brotli/research/dictionary_generator.cc
Eugene Kliuchnikov 8376f72ed6
Prepare for copybara (#939)
Co-authored-by: Eugene Kliuchnikov <eustas@chromium.org>
2021-11-10 10:34:39 +01:00

353 lines
12 KiB
C++

#include <climits>
#include <cstddef>
#include <cstdio>
#include <cstring>
#include <fstream>
#if !defined(_MSC_VER)
#include <glob.h>
#endif
#include <vector>
#include "deorummolae.h"
#include "durchschlag.h"
#include "sieve.h"
/* This isn't a definitive list of "--foo" arguments, only those that take an
* additional "=#" integer parameter, like "--foo=20" or "--foo=32K".
*/
#define LONG_ARG_BLOCK_LEN "--block_len="
#define LONG_ARG_SLICE_LEN "--slice_len="
#define LONG_ARG_TARGET_DICT_LEN "--target_dict_len="
#define LONG_ARG_MIN_SLICE_POP "--min_slice_pop="
#define LONG_ARG_CHUNK_LEN "--chunk_len="
#define LONG_ARG_OVERLAP_LEN "--overlap_len="
#define METHOD_DM 0
#define METHOD_SIEVE 1
#define METHOD_DURCHSCHLAG 2
#define METHOD_DISTILL 3
#define METHOD_PURIFY 4
static size_t readInt(const char* str) {
size_t result = 0;
if (str[0] == 0 || str[0] == '0') {
return 0;
}
for (size_t i = 0; i < 13; ++i) {
if (str[i] == 0) {
return result;
}
if (str[i] == 'k' || str[i] == 'K') {
if ((str[i + 1] == 0) && ((result << 10) > result)) {
return result << 10;
}
return 0;
}
if (str[i] == 'm' || str[i] == 'M') {
if ((str[i + 1] == 0) && ((result << 20) > result)) {
return result << 20;
}
return 0;
}
if (str[i] < '0' || str[i] > '9') {
return 0;
}
size_t next = (10 * result) + (str[i] - '0');
if (next <= result) {
return 0;
}
result = next;
}
return 0;
}
static std::string readFile(const std::string& path) {
std::ifstream file(path);
std::string content(
(std::istreambuf_iterator<char>(file)), std::istreambuf_iterator<char>());
return content;
}
static void writeFile(const char* file, const std::string& content) {
std::ofstream outfile(file, std::ofstream::binary);
outfile.write(content.c_str(), static_cast<std::streamsize>(content.size()));
outfile.close();
}
static void writeSamples(const std::vector<std::string>& paths,
const std::vector<size_t>& sizes, const uint8_t* data) {
size_t offset = 0;
for (size_t i = 0; i < paths.size(); ++i) {
const char* path = paths[i].c_str();
size_t sampleSize = sizes[i];
std::ofstream outfile(path, std::ofstream::binary);
outfile.write(reinterpret_cast<const char*>(data + offset),
static_cast<std::streamsize>(sampleSize));
outfile.close();
offset += sampleSize;
}
}
/* Returns "base file name" or its tail, if it contains '/' or '\'. */
static const char* fileName(const char* path) {
const char* separator_position = strrchr(path, '/');
if (separator_position) path = separator_position + 1;
separator_position = strrchr(path, '\\');
if (separator_position) path = separator_position + 1;
return path;
}
static void printHelp(const char* name) {
fprintf(stderr, "Usage: %s [OPTION]... DICTIONARY [SAMPLE]...\n", name);
fprintf(stderr,
"Options:\n"
" --dm use 'deorummolae' engine\n"
" --distill rewrite samples; unique text parts are removed\n"
" --dsh use 'durchschlag' engine (default)\n"
" --purify rewrite samples; unique text parts are zeroed out\n"
" --sieve use 'sieve' engine\n"
" -b#, --block_len=#\n"
" set block length for 'durchschlag'; default: 1024\n"
" -s#, --slice_len=#\n"
" set slice length for 'distill', 'durchschlag', 'purify'\n"
" and 'sieve'; default: 16\n"
" -t#, --target_dict_len=#\n"
" set target dictionary length (limit); default: 16K\n"
" -u#, --min_slice_pop=#\n"
" set minimum slice population (for rewrites); default: 2\n"
" -c#, --chunk_len=#\n"
" if positive, samples are cut into chunks of this length;\n"
" default: 0; cannot mix with 'rewrite samples'\n"
" -o#, --overlap_len=#\n"
" set chunk overlap length; default 0\n"
"# is a decimal number with optional k/K/m/M suffix.\n"
"WARNING: 'distill' and 'purify' will overwrite original samples!\n"
" Completely unique samples might become empty files.\n\n");
}
int main(int argc, char const* argv[]) {
int dictionaryArg = -1;
int method = METHOD_DURCHSCHLAG;
size_t sliceLen = 16;
size_t targetSize = 16 << 10;
size_t blockSize = 1024;
size_t minimumPopulation = 2;
size_t chunkLen = 0;
size_t overlapLen = 0;
std::vector<uint8_t> data;
std::vector<size_t> sizes;
std::vector<std::string> paths;
size_t total = 0;
for (int i = 1; i < argc; ++i) {
if (argv[i] == nullptr) {
continue;
}
if (argv[i][0] == '-') {
char arg1 = argv[i][1];
const char* arg2 = arg1 ? &argv[i][2] : nullptr;
if (arg1 == '-') {
if (dictionaryArg != -1) {
fprintf(stderr,
"Method should be specified before dictionary / sample '%s'\n",
argv[i]);
exit(1);
}
/* Look for "--long_arg" via exact match. */
if (std::strcmp(argv[i], "--sieve") == 0) {
method = METHOD_SIEVE;
continue;
}
if (std::strcmp(argv[i], "--dm") == 0) {
method = METHOD_DM;
continue;
}
if (std::strcmp(argv[i], "--dsh") == 0) {
method = METHOD_DURCHSCHLAG;
continue;
}
if (std::strcmp(argv[i], "--distill") == 0) {
method = METHOD_DISTILL;
continue;
}
if (std::strcmp(argv[i], "--purify") == 0) {
method = METHOD_PURIFY;
continue;
}
/* Look for "--long_arg=#" via prefix match. */
if (std::strncmp(argv[i], LONG_ARG_BLOCK_LEN,
std::strlen(LONG_ARG_BLOCK_LEN)) == 0) {
arg1 = 'b';
arg2 = &argv[i][std::strlen(LONG_ARG_BLOCK_LEN)];
} else if (std::strncmp(argv[i], LONG_ARG_SLICE_LEN,
std::strlen(LONG_ARG_SLICE_LEN)) == 0) {
arg1 = 's';
arg2 = &argv[i][std::strlen(LONG_ARG_SLICE_LEN)];
} else if (std::strncmp(argv[i], LONG_ARG_TARGET_DICT_LEN,
std::strlen(LONG_ARG_TARGET_DICT_LEN)) == 0) {
arg1 = 't';
arg2 = &argv[i][std::strlen(LONG_ARG_TARGET_DICT_LEN)];
} else if (std::strncmp(argv[i], LONG_ARG_MIN_SLICE_POP,
std::strlen(LONG_ARG_MIN_SLICE_POP)) == 0) {
arg1 = 'u';
arg2 = &argv[i][std::strlen(LONG_ARG_MIN_SLICE_POP)];
} else if (std::strncmp(argv[i], LONG_ARG_CHUNK_LEN,
std::strlen(LONG_ARG_CHUNK_LEN)) == 0) {
arg1 = 'c';
arg2 = &argv[i][std::strlen(LONG_ARG_CHUNK_LEN)];
} else if (std::strncmp(argv[i], LONG_ARG_OVERLAP_LEN,
std::strlen(LONG_ARG_OVERLAP_LEN)) == 0) {
arg1 = 'o';
arg2 = &argv[i][std::strlen(LONG_ARG_OVERLAP_LEN)];
} else {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid option '%s'\n", argv[i]);
exit(1);
}
}
/* Look for "-f" short args or "--foo=#" long args. */
if (arg1 == 'b') {
blockSize = readInt(arg2);
if (blockSize < 16 || blockSize > 65536) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid option '%s'\n", argv[i]);
exit(1);
}
} else if (arg1 == 's') {
sliceLen = readInt(arg2);
// TODO(eustas): investigate why sliceLen == 4..5 greatly slows down
// durschlag engine, but only from command line;
// durschlag_runner seems to work fine with those.
if (sliceLen < 4 || sliceLen > 256) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid option '%s'\n", argv[i]);
exit(1);
}
} else if (arg1 == 't') {
targetSize = readInt(arg2);
if (targetSize < 256 || targetSize > (1 << 25)) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid option '%s'\n", argv[i]);
exit(1);
}
} else if (arg1 == 'u') {
minimumPopulation = readInt(arg2);
if (minimumPopulation < 256 || minimumPopulation > 65536) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid option '%s'\n", argv[i]);
exit(1);
}
} else if (arg1 == 'c') {
chunkLen = readInt(arg2);
if (chunkLen < 0 || chunkLen > INT_MAX) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid option '%s'\n", argv[i]);
exit(1);
}
} else if (arg1 == 'o') {
overlapLen = readInt(arg2);
if (overlapLen < 0 || overlapLen > INT_MAX) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid option '%s'\n", argv[i]);
exit(1);
}
} else {
printHelp(fileName(argv[0]));
fprintf(stderr, "Unrecognized option '%s'\n", argv[i]);
exit(1);
}
continue;
}
if (dictionaryArg == -1) {
if (method != METHOD_DISTILL && method != METHOD_PURIFY) {
dictionaryArg = i;
continue;
}
}
bool ok = true;
#if defined(_MSC_VER)
const char* resolved_path = argv[i];
#else
glob_t resolved_paths;
memset(&resolved_paths, 0, sizeof(resolved_paths));
if (glob(argv[i], GLOB_TILDE, NULL, &resolved_paths) == 0) {
for(size_t j = 0; j < resolved_paths.gl_pathc; ++j) {
const char* resolved_path = resolved_paths.gl_pathv[j];
#endif
std::string content = readFile(resolved_path);
if (chunkLen == 0) {
paths.emplace_back(resolved_path);
data.insert(data.end(), content.begin(), content.end());
total += content.size();
sizes.push_back(content.size());
continue;
} else if (chunkLen <= overlapLen) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Invalid chunkLen - overlapLen combination\n");
exit(1);
}
for (size_t chunkStart = 0;
chunkStart < content.size();
chunkStart += chunkLen - overlapLen) {
std::string chunk = content.substr(chunkStart, chunkLen);
data.insert(data.end(), chunk.begin(), chunk.end());
total += chunk.size();
sizes.push_back(chunk.size());
}
#if !defined(_MSC_VER)
}
} else {
ok = false;
}
globfree(&resolved_paths);
#endif
if (!ok) exit(1);
}
fprintf(stderr, "Number of chunks: %zu; total size: %zu\n", sizes.size(),
total);
bool wantDictionary = (dictionaryArg == -1);
if (method == METHOD_DISTILL || method == METHOD_PURIFY) {
wantDictionary = false;
if (chunkLen != 0) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Cannot mix 'rewrite samples' with positive chunk_len\n");
exit(1);
}
}
if (wantDictionary || total == 0) {
printHelp(fileName(argv[0]));
fprintf(stderr, "Not enough arguments\n");
exit(1);
}
if (method == METHOD_SIEVE) {
writeFile(argv[dictionaryArg], sieve_generate(
targetSize, sliceLen, sizes, data.data()));
} else if (method == METHOD_DM) {
writeFile(argv[dictionaryArg], DM_generate(
targetSize, sizes, data.data()));
} else if (method == METHOD_DURCHSCHLAG) {
writeFile(argv[dictionaryArg], durchschlag_generate(
targetSize, sliceLen, blockSize, sizes, data.data()));
} else if (method == METHOD_DISTILL) {
durchschlag_distill(sliceLen, minimumPopulation, &sizes, data.data());
writeSamples(paths, sizes, data.data());
} else if (method == METHOD_PURIFY) {
durchschlag_purify(sliceLen, minimumPopulation, sizes, data.data());
writeSamples(paths, sizes, data.data());
} else {
printHelp(fileName(argv[0]));
fprintf(stderr, "Unknown generator\n");
exit(1);
}
return 0;
}