 U32 end, COVER_params_t parameters) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
372 /* Constants */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
373 const U32 k = parameters.k;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
374 const U32 d = parameters.d;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
375 const U32 dmersInK = k - d + 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
376 /* Try each segment (activeSegment) and save the best (bestSegment) */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
377 COVER_segment_t bestSegment = {0, 0, 0};
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
378 COVER_segment_t activeSegment;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
379 /* Reset the activeDmers in the segment */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
380 COVER_map_clear(activeDmers);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
381 /* The activeSegment starts at the beginning of the epoch. */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
382 activeSegment.begin = begin;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
383 activeSegment.end = begin;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
384 activeSegment.score = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
385 /* Slide the activeSegment through the whole epoch.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
386 * Save the best segment in bestSegment.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
387 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
388 while (activeSegment.end < end) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
389 /* The dmerId for the dmer at the next position */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
390 U32 newDmer = ctx->dmerAt[activeSegment.end];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
391 /* The entry in activeDmers for this dmerId */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
392 U32 *newDmerOcc = COVER_map_at(activeDmers, newDmer);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
393 /* If the dmer isn't already present in the segment add its score. */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
394 if (*newDmerOcc == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
395 /* The paper suggest using the L-0.5 norm, but experiments show that it
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
396 * doesn't help.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
397 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
398 activeSegment.score += freqs[newDmer];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
399 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
400 /* Add the dmer to the segment */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
401 activeSegment.end += 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
402 *newDmerOcc += 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
403
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
404 /* If the window is now too large, drop the first position */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
405 if (activeSegment.end - activeSegment.begin == dmersInK + 1) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
406 U32 delDmer = ctx->dmerAt[activeSegment.begin];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
407 U32 *delDmerOcc = COVER_map_at(activeDmers, delDmer);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
408 activeSegment.begin += 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
409 *delDmerOcc -= 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
410 /* If this is the last occurence of the dmer, subtract its score */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
411 if (*delDmerOcc == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
412 COVER_map_remove(activeDmers, delDmer);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
413 activeSegment.score -= freqs[delDmer];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
414 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
415 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
416
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
417 /* If this segment is the best so far save it */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
418 if (activeSegment.score > bestSegment.score) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
419 bestSegment = activeSegment;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
420 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
421 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
422 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
423 /* Trim off the zero frequency head and tail from the segment. */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
424 U32 newBegin = bestSegment.end;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
425 U32 newEnd = bestSegment.begin;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
426 U32 pos;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
427 for (pos = bestSegment.begin; pos != bestSegment.end; ++pos) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
428 U32 freq = freqs[ctx->dmerAt[pos]];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
429 if (freq != 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
430 newBegin = MIN(newBegin, pos);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
431 newEnd = pos + 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
432 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
433 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
434 bestSegment.begin = newBegin;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
435 bestSegment.end = newEnd;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
436 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
437 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
438 /* Zero out the frequency of each dmer covered by the chosen segment. */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
439 U32 pos;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
440 for (pos = bestSegment.begin; pos != bestSegment.end; ++pos) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
441 freqs[ctx->dmerAt[pos]] = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
442 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
443 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
444 return bestSegment;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
445 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
446
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
447 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
448 * Check the validity of the parameters.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
449 * Returns non-zero if the parameters are valid and 0 otherwise.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
450 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
451 static int COVER_checkParameters(COVER_params_t parameters) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
452 /* k and d are required parameters */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
453 if (parameters.d == 0 || parameters.k == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
454 return 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
455 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
456 /* d <= k */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
457 if (parameters.d > parameters.k) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
458 return 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
459 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
460 return 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
461 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
462
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
463 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
464 * Clean up a context initialized with `COVER_ctx_init()`.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
465 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
466 static void COVER_ctx_destroy(COVER_ctx_t *ctx) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
467 if (!ctx) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
468 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
469 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
470 if (ctx->suffix) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
471 free(ctx->suffix);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
472 ctx->suffix = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
473 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
474 if (ctx->freqs) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
475 free(ctx->freqs);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
476 ctx->freqs = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
477 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
478 if (ctx->dmerAt) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
479 free(ctx->dmerAt);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
480 ctx->dmerAt = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
481 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
482 if (ctx->offsets) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
483 free(ctx->offsets);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
484 ctx->offsets = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
485 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
486 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
487
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
488 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
489 * Prepare a context for dictionary building.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
490 * The context is only dependent on the parameter `d` and can used multiple
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
491 * times.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
492 * Returns 1 on success or zero on error.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
493 * The context must be destroyed with `COVER_ctx_destroy()`.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
494 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
495 static int COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
496 const size_t *samplesSizes, unsigned nbSamples,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
497 unsigned d) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
498 const BYTE *const samples = (const BYTE *)samplesBuffer;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
499 const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
500 /* Checks */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
501 if (totalSamplesSize < d ||
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
502 totalSamplesSize >= (size_t)COVER_MAX_SAMPLES_SIZE) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
503 DISPLAYLEVEL(1, "Total samples size is too large, maximum size is %u MB\n",
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
504 (COVER_MAX_SAMPLES_SIZE >> 20));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
505 return 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
506 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
507 /* Zero the context */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
508 memset(ctx, 0, sizeof(*ctx));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
509 DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbSamples,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
510 (U32)totalSamplesSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
511 ctx->samples = samples;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
512 ctx->samplesSizes = samplesSizes;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
513 ctx->nbSamples = nbSamples;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
514 /* Partial suffix array */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
515 ctx->suffixSize = totalSamplesSize - d + 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
516 ctx->suffix = (U32 *)malloc(ctx->suffixSize * sizeof(U32));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
517 /* Maps index to the dmerID */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
518 ctx->dmerAt = (U32 *)malloc(ctx->suffixSize * sizeof(U32));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
519 /* The offsets of each file */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
520 ctx->offsets = (size_t *)malloc((nbSamples + 1) * sizeof(size_t));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
521 if (!ctx->suffix || !ctx->dmerAt || !ctx->offsets) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
522 DISPLAYLEVEL(1, "Failed to allocate scratch buffers\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
523 COVER_ctx_destroy(ctx);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
524 return 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
525 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
526 ctx->freqs = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
527 ctx->d = d;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
528
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
529 /* Fill offsets from the samlesSizes */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
530 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
531 U32 i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
532 ctx->offsets[0] = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
533 for (i = 1; i <= nbSamples; ++i) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
534 ctx->offsets[i] = ctx->offsets[i - 1] + samplesSizes[i - 1];
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
535 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
536 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
537 DISPLAYLEVEL(2, "Constructing partial suffix array\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
538 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
539 /* suffix is a partial suffix array.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
540 * It only sorts suffixes by their first parameters.d bytes.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
541 * The sort is stable, so each dmer group is sorted by position in input.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
542 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
543 U32 i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
544 for (i = 0; i < ctx->suffixSize; ++i) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
545 ctx->suffix[i] = i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
546 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
547 /* qsort doesn't take an opaque pointer, so pass as a global */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
548 g_ctx = ctx;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
549 qsort(ctx->suffix, ctx->suffixSize, sizeof(U32), &COVER_strict_cmp);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
550 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
551 DISPLAYLEVEL(2, "Computing frequencies\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
552 /* For each dmer group (group of positions with the same first d bytes):
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
553 * 1. For each position we set dmerAt[position] = dmerID. The dmerID is
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
554 * (groupBeginPtr - suffix). This allows us to go from position to
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
555 * dmerID so we can look up values in freq.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
556 * 2. We calculate how many samples the dmer occurs in and save it in
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
557 * freqs[dmerId].
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
558 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
559 COVER_groupBy(ctx->suffix, ctx->suffixSize, sizeof(U32), ctx, &COVER_cmp,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
560 &COVER_group);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
561 ctx->freqs = ctx->suffix;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
562 ctx->suffix = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
563 return 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
564 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
565
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
566 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
567 * Given the prepared context build the dictionary.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
568 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
569 static size_t COVER_buildDictionary(const COVER_ctx_t *ctx, U32 *freqs,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
570 COVER_map_t *activeDmers, void *dictBuffer,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
571 size_t dictBufferCapacity,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
572 COVER_params_t parameters) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
573 BYTE *const dict = (BYTE *)dictBuffer;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
574 size_t tail = dictBufferCapacity;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
575 /* Divide the data up into epochs of equal size.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
576 * We will select at least one segment from each epoch.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
577 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
578 const U32 epochs = (U32)(dictBufferCapacity / parameters.k);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
579 const U32 epochSize = (U32)(ctx->suffixSize / epochs);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
580 size_t epoch;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
581 DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n", epochs,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
582 epochSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
583 /* Loop through the epochs until there are no more segments or the dictionary
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
584 * is full.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
585 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
586 for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
587 const U32 epochBegin = (U32)(epoch * epochSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
588 const U32 epochEnd = epochBegin + epochSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
589 size_t segmentSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
590 /* Select a segment */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
591 COVER_segment_t segment = COVER_selectSegment(
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
592 ctx, freqs, activeDmers, epochBegin, epochEnd, parameters);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
593 /* Trim the segment if necessary and if it is empty then we are done */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
594 segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
595 if (segmentSize == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
596 break;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
597 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
598 /* We fill the dictionary from the back to allow the best segments to be
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
599 * referenced with the smallest offsets.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
600 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
601 tail -= segmentSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
602 memcpy(dict + tail, ctx->samples + segment.begin, segmentSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
603 DISPLAYUPDATE(
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
604 2, "\r%u%% ",
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
605 (U32)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
606 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
607 DISPLAYLEVEL(2, "\r%79s\r", "");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
608 return tail;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
609 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
610
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
611 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
612 * Translate from COVER_params_t to ZDICT_params_t required for finalizing the
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
613 * dictionary.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
614 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
615 static ZDICT_params_t COVER_translateParams(COVER_params_t parameters) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
616 ZDICT_params_t zdictParams;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
617 memset(&zdictParams, 0, sizeof(zdictParams));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
618 zdictParams.notificationLevel = 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
619 zdictParams.dictID = parameters.dictID;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
620 zdictParams.compressionLevel = parameters.compressionLevel;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
621 return zdictParams;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
622 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
623
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
624 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
625 * Constructs a dictionary using a heuristic based on the following paper:
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
626 *
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
627 * Liao, Petri, Moffat, Wirth
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
628 * Effective Construction of Relative Lempel-Ziv Dictionaries
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
629 * Published in WWW 2016.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
630 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
631 ZDICTLIB_API size_t COVER_trainFromBuffer(
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
632 void *dictBuffer, size_t dictBufferCapacity, const void *samplesBuffer,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
633 const size_t *samplesSizes, unsigned nbSamples, COVER_params_t parameters) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
634 BYTE *const dict = (BYTE *)dictBuffer;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
635 COVER_ctx_t ctx;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
636 COVER_map_t activeDmers;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
637 /* Checks */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
638 if (!COVER_checkParameters(parameters)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
639 DISPLAYLEVEL(1, "Cover parameters incorrect\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
640 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
641 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
642 if (nbSamples == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
643 DISPLAYLEVEL(1, "Cover must have at least one input file\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
644 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
645 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
646 if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
647 DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
648 ZDICT_DICTSIZE_MIN);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
649 return ERROR(dstSize_tooSmall);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
650 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
651 /* Initialize global data */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
652 g_displayLevel = parameters.notificationLevel;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
653 /* Initialize context and activeDmers */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
654 if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
655 parameters.d)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
656 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
657 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
658 if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
659 DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
660 COVER_ctx_destroy(&ctx);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
661 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
662 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
663
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
664 DISPLAYLEVEL(2, "Building dictionary\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
665 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
666 const size_t tail =
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
667 COVER_buildDictionary(&ctx, ctx.freqs, &activeDmers, dictBuffer,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
668 dictBufferCapacity, parameters);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
669 ZDICT_params_t zdictParams = COVER_translateParams(parameters);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
670 const size_t dictionarySize = ZDICT_finalizeDictionary(
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
671 dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
672 samplesBuffer, samplesSizes, nbSamples, zdictParams);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
673 if (!ZSTD_isError(dictionarySize)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
674 DISPLAYLEVEL(2, "Constructed dictionary of size %u\n",
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
675 (U32)dictionarySize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
676 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
677 COVER_ctx_destroy(&ctx);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
678 COVER_map_destroy(&activeDmers);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
679 return dictionarySize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
680 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
681 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
682
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
683 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
684 * COVER_best_t is used for two purposes:
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
685 * 1. Synchronizing threads.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
686 * 2. Saving the best parameters and dictionary.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
687 *
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
688 * All of the methods except COVER_best_init() are thread safe if zstd is
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
689 * compiled with multithreaded support.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
690 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
691 typedef struct COVER_best_s {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
692 pthread_mutex_t mutex;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
693 pthread_cond_t cond;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
694 size_t liveJobs;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
695 void *dict;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
696 size_t dictSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
697 COVER_params_t parameters;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
698 size_t compressedSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
699 } COVER_best_t;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
700
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
701 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
702 * Initialize the `COVER_best_t`.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
703 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
704 static void COVER_best_init(COVER_best_t *best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
705 if (!best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
706 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
707 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
708 pthread_mutex_init(&best->mutex, NULL);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
709 pthread_cond_init(&best->cond, NULL);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
710 best->liveJobs = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
711 best->dict = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
712 best->dictSize = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
713 best->compressedSize = (size_t)-1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
714 memset(&best->parameters, 0, sizeof(best->parameters));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
715 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
716
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
717 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
718 * Wait until liveJobs == 0.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
719 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
720 static void COVER_best_wait(COVER_best_t *best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
721 if (!best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
722 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
723 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
724 pthread_mutex_lock(&best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
725 while (best->liveJobs != 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
726 pthread_cond_wait(&best->cond, &best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
727 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
728 pthread_mutex_unlock(&best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
729 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
730
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
731 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
732 * Call COVER_best_wait() and then destroy the COVER_best_t.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
733 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
734 static void COVER_best_destroy(COVER_best_t *best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
735 if (!best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
736 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
737 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
738 COVER_best_wait(best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
739 if (best->dict) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
740 free(best->dict);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
741 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
742 pthread_mutex_destroy(&best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
743 pthread_cond_destroy(&best->cond);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
744 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
745
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
746 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
747 * Called when a thread is about to be launched.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
748 * Increments liveJobs.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
749 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
750 static void COVER_best_start(COVER_best_t *best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
751 if (!best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
752 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
753 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
754 pthread_mutex_lock(&best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
755 ++best->liveJobs;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
756 pthread_mutex_unlock(&best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
757 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
758
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
759 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
760 * Called when a thread finishes executing, both on error or success.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
761 * Decrements liveJobs and signals any waiting threads if liveJobs == 0.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
762 * If this dictionary is the best so far save it and its parameters.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
763 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
764 static void COVER_best_finish(COVER_best_t *best, size_t compressedSize,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
765 COVER_params_t parameters, void *dict,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
766 size_t dictSize) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
767 if (!best) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
768 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
769 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
770 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
771 size_t liveJobs;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
772 pthread_mutex_lock(&best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
773 --best->liveJobs;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
774 liveJobs = best->liveJobs;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
775 /* If the new dictionary is better */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
776 if (compressedSize < best->compressedSize) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
777 /* Allocate space if necessary */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
778 if (!best->dict || best->dictSize < dictSize) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
779 if (best->dict) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
780 free(best->dict);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
781 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
782 best->dict = malloc(dictSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
783 if (!best->dict) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
784 best->compressedSize = ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
785 best->dictSize = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
786 return;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
787 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
788 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
789 /* Save the dictionary, parameters, and size */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
790 memcpy(best->dict, dict, dictSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
791 best->dictSize = dictSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
792 best->parameters = parameters;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
793 best->compressedSize = compressedSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
794 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
795 pthread_mutex_unlock(&best->mutex);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
796 if (liveJobs == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
797 pthread_cond_broadcast(&best->cond);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
798 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
799 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
800 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
801
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
802 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
803 * Parameters for COVER_tryParameters().
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
804 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
805 typedef struct COVER_tryParameters_data_s {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
806 const COVER_ctx_t *ctx;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
807 COVER_best_t *best;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
808 size_t dictBufferCapacity;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
809 COVER_params_t parameters;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
810 } COVER_tryParameters_data_t;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
811
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
812 /**
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
813 * Tries a set of parameters and upates the COVER_best_t with the results.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
814 * This function is thread safe if zstd is compiled with multithreaded support.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
815 * It takes its parameters as an *OWNING* opaque pointer to support threading.
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
816 */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
817 static void COVER_tryParameters(void *opaque) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
818 /* Save parameters as local variables */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
819 COVER_tryParameters_data_t *const data = (COVER_tryParameters_data_t *)opaque;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
820 const COVER_ctx_t *const ctx = data->ctx;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
821 const COVER_params_t parameters = data->parameters;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
822 size_t dictBufferCapacity = data->dictBufferCapacity;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
823 size_t totalCompressedSize = ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
824 /* Allocate space for hash table, dict, and freqs */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
825 COVER_map_t activeDmers;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
826 BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
827 U32 *freqs = (U32 *)malloc(ctx->suffixSize * sizeof(U32));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
828 if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
829 DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
830 goto _cleanup;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
831 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
832 if (!dict || !freqs) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
833 DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
834 goto _cleanup;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
835 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
836 /* Copy the frequencies because we need to modify them */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
837 memcpy(freqs, ctx->freqs, ctx->suffixSize * sizeof(U32));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
838 /* Build the dictionary */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
839 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
840 const size_t tail = COVER_buildDictionary(ctx, freqs, &activeDmers, dict,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
841 dictBufferCapacity, parameters);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
842 const ZDICT_params_t zdictParams = COVER_translateParams(parameters);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
843 dictBufferCapacity = ZDICT_finalizeDictionary(
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
844 dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
845 ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbSamples, zdictParams);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
846 if (ZDICT_isError(dictBufferCapacity)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
847 DISPLAYLEVEL(1, "Failed to finalize dictionary\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
848 goto _cleanup;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
849 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
850 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
851 /* Check total compressed size */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
852 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
853 /* Pointers */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
854 ZSTD_CCtx *cctx;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
855 ZSTD_CDict *cdict;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
856 void *dst;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
857 /* Local variables */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
858 size_t dstCapacity;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
859 size_t i;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
860 /* Allocate dst with enough space to compress the maximum sized sample */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
861 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
862 size_t maxSampleSize = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
863 for (i = 0; i < ctx->nbSamples; ++i) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
864 maxSampleSize = MAX(ctx->samplesSizes[i], maxSampleSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
865 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
866 dstCapacity = ZSTD_compressBound(maxSampleSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
867 dst = malloc(dstCapacity);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
868 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
869 /* Create the cctx and cdict */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
870 cctx = ZSTD_createCCtx();
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
871 cdict =
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
872 ZSTD_createCDict(dict, dictBufferCapacity, parameters.compressionLevel);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
873 if (!dst || !cctx || !cdict) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
874 goto _compressCleanup;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
875 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
876 /* Compress each sample and sum their sizes (or error) */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
877 totalCompressedSize = 0;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
878 for (i = 0; i < ctx->nbSamples; ++i) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
879 const size_t size = ZSTD_compress_usingCDict(
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
880 cctx, dst, dstCapacity, ctx->samples + ctx->offsets[i],
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
881 ctx->samplesSizes[i], cdict);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
882 if (ZSTD_isError(size)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
883 totalCompressedSize = ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
884 goto _compressCleanup;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
885 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
886 totalCompressedSize += size;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
887 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
888 _compressCleanup:
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
889 ZSTD_freeCCtx(cctx);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
890 ZSTD_freeCDict(cdict);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
891 if (dst) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
892 free(dst);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
893 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
894 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
895
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
896 _cleanup:
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
897 COVER_best_finish(data->best, totalCompressedSize, parameters, dict,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
898 dictBufferCapacity);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
899 free(data);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
900 COVER_map_destroy(&activeDmers);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
901 if (dict) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
902 free(dict);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
903 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
904 if (freqs) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
905 free(freqs);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
906 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
907 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
908
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
909 ZDICTLIB_API size_t COVER_optimizeTrainFromBuffer(void *dictBuffer,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
910 size_t dictBufferCapacity,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
911 const void *samplesBuffer,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
912 const size_t *samplesSizes,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
913 unsigned nbSamples,
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
914 COVER_params_t *parameters) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
915 /* constants */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
916 const unsigned nbThreads = parameters->nbThreads;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
917 const unsigned kMinD = parameters->d == 0 ? 6 : parameters->d;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
918 const unsigned kMaxD = parameters->d == 0 ? 16 : parameters->d;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
919 const unsigned kMinK = parameters->k == 0 ? kMaxD : parameters->k;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
920 const unsigned kMaxK = parameters->k == 0 ? 2048 : parameters->k;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
921 const unsigned kSteps = parameters->steps == 0 ? 32 : parameters->steps;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
922 const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
923 const unsigned kIterations =
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
924 (1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
925 /* Local variables */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
926 const int displayLevel = parameters->notificationLevel;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
927 unsigned iteration = 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
928 unsigned d;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
929 unsigned k;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
930 COVER_best_t best;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
931 POOL_ctx *pool = NULL;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
932 /* Checks */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
933 if (kMinK < kMaxD || kMaxK < kMinK) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
934 LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
935 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
936 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
937 if (nbSamples == 0) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
938 DISPLAYLEVEL(1, "Cover must have at least one input file\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
939 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
940 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
941 if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
942 DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
943 ZDICT_DICTSIZE_MIN);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
944 return ERROR(dstSize_tooSmall);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
945 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
946 if (nbThreads > 1) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
947 pool = POOL_create(nbThreads, 1);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
948 if (!pool) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
949 return ERROR(memory_allocation);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
950 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
951 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
952 /* Initialization */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
953 COVER_best_init(&best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
954 /* Turn down global display level to clean up display at level 2 and below */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
955 g_displayLevel = parameters->notificationLevel - 1;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
956 /* Loop through d first because each new value needs a new context */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
957 LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n",
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
958 kIterations);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
959 for (d = kMinD; d <= kMaxD; d += 2) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
960 /* Initialize the context for this value of d */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
961 COVER_ctx_t ctx;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
962 LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
963 if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
964 LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
965 COVER_best_destroy(&best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
966 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
967 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
968 /* Loop through k reusing the same context */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
969 for (k = kMinK; k <= kMaxK; k += kStepSize) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
970 /* Prepare the arguments */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
971 COVER_tryParameters_data_t *data = (COVER_tryParameters_data_t *)malloc(
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
972 sizeof(COVER_tryParameters_data_t));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
973 LOCALDISPLAYLEVEL(displayLevel, 3, "k=%u\n", k);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
974 if (!data) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
975 LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
976 COVER_best_destroy(&best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
977 COVER_ctx_destroy(&ctx);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
978 return ERROR(GENERIC);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
979 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
980 data->ctx = &ctx;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
981 data->best = &best;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
982 data->dictBufferCapacity = dictBufferCapacity;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
983 data->parameters = *parameters;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
984 data->parameters.k = k;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
985 data->parameters.d = d;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
986 data->parameters.steps = kSteps;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
987 /* Check the parameters */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
988 if (!COVER_checkParameters(data->parameters)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
989 DISPLAYLEVEL(1, "Cover parameters incorrect\n");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
990 continue;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
991 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
992 /* Call the function and pass ownership of data to it */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
993 COVER_best_start(&best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
994 if (pool) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
995 POOL_add(pool, &COVER_tryParameters, data);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
996 } else {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
997 COVER_tryParameters(data);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
998 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
999 /* Print status */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1000 LOCALDISPLAYUPDATE(displayLevel, 2, "\r%u%% ",
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1001 (U32)((iteration * 100) / kIterations));
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1002 ++iteration;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1003 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1004 COVER_best_wait(&best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1005 COVER_ctx_destroy(&ctx);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1006 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1007 LOCALDISPLAYLEVEL(displayLevel, 2, "\r%79s\r", "");
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1008 /* Fill the output buffer and parameters with output of the best parameters */
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1009 {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1010 const size_t dictSize = best.dictSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1011 if (ZSTD_isError(best.compressedSize)) {
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1012 COVER_best_destroy(&best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1013 return best.compressedSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1014 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1015 *parameters = best.parameters;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1016 memcpy(dictBuffer, best.dict, dictSize);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1017 COVER_best_destroy(&best);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1018 POOL_free(pool);
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1019 return dictSize;
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1020 }
c32454d69b85 zstd: vendor python-zstandard 0.7.0
Gregory Szorc <gregory.szorc@gmail.com>
parents:
diff changeset
1021 }