diff options
Diffstat (limited to 'contrib/experimental_dict_builders/fastCover/main.c')
| -rw-r--r-- | contrib/experimental_dict_builders/fastCover/main.c | 183 |
1 files changed, 183 insertions, 0 deletions
diff --git a/contrib/experimental_dict_builders/fastCover/main.c b/contrib/experimental_dict_builders/fastCover/main.c new file mode 100644 index 000000000000..df7d91812e29 --- /dev/null +++ b/contrib/experimental_dict_builders/fastCover/main.c @@ -0,0 +1,183 @@ +#include <stdio.h> /* fprintf */ +#include <stdlib.h> /* malloc, free, qsort */ +#include <string.h> /* strcmp, strlen */ +#include <errno.h> /* errno */ +#include <ctype.h> +#include "fastCover.h" +#include "io.h" +#include "util.h" +#include "zdict.h" + + +/*-************************************* +* Console display +***************************************/ +#define DISPLAY(...) fprintf(stderr, __VA_ARGS__) +#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); } + +static const U64 g_refreshRate = SEC_TO_MICRO / 6; +static UTIL_time_t g_displayClock = UTIL_TIME_INITIALIZER; + +#define DISPLAYUPDATE(l, ...) { if (displayLevel>=l) { \ + if ((UTIL_clockSpanMicro(g_displayClock) > g_refreshRate) || (displayLevel>=4)) \ + { g_displayClock = UTIL_getTime(); DISPLAY(__VA_ARGS__); \ + if (displayLevel>=4) fflush(stderr); } } } + + +/*-************************************* +* Exceptions +***************************************/ +#ifndef DEBUG +# define DEBUG 0 +#endif +#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__); +#define EXM_THROW(error, ...) \ +{ \ + DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \ + DISPLAY("Error %i : ", error); \ + DISPLAY(__VA_ARGS__); \ + DISPLAY("\n"); \ + exit(error); \ +} + + +/*-************************************* +* Constants +***************************************/ +static const unsigned g_defaultMaxDictSize = 110 KB; +#define DEFAULT_CLEVEL 3 + + +/*-************************************* +* FASTCOVER +***************************************/ +int FASTCOVER_trainFromFiles(const char* dictFileName, sampleInfo *info, + unsigned maxDictSize, + ZDICT_fastCover_params_t *params) { + unsigned const displayLevel = params->zParams.notificationLevel; + void* const dictBuffer = malloc(maxDictSize); + + int result = 0; + + /* Checks */ + if (!dictBuffer) + EXM_THROW(12, "not enough memory for trainFromFiles"); /* should not happen */ + + { size_t dictSize; + /* Run the optimize version if either k or d is not provided */ + if (!params->d || !params->k) { + dictSize = ZDICT_optimizeTrainFromBuffer_fastCover(dictBuffer, maxDictSize, info->srcBuffer, + info->samplesSizes, info->nbSamples, params); + } else { + dictSize = ZDICT_trainFromBuffer_fastCover(dictBuffer, maxDictSize, info->srcBuffer, + info->samplesSizes, info->nbSamples, *params); + } + DISPLAYLEVEL(2, "k=%u\nd=%u\nf=%u\nsteps=%u\nsplit=%u\n", params->k, params->d, params->f, params->steps, (unsigned)(params->splitPoint*100)); + if (ZDICT_isError(dictSize)) { + DISPLAYLEVEL(1, "dictionary training failed : %s \n", ZDICT_getErrorName(dictSize)); /* should not happen */ + result = 1; + goto _done; + } + /* save dict */ + DISPLAYLEVEL(2, "Save dictionary of size %u into file %s \n", (U32)dictSize, dictFileName); + saveDict(dictFileName, dictBuffer, dictSize); + } + + /* clean up */ +_done: + free(dictBuffer); + return result; +} + + + +int main(int argCount, const char* argv[]) +{ + int displayLevel = 2; + const char* programName = argv[0]; + int operationResult = 0; + + /* Initialize arguments to default values */ + unsigned k = 0; + unsigned d = 0; + unsigned f = 23; + unsigned steps = 32; + unsigned nbThreads = 1; + unsigned split = 100; + const char* outputFile = "fastCoverDict"; + unsigned dictID = 0; + unsigned maxDictSize = g_defaultMaxDictSize; + + /* Initialize table to store input files */ + const char** filenameTable = (const char**)malloc(argCount * sizeof(const char*)); + unsigned filenameIdx = 0; + + char* fileNamesBuf = NULL; + unsigned fileNamesNb = filenameIdx; + int followLinks = 0; /* follow directory recursively */ + const char** extendedFileList = NULL; + + /* Parse arguments */ + for (int i = 1; i < argCount; i++) { + const char* argument = argv[i]; + if (longCommandWArg(&argument, "k=")) { k = readU32FromChar(&argument); continue; } + if (longCommandWArg(&argument, "d=")) { d = readU32FromChar(&argument); continue; } + if (longCommandWArg(&argument, "f=")) { f = readU32FromChar(&argument); continue; } + if (longCommandWArg(&argument, "steps=")) { steps = readU32FromChar(&argument); continue; } + if (longCommandWArg(&argument, "split=")) { split = readU32FromChar(&argument); continue; } + if (longCommandWArg(&argument, "dictID=")) { dictID = readU32FromChar(&argument); continue; } + if (longCommandWArg(&argument, "maxdict=")) { maxDictSize = readU32FromChar(&argument); continue; } + if (longCommandWArg(&argument, "in=")) { + filenameTable[filenameIdx] = argument; + filenameIdx++; + continue; + } + if (longCommandWArg(&argument, "out=")) { + outputFile = argument; + continue; + } + DISPLAYLEVEL(1, "Incorrect parameters\n"); + operationResult = 1; + return operationResult; + } + + /* Get the list of all files recursively (because followLinks==0)*/ + extendedFileList = UTIL_createFileList(filenameTable, filenameIdx, &fileNamesBuf, + &fileNamesNb, followLinks); + if (extendedFileList) { + unsigned u; + for (u=0; u<fileNamesNb; u++) DISPLAYLEVEL(4, "%u %s\n", u, extendedFileList[u]); + free((void*)filenameTable); + filenameTable = extendedFileList; + filenameIdx = fileNamesNb; + } + + size_t blockSize = 0; + + /* Set up zParams */ + ZDICT_params_t zParams; + zParams.compressionLevel = DEFAULT_CLEVEL; + zParams.notificationLevel = displayLevel; + zParams.dictID = dictID; + + /* Set up fastCover params */ + ZDICT_fastCover_params_t params; + params.zParams = zParams; + params.k = k; + params.d = d; + params.f = f; + params.steps = steps; + params.nbThreads = nbThreads; + params.splitPoint = (double)split/100; + + /* Build dictionary */ + sampleInfo* info = getSampleInfo(filenameTable, + filenameIdx, blockSize, maxDictSize, zParams.notificationLevel); + operationResult = FASTCOVER_trainFromFiles(outputFile, info, maxDictSize, ¶ms); + + /* Free allocated memory */ + UTIL_freeFileList(extendedFileList, fileNamesBuf); + freeSampleInfo(info); + + return operationResult; +} |