Yann Collet | 71eafdd | 2016-02-12 02:31:57 +0100 | [diff] [blame] | 1 | /* |
| 2 | dibio - I/O API for dictionary builder |
| 3 | Copyright (C) Yann Collet 2016 |
| 4 | |
| 5 | GPL v2 License |
| 6 | |
| 7 | This program is free software; you can redistribute it and/or modify |
| 8 | it under the terms of the GNU General Public License as published by |
| 9 | the Free Software Foundation; either version 2 of the License, or |
| 10 | (at your option) any later version. |
| 11 | |
| 12 | This program is distributed in the hope that it will be useful, |
| 13 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 15 | GNU General Public License for more details. |
| 16 | |
| 17 | You should have received a copy of the GNU General Public License along |
| 18 | with this program; if not, write to the Free Software Foundation, Inc., |
| 19 | 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. |
| 20 | |
| 21 | You can contact the author at : |
| 22 | - zstd homepage : http://www.zstd.net/ |
| 23 | */ |
| 24 | |
| 25 | /*-************************************** |
| 26 | * Compiler Options |
| 27 | ****************************************/ |
| 28 | /* Disable some Visual warning messages */ |
| 29 | #ifdef _MSC_VER |
| 30 | # define _CRT_SECURE_NO_WARNINGS /* fopen */ |
| 31 | # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ |
| 32 | #endif |
| 33 | |
| 34 | /* Unix Large Files support (>4GB) */ |
| 35 | #define _FILE_OFFSET_BITS 64 |
| 36 | #if (defined(__sun__) && (!defined(__LP64__))) /* Sun Solaris 32-bits requires specific definitions */ |
| 37 | # define _LARGEFILE_SOURCE |
| 38 | #elif ! defined(__LP64__) /* No point defining Large file for 64 bit */ |
| 39 | # define _LARGEFILE64_SOURCE |
| 40 | #endif |
| 41 | |
| 42 | |
| 43 | /*-************************************* |
| 44 | * Includes |
| 45 | ***************************************/ |
| 46 | #include <stdlib.h> /* malloc, free */ |
| 47 | #include <string.h> /* memset */ |
| 48 | #include <stdio.h> /* fprintf, fopen, ftello64 */ |
| 49 | #include <sys/types.h> /* stat64 */ |
| 50 | #include <sys/stat.h> /* stat64 */ |
| 51 | #include <time.h> /* clock */ |
| 52 | |
| 53 | #include "mem.h" /* read */ |
| 54 | #include "error_private.h" |
Yann Collet | f4c9d75 | 2016-02-12 18:45:02 +0100 | [diff] [blame] | 55 | #include "zdict_static.h" |
| 56 | |
Yann Collet | 71eafdd | 2016-02-12 02:31:57 +0100 | [diff] [blame] | 57 | |
| 58 | /*-************************************* |
| 59 | * Compiler specifics |
| 60 | ***************************************/ |
| 61 | #if !defined(S_ISREG) |
| 62 | # define S_ISREG(x) (((x) & S_IFMT) == S_IFREG) |
| 63 | #endif |
| 64 | |
| 65 | |
| 66 | /*-************************************* |
| 67 | * Constants |
| 68 | ***************************************/ |
| 69 | #define KB *(1 <<10) |
| 70 | #define MB *(1 <<20) |
| 71 | #define GB *(1U<<30) |
| 72 | |
| 73 | #define DICTLISTSIZE 10000 |
| 74 | #define MEMMULT 11 |
| 75 | static const size_t maxMemory = (sizeof(size_t) == 4) ? (2 GB - 64 MB) : ((size_t)(512 MB) << sizeof(size_t)); |
| 76 | |
| 77 | #define NOISELENGTH 32 |
| 78 | #define PRIME1 2654435761U |
| 79 | #define PRIME2 2246822519U |
| 80 | |
| 81 | |
| 82 | /*-************************************* |
| 83 | * Console display |
| 84 | ***************************************/ |
| 85 | #define DISPLAY(...) fprintf(stderr, __VA_ARGS__) |
| 86 | #define DISPLAYLEVEL(l, ...) if (g_displayLevel>=l) { DISPLAY(__VA_ARGS__); } |
| 87 | static unsigned g_displayLevel = 0; /* 0 : no display; 1: errors; 2: default; 4: full information */ |
Yann Collet | 71eafdd | 2016-02-12 02:31:57 +0100 | [diff] [blame] | 88 | |
Yann Collet | 71eafdd | 2016-02-12 02:31:57 +0100 | [diff] [blame] | 89 | |
| 90 | /*-************************************* |
| 91 | * Exceptions |
| 92 | ***************************************/ |
| 93 | #ifndef DEBUG |
| 94 | # define DEBUG 0 |
| 95 | #endif |
| 96 | #define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__); |
| 97 | #define EXM_THROW(error, ...) \ |
| 98 | { \ |
| 99 | DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \ |
| 100 | DISPLAYLEVEL(1, "Error %i : ", error); \ |
| 101 | DISPLAYLEVEL(1, __VA_ARGS__); \ |
| 102 | DISPLAYLEVEL(1, "\n"); \ |
| 103 | exit(error); \ |
| 104 | } |
| 105 | |
| 106 | |
| 107 | /* ******************************************************** |
| 108 | * Helper functions |
| 109 | **********************************************************/ |
| 110 | unsigned DiB_isError(size_t errorCode) { return ERR_isError(errorCode); } |
| 111 | |
| 112 | const char* DiB_getErrorName(size_t errorCode) { return ERR_getErrorName(errorCode); } |
| 113 | |
| 114 | |
| 115 | /* ******************************************************** |
| 116 | * File related operations |
| 117 | **********************************************************/ |
| 118 | static unsigned long long DiB_getFileSize(const char* infilename) |
| 119 | { |
| 120 | int r; |
| 121 | #if defined(_MSC_VER) |
| 122 | struct _stat64 statbuf; |
| 123 | r = _stat64(infilename, &statbuf); |
| 124 | #else |
| 125 | struct stat statbuf; |
| 126 | r = stat(infilename, &statbuf); |
| 127 | #endif |
| 128 | if (r || !S_ISREG(statbuf.st_mode)) return 0; /* No good... */ |
| 129 | return (unsigned long long)statbuf.st_size; |
| 130 | } |
| 131 | |
| 132 | |
| 133 | static unsigned long long DiB_getTotalFileSize(const char** fileNamesTable, unsigned nbFiles) |
| 134 | { |
| 135 | unsigned long long total = 0; |
| 136 | unsigned n; |
| 137 | for (n=0; n<nbFiles; n++) |
| 138 | total += DiB_getFileSize(fileNamesTable[n]); |
| 139 | return total; |
| 140 | } |
| 141 | |
| 142 | |
| 143 | static void DiB_loadFiles(void* buffer, size_t bufferSize, |
| 144 | size_t* fileSizes, |
| 145 | const char** fileNamesTable, unsigned nbFiles) |
| 146 | { |
| 147 | char* buff = (char*)buffer; |
| 148 | size_t pos = 0; |
| 149 | unsigned n; |
| 150 | |
| 151 | for (n=0; n<nbFiles; n++) { |
| 152 | size_t readSize; |
| 153 | unsigned long long fileSize = DiB_getFileSize(fileNamesTable[n]); |
| 154 | FILE* f = fopen(fileNamesTable[n], "rb"); |
| 155 | if (f==NULL) EXM_THROW(10, "impossible to open file %s", fileNamesTable[n]); |
| 156 | DISPLAYLEVEL(2, "Loading %s... \r", fileNamesTable[n]); |
| 157 | if (fileSize > bufferSize-pos) fileSize = 0; /* stop there, not enough memory to load all files */ |
| 158 | readSize = fread(buff+pos, 1, (size_t)fileSize, f); |
| 159 | if (readSize != (size_t)fileSize) EXM_THROW(11, "could not read %s", fileNamesTable[n]); |
| 160 | pos += readSize; |
| 161 | fileSizes[n] = (size_t)fileSize; |
| 162 | fclose(f); |
| 163 | } |
| 164 | } |
| 165 | |
| 166 | |
| 167 | /*-******************************************************** |
| 168 | * Dictionary training functions |
| 169 | **********************************************************/ |
| 170 | static size_t DiB_findMaxMem(unsigned long long requiredMem) |
| 171 | { |
| 172 | size_t step = 8 MB; |
| 173 | void* testmem = NULL; |
| 174 | |
| 175 | requiredMem = (((requiredMem >> 23) + 1) << 23); |
| 176 | requiredMem += 2 * step; |
| 177 | if (requiredMem > maxMemory) requiredMem = maxMemory; |
| 178 | |
| 179 | while (!testmem) { |
| 180 | requiredMem -= step; |
| 181 | testmem = malloc((size_t)requiredMem); |
| 182 | } |
| 183 | |
| 184 | free(testmem); |
| 185 | return (size_t)(requiredMem - step); |
| 186 | } |
| 187 | |
| 188 | |
| 189 | static void DiB_fillNoise(void* buffer, size_t length) |
| 190 | { |
| 191 | unsigned acc = PRIME1; |
| 192 | size_t p=0;; |
| 193 | |
| 194 | for (p=0; p<length; p++) { |
| 195 | acc *= PRIME2; |
| 196 | ((unsigned char*)buffer)[p] = (unsigned char)(acc >> 21); |
| 197 | } |
| 198 | } |
| 199 | |
| 200 | |
| 201 | static void DiB_saveDict(const char* dictFileName, |
| 202 | const void* buff, size_t buffSize) |
| 203 | { |
| 204 | FILE* f; |
| 205 | size_t n; |
| 206 | |
| 207 | f = fopen(dictFileName, "wb"); |
| 208 | if (f==NULL) EXM_THROW(3, "cannot open %s ", dictFileName); |
| 209 | |
| 210 | n = fwrite(buff, 1, buffSize, f); |
| 211 | if (n!=buffSize) EXM_THROW(4, "%s : write error", dictFileName) |
| 212 | |
| 213 | n = (size_t)fclose(f); |
| 214 | if (n!=0) EXM_THROW(5, "%s : flush error", dictFileName) |
| 215 | } |
| 216 | |
| 217 | |
Yann Collet | 6f3acba | 2016-02-12 20:19:48 +0100 | [diff] [blame^] | 218 | /*! ZDICT_trainFromBuffer_unsafe() : |
| 219 | Strictly Internal use only !! |
| 220 | Same as ZDICT_trainFromBuffer_advanced(), but does not control `samplesBuffer`. |
| 221 | `samplesBuffer` must be followed by noisy guard band to avoid out-of-buffer reads. |
| 222 | @return : size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`) |
| 223 | or an error code. |
| 224 | */ |
| 225 | size_t ZDICT_trainFromBuffer_unsafe(void* dictBuffer, size_t dictBufferCapacity, |
| 226 | const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples, |
| 227 | ZDICT_params_t parameters); |
| 228 | |
| 229 | |
Yann Collet | 71eafdd | 2016-02-12 02:31:57 +0100 | [diff] [blame] | 230 | int DiB_trainFromFiles(const char* dictFileName, unsigned maxDictSize, |
| 231 | const char** fileNamesTable, unsigned nbFiles, |
| 232 | ZDICT_params_t params) |
| 233 | { |
| 234 | void* srcBuffer; |
| 235 | size_t benchedSize; |
| 236 | size_t* fileSizes = (size_t*)malloc(nbFiles * sizeof(size_t)); |
| 237 | unsigned long long totalSizeToLoad = DiB_getTotalFileSize(fileNamesTable, nbFiles); |
| 238 | void* dictBuffer = malloc(maxDictSize); |
| 239 | size_t dictSize; |
| 240 | int result = 0; |
| 241 | |
| 242 | /* init */ |
Yann Collet | 6f3acba | 2016-02-12 20:19:48 +0100 | [diff] [blame^] | 243 | g_displayLevel = params.notificationLevel; |
Yann Collet | 71eafdd | 2016-02-12 02:31:57 +0100 | [diff] [blame] | 244 | benchedSize = DiB_findMaxMem(totalSizeToLoad * MEMMULT) / MEMMULT; |
| 245 | if ((unsigned long long)benchedSize > totalSizeToLoad) benchedSize = (size_t)totalSizeToLoad; |
| 246 | if (benchedSize < totalSizeToLoad) |
| 247 | DISPLAYLEVEL(1, "Not enough memory; training on %u MB only...\n", (unsigned)(benchedSize >> 20)); |
| 248 | |
| 249 | /* Memory allocation & restrictions */ |
| 250 | srcBuffer = malloc(benchedSize+NOISELENGTH); /* + noise */ |
| 251 | if ((!fileSizes) || (!srcBuffer) || (!dictBuffer)) EXM_THROW(12, "not enough memory for DiB_trainFiles"); /* should not happen */ |
| 252 | |
| 253 | /* Load input buffer */ |
| 254 | DiB_loadFiles(srcBuffer, benchedSize, fileSizes, fileNamesTable, nbFiles); |
| 255 | DiB_fillNoise((char*)srcBuffer + benchedSize, NOISELENGTH); /* guard band, for end of buffer condition */ |
| 256 | |
| 257 | /* call buffer version */ |
| 258 | dictSize = ZDICT_trainFromBuffer_unsafe(dictBuffer, maxDictSize, |
| 259 | srcBuffer, fileSizes, nbFiles, |
| 260 | params); |
| 261 | if (ZDICT_isError(dictSize)) { |
| 262 | DISPLAYLEVEL(1, "dictionary training failed : %s", ZDICT_getErrorName(dictSize)); /* should not happen */ |
| 263 | result = 1; |
| 264 | goto _cleanup; |
| 265 | } |
| 266 | |
| 267 | /* save dict */ |
| 268 | DISPLAYLEVEL(2, "Save dictionary of size %u into file %s \n", (U32)dictSize, dictFileName); |
| 269 | DiB_saveDict(dictFileName, dictBuffer, dictSize); |
| 270 | |
| 271 | /* clean up */ |
| 272 | _cleanup: |
| 273 | free(srcBuffer); |
| 274 | free(dictBuffer); |
| 275 | free(fileSizes); |
| 276 | return result; |
| 277 | } |