xref: /freebsd-src/sys/contrib/zstd/examples/dictionary_compression.c (revision 5ff13fbc199bdf5f0572845351c68ee5ca828e71)
12b9c00cbSConrad Meyer /*
2*5ff13fbcSAllan Jude  * Copyright (c) Yann Collet, Facebook, Inc.
32b9c00cbSConrad Meyer  * All rights reserved.
42b9c00cbSConrad Meyer  *
52b9c00cbSConrad Meyer  * This source code is licensed under both the BSD-style license (found in the
62b9c00cbSConrad Meyer  * LICENSE file in the root directory of this source tree) and the GPLv2 (found
72b9c00cbSConrad Meyer  * in the COPYING file in the root directory of this source tree).
82b9c00cbSConrad Meyer  * You may select, at your option, one of the above-listed licenses.
92b9c00cbSConrad Meyer  */
102b9c00cbSConrad Meyer #include <stdio.h>     // printf
112b9c00cbSConrad Meyer #include <stdlib.h>    // free
122b9c00cbSConrad Meyer #include <string.h>    // memset, strcat
132b9c00cbSConrad Meyer #include <zstd.h>      // presumes zstd library is installed
142b9c00cbSConrad Meyer #include "common.h"    // Helper functions, CHECK(), and CHECK_ZSTD()
152b9c00cbSConrad Meyer 
162b9c00cbSConrad Meyer /* createDict() :
172b9c00cbSConrad Meyer    `dictFileName` is supposed to have been created using `zstd --train` */
createCDict_orDie(const char * dictFileName,int cLevel)182b9c00cbSConrad Meyer static ZSTD_CDict* createCDict_orDie(const char* dictFileName, int cLevel)
192b9c00cbSConrad Meyer {
202b9c00cbSConrad Meyer     size_t dictSize;
212b9c00cbSConrad Meyer     printf("loading dictionary %s \n", dictFileName);
222b9c00cbSConrad Meyer     void* const dictBuffer = mallocAndLoadFile_orDie(dictFileName, &dictSize);
232b9c00cbSConrad Meyer     ZSTD_CDict* const cdict = ZSTD_createCDict(dictBuffer, dictSize, cLevel);
242b9c00cbSConrad Meyer     CHECK(cdict != NULL, "ZSTD_createCDict() failed!");
252b9c00cbSConrad Meyer     free(dictBuffer);
262b9c00cbSConrad Meyer     return cdict;
272b9c00cbSConrad Meyer }
282b9c00cbSConrad Meyer 
292b9c00cbSConrad Meyer 
compress(const char * fname,const char * oname,const ZSTD_CDict * cdict)302b9c00cbSConrad Meyer static void compress(const char* fname, const char* oname, const ZSTD_CDict* cdict)
312b9c00cbSConrad Meyer {
322b9c00cbSConrad Meyer     size_t fSize;
332b9c00cbSConrad Meyer     void* const fBuff = mallocAndLoadFile_orDie(fname, &fSize);
342b9c00cbSConrad Meyer     size_t const cBuffSize = ZSTD_compressBound(fSize);
352b9c00cbSConrad Meyer     void* const cBuff = malloc_orDie(cBuffSize);
362b9c00cbSConrad Meyer 
372b9c00cbSConrad Meyer     /* Compress using the dictionary.
382b9c00cbSConrad Meyer      * This function writes the dictionary id, and content size into the header.
392b9c00cbSConrad Meyer      * But, it doesn't use a checksum. You can control these options using the
402b9c00cbSConrad Meyer      * advanced API: ZSTD_CCtx_setParameter(), ZSTD_CCtx_refCDict(),
412b9c00cbSConrad Meyer      * and ZSTD_compress2().
422b9c00cbSConrad Meyer      */
432b9c00cbSConrad Meyer     ZSTD_CCtx* const cctx = ZSTD_createCCtx();
442b9c00cbSConrad Meyer     CHECK(cctx != NULL, "ZSTD_createCCtx() failed!");
452b9c00cbSConrad Meyer     size_t const cSize = ZSTD_compress_usingCDict(cctx, cBuff, cBuffSize, fBuff, fSize, cdict);
462b9c00cbSConrad Meyer     CHECK_ZSTD(cSize);
472b9c00cbSConrad Meyer 
482b9c00cbSConrad Meyer     saveFile_orDie(oname, cBuff, cSize);
492b9c00cbSConrad Meyer 
502b9c00cbSConrad Meyer     /* success */
512b9c00cbSConrad Meyer     printf("%25s : %6u -> %7u - %s \n", fname, (unsigned)fSize, (unsigned)cSize, oname);
522b9c00cbSConrad Meyer 
532b9c00cbSConrad Meyer     ZSTD_freeCCtx(cctx);   /* never fails */
542b9c00cbSConrad Meyer     free(fBuff);
552b9c00cbSConrad Meyer     free(cBuff);
562b9c00cbSConrad Meyer }
572b9c00cbSConrad Meyer 
582b9c00cbSConrad Meyer 
createOutFilename_orDie(const char * filename)592b9c00cbSConrad Meyer static char* createOutFilename_orDie(const char* filename)
602b9c00cbSConrad Meyer {
612b9c00cbSConrad Meyer     size_t const inL = strlen(filename);
622b9c00cbSConrad Meyer     size_t const outL = inL + 5;
632b9c00cbSConrad Meyer     void* outSpace = malloc_orDie(outL);
642b9c00cbSConrad Meyer     memset(outSpace, 0, outL);
652b9c00cbSConrad Meyer     strcat(outSpace, filename);
662b9c00cbSConrad Meyer     strcat(outSpace, ".zst");
672b9c00cbSConrad Meyer     return (char*)outSpace;
682b9c00cbSConrad Meyer }
692b9c00cbSConrad Meyer 
main(int argc,const char ** argv)702b9c00cbSConrad Meyer int main(int argc, const char** argv)
712b9c00cbSConrad Meyer {
722b9c00cbSConrad Meyer     const char* const exeName = argv[0];
732b9c00cbSConrad Meyer     int const cLevel = 3;
742b9c00cbSConrad Meyer 
752b9c00cbSConrad Meyer     if (argc<3) {
762b9c00cbSConrad Meyer         fprintf(stderr, "wrong arguments\n");
772b9c00cbSConrad Meyer         fprintf(stderr, "usage:\n");
782b9c00cbSConrad Meyer         fprintf(stderr, "%s [FILES] dictionary\n", exeName);
792b9c00cbSConrad Meyer         return 1;
802b9c00cbSConrad Meyer     }
812b9c00cbSConrad Meyer 
822b9c00cbSConrad Meyer     /* load dictionary only once */
832b9c00cbSConrad Meyer     const char* const dictName = argv[argc-1];
842b9c00cbSConrad Meyer     ZSTD_CDict* const dictPtr = createCDict_orDie(dictName, cLevel);
852b9c00cbSConrad Meyer 
862b9c00cbSConrad Meyer     int u;
872b9c00cbSConrad Meyer     for (u=1; u<argc-1; u++) {
882b9c00cbSConrad Meyer         const char* inFilename = argv[u];
892b9c00cbSConrad Meyer         char* const outFilename = createOutFilename_orDie(inFilename);
902b9c00cbSConrad Meyer         compress(inFilename, outFilename, dictPtr);
912b9c00cbSConrad Meyer         free(outFilename);
922b9c00cbSConrad Meyer     }
932b9c00cbSConrad Meyer 
942b9c00cbSConrad Meyer     ZSTD_freeCDict(dictPtr);
952b9c00cbSConrad Meyer     printf("All %u files compressed. \n", argc-2);
962b9c00cbSConrad Meyer     return 0;
972b9c00cbSConrad Meyer }
98