zdict.h

Go to the documentation of this file.

/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * All rights reserved.
 *
 * This source code is licensed under both the BSD-style license (found in the
 * LICENSE file in the root directory of this source tree) and the GPLv2 (found
 * in the COPYING file in the root directory of this source tree).
 * You may select, at your option, one of the above-listed licenses.
 */
 
#if defined (__cplusplus)
extern "C" {
#endif
 
#ifndef ZSTD_ZDICT_H
#define ZSTD_ZDICT_H
 
/*======  Dependencies  ======*/
#include <stddef.h>  /* size_t */
 
 
/* =====   ZDICTLIB_API : control library symbols visibility   ===== */
#ifndef ZDICTLIB_VISIBLE
   /* Backwards compatibility with old macro name */
#  ifdef ZDICTLIB_VISIBILITY
#    define ZDICTLIB_VISIBLE ZDICTLIB_VISIBILITY
#  elif defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
#    define ZDICTLIB_VISIBLE __attribute__ ((visibility ("default")))
#  else
#    define ZDICTLIB_VISIBLE
#  endif
#endif
 
#ifndef ZDICTLIB_HIDDEN
#  if defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
#    define ZDICTLIB_HIDDEN __attribute__ ((visibility ("hidden")))
#  else
#    define ZDICTLIB_HIDDEN
#  endif
#endif
 
#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
#  define ZDICTLIB_API __declspec(dllexport) ZDICTLIB_VISIBLE
#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
#  define ZDICTLIB_API __declspec(dllimport) ZDICTLIB_VISIBLE /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
#else
#  define ZDICTLIB_API ZDICTLIB_VISIBLE
#endif
 
/*******************************************************************************
 * Zstd dictionary builder
 *
 * FAQ
 * ===
 * Why should I use a dictionary?
 * ------------------------------
 *
 * Zstd can use dictionaries to improve compression ratio of small data.
 * Traditionally small files don't compress well because there is very little
 * repetition in a single sample, since it is small. But, if you are compressing
 * many similar files, like a bunch of JSON records that share the same
 * structure, you can train a dictionary on ahead of time on some samples of
 * these files. Then, zstd can use the dictionary to find repetitions that are
 * present across samples. This can vastly improve compression ratio.
 *
 * When is a dictionary useful?
 * ----------------------------
 *
 * Dictionaries are useful when compressing many small files that are similar.
 * The larger a file is, the less benefit a dictionary will have. Generally,
 * we don't expect dictionary compression to be effective past 100KB. And the
 * smaller a file is, the more we would expect the dictionary to help.
 *
 * How do I use a dictionary?
 * --------------------------
 *
 * Simply pass the dictionary to the zstd compressor with
 * `ZSTD_CCtx_loadDictionary()`. The same dictionary must then be passed to
 * the decompressor, using `ZSTD_DCtx_loadDictionary()`. There are other
 * more advanced functions that allow selecting some options, see zstd.h for
 * complete documentation.
 *
 * What is a zstd dictionary?
 * --------------------------
 *
 * A zstd dictionary has two pieces: Its header, and its content. The header
 * contains a magic number, the dictionary ID, and entropy tables. These
 * entropy tables allow zstd to save on header costs in the compressed file,
 * which really matters for small data. The content is just bytes, which are
 * repeated content that is common across many samples.
 *
 * What is a raw content dictionary?
 * ---------------------------------
 *
 * A raw content dictionary is just bytes. It doesn't have a zstd dictionary
 * header, a dictionary ID, or entropy tables. Any buffer is a valid raw
 * content dictionary.
 *
 * How do I train a dictionary?
 * ----------------------------
 *
 * Gather samples from your use case. These samples should be similar to each
 * other. If you have several use cases, you could try to train one dictionary
 * per use case.
 *
 * Pass those samples to `ZDICT_trainFromBuffer()` and that will train your
 * dictionary. There are a few advanced versions of this function, but this
 * is a great starting point. If you want to further tune your dictionary
 * you could try `ZDICT_optimizeTrainFromBuffer_cover()`. If that is too slow
 * you can try `ZDICT_optimizeTrainFromBuffer_fastCover()`.
 *
 * If the dictionary training function fails, that is likely because you
 * either passed too few samples, or a dictionary would not be effective
 * for your data. Look at the messages that the dictionary trainer printed,
 * if it doesn't say too few samples, then a dictionary would not be effective.
 *
 * How large should my dictionary be?
 * ----------------------------------
 *
 * A reasonable dictionary size, the `dictBufferCapacity`, is about 100KB.
 * The zstd CLI defaults to a 110KB dictionary. You likely don't need a
 * dictionary larger than that. But, most use cases can get away with a
 * smaller dictionary. The advanced dictionary builders can automatically
 * shrink the dictionary for you, and select the smallest size that doesn't
 * hurt compression ratio too much. See the `shrinkDict` parameter.
 * A smaller dictionary can save memory, and potentially speed up
 * compression.
 *
 * How many samples should I provide to the dictionary builder?
 * ------------------------------------------------------------
 *
 * We generally recommend passing ~100x the size of the dictionary
 * in samples. A few thousand should suffice. Having too few samples
 * can hurt the dictionaries effectiveness. Having more samples will
 * only improve the dictionaries effectiveness. But having too many
 * samples can slow down the dictionary builder.
 *
 * How do I determine if a dictionary will be effective?
 * -----------------------------------------------------
 *
 * Simply train a dictionary and try it out. You can use zstd's built in
 * benchmarking tool to test the dictionary effectiveness.
 *
 *   # Benchmark levels 1-3 without a dictionary
 *   zstd -b1e3 -r /path/to/my/files
 *   # Benchmark levels 1-3 with a dictionary
 *   zstd -b1e3 -r /path/to/my/files -D /path/to/my/dictionary
 *
 * When should I retrain a dictionary?
 * -----------------------------------
 *
 * You should retrain a dictionary when its effectiveness drops. Dictionary
 * effectiveness drops as the data you are compressing changes. Generally, we do
 * expect dictionaries to "decay" over time, as your data changes, but the rate
 * at which they decay depends on your use case. Internally, we regularly
 * retrain dictionaries, and if the new dictionary performs significantly
 * better than the old dictionary, we will ship the new dictionary.
 *
 * I have a raw content dictionary, how do I turn it into a zstd dictionary?
 * -------------------------------------------------------------------------
 *
 * If you have a raw content dictionary, e.g. by manually constructing it, or
 * using a third-party dictionary builder, you can turn it into a zstd
 * dictionary by using `ZDICT_finalizeDictionary()`. You'll also have to
 * provide some samples of the data. It will add the zstd header to the
 * raw content, which contains a dictionary ID and entropy tables, which
 * will improve compression ratio, and allow zstd to write the dictionary ID
 * into the frame, if you so choose.
 *
 * Do I have to use zstd's dictionary builder?
 * -------------------------------------------
 *
 * No! You can construct dictionary content however you please, it is just
 * bytes. It will always be valid as a raw content dictionary. If you want
 * a zstd dictionary, which can improve compression ratio, use
 * `ZDICT_finalizeDictionary()`.
 *
 * What is the attack surface of a zstd dictionary?
 * ------------------------------------------------
 *
 * Zstd is heavily fuzz tested, including loading fuzzed dictionaries, so
 * zstd should never crash, or access out-of-bounds memory no matter what
 * the dictionary is. However, if an attacker can control the dictionary
 * during decompression, they can cause zstd to generate arbitrary bytes,
 * just like if they controlled the compressed data.
 *
 ******************************************************************************/
 
 
ZDICTLIB_API size_t ZDICT_trainFromBuffer(void* dictBuffer, size_t dictBufferCapacity,
                                    const void* samplesBuffer,
                                    const size_t* samplesSizes, unsigned nbSamples);
 
typedef struct {
    int      compressionLevel;   
    unsigned notificationLevel;  
    unsigned dictID;             
} ZDICT_params_t;
 
ZDICTLIB_API size_t ZDICT_finalizeDictionary(void* dstDictBuffer, size_t maxDictSize,
                                const void* dictContent, size_t dictContentSize,
                                const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
                                ZDICT_params_t parameters);
 
 
/*======   Helper functions   ======*/
ZDICTLIB_API unsigned ZDICT_getDictID(const void* dictBuffer, size_t dictSize);  
ZDICTLIB_API size_t ZDICT_getDictHeaderSize(const void* dictBuffer, size_t dictSize);  /* returns dict header size; returns a ZSTD error code on failure */
ZDICTLIB_API unsigned ZDICT_isError(size_t errorCode);
ZDICTLIB_API const char* ZDICT_getErrorName(size_t errorCode);
 
#endif   /* ZSTD_ZDICT_H */
 
#if defined(ZDICT_STATIC_LINKING_ONLY) && !defined(ZSTD_ZDICT_H_STATIC)
#define ZSTD_ZDICT_H_STATIC
 
/* This can be overridden externally to hide static symbols. */
#ifndef ZDICTLIB_STATIC_API
#  if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
#    define ZDICTLIB_STATIC_API __declspec(dllexport) ZDICTLIB_VISIBLE
#  elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
#    define ZDICTLIB_STATIC_API __declspec(dllimport) ZDICTLIB_VISIBLE
#  else
#    define ZDICTLIB_STATIC_API ZDICTLIB_VISIBLE
#  endif
#endif
 
/* ====================================================================================
 * The definitions in this section are considered experimental.
 * They should never be used with a dynamic library, as they may change in the future.
 * They are provided for advanced usages.
 * Use them only in association with static linking.
 * ==================================================================================== */
 
#define ZDICT_DICTSIZE_MIN    256
/* Deprecated: Remove in v1.6.0 */
#define ZDICT_CONTENTSIZE_MIN 128
 
typedef struct {
    unsigned k;                  /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */
    unsigned d;                  /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */
    unsigned steps;              /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */
    unsigned nbThreads;          /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */
    double splitPoint;           /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (1.0), 1.0 when all samples are used for both training and testing */
    unsigned shrinkDict;         /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking  */
    unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
    ZDICT_params_t zParams;
} ZDICT_cover_params_t;
 
typedef struct {
    unsigned k;                  /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */
    unsigned d;                  /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */
    unsigned f;                  /* log of size of frequency array : constraint: 0 < f <= 31 : 1 means default(20)*/
    unsigned steps;              /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */
    unsigned nbThreads;          /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */
    double splitPoint;           /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (0.75), 1.0 when all samples are used for both training and testing */
    unsigned accel;              /* Acceleration level: constraint: 0 < accel <= 10, higher means faster and less accurate, 0 means default(1) */
    unsigned shrinkDict;         /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking  */
    unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
 
    ZDICT_params_t zParams;
} ZDICT_fastCover_params_t;
 
ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_cover(
          void *dictBuffer, size_t dictBufferCapacity,
    const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,
          ZDICT_cover_params_t parameters);
 
ZDICTLIB_STATIC_API size_t ZDICT_optimizeTrainFromBuffer_cover(
          void* dictBuffer, size_t dictBufferCapacity,
    const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
          ZDICT_cover_params_t* parameters);
 
ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_fastCover(void *dictBuffer,
                    size_t dictBufferCapacity, const void *samplesBuffer,
                    const size_t *samplesSizes, unsigned nbSamples,
                    ZDICT_fastCover_params_t parameters);
 
ZDICTLIB_STATIC_API size_t ZDICT_optimizeTrainFromBuffer_fastCover(void* dictBuffer,
                    size_t dictBufferCapacity, const void* samplesBuffer,
                    const size_t* samplesSizes, unsigned nbSamples,
                    ZDICT_fastCover_params_t* parameters);
 
typedef struct {
    unsigned selectivityLevel;   /* 0 means default; larger => select more => larger dictionary */
    ZDICT_params_t zParams;
} ZDICT_legacy_params_t;
 
ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_legacy(
    void* dictBuffer, size_t dictBufferCapacity,
    const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
    ZDICT_legacy_params_t parameters);
 
 
/* Deprecation warnings */
/* It is generally possible to disable deprecation warnings from compiler,
   for example with -Wno-deprecated-declarations for gcc
   or _CRT_SECURE_NO_WARNINGS in Visual.
   Otherwise, it's also possible to manually define ZDICT_DISABLE_DEPRECATE_WARNINGS */
#ifdef ZDICT_DISABLE_DEPRECATE_WARNINGS
#  define ZDICT_DEPRECATED(message) /* disable deprecation warnings */
#else
#  define ZDICT_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
#  if defined (__cplusplus) && (__cplusplus >= 201402) /* C++14 or greater */
#    define ZDICT_DEPRECATED(message) [[deprecated(message)]]
#  elif defined(__clang__) || (ZDICT_GCC_VERSION >= 405)
#    define ZDICT_DEPRECATED(message) __attribute__((deprecated(message)))
#  elif (ZDICT_GCC_VERSION >= 301)
#    define ZDICT_DEPRECATED(message) __attribute__((deprecated))
#  elif defined(_MSC_VER)
#    define ZDICT_DEPRECATED(message) __declspec(deprecated(message))
#  else
#    pragma message("WARNING: You need to implement ZDICT_DEPRECATED for this compiler")
#    define ZDICT_DEPRECATED(message)
#  endif
#endif /* ZDICT_DISABLE_DEPRECATE_WARNINGS */
 
ZDICT_DEPRECATED("use ZDICT_finalizeDictionary() instead")
ZDICTLIB_STATIC_API
size_t ZDICT_addEntropyTablesFromBuffer(void* dictBuffer, size_t dictContentSize, size_t dictBufferCapacity,
                                  const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples);
 
 
#endif   /* ZSTD_ZDICT_H_STATIC */
 
#if defined (__cplusplus)
}
#endif