nicolas
/
ik_llama_opt
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Revert "Rpc improvement (#480)" 

This reverts commit 8a5f8573aefc23282200041abbfa12886083334a.
This commit is contained in:
Kawrakow 2025-06-08 14:49:50 +03:00
parent ed9e8ecc9b
commit 7fc39ae7e8
11 changed files with 508 additions and 1258 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
CMakeLists.txt
View File

@ -6,6 +6,7 @@ include(CheckIncludeFileCXX)
set(CMAKE_WARN_UNUSED_CLI YES) set(CMAKE_WARN_UNUSED_CLI YES)
set(CMAKE_EXPORT_COMPILE_COMMANDS ON) set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
set(CMAKE_CXX_STANDARD 17) set(CMAKE_CXX_STANDARD 17)
set(CMAKE_CXX_STANDARD_REQUIRED true) set(CMAKE_CXX_STANDARD_REQUIRED true)

51
common/common.cpp
View File

@ -81,9 +81,7 @@
#endif #endif
#define LLAMA_CURL_MAX_URL_LENGTH 2084 // Maximum URL Length in Chrome: 2083 #define LLAMA_CURL_MAX_URL_LENGTH 2084 // Maximum URL Length in Chrome: 2083
#endif // LLAMA_USE_CURL #endif // LLAMA_USE_CURL
#ifdef GGML_USE_RPC
# include "ggml-rpc.h"
#endif
using json = nlohmann::ordered_json; using json = nlohmann::ordered_json;
// //
@ -1006,35 +1004,6 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
if (arg == "--rpc") { if (arg == "--rpc") {
CHECK_ARG CHECK_ARG
params.rpc_servers = argv[i]; params.rpc_servers = argv[i];
std::string servers(params.rpc_servers);
size_t pos = 0;
while ((pos = servers.find(",")) != std::string::npos) {
std::string server = servers.substr(0, pos);
ggml_backend_rpc_buffer_type(server.c_str());
servers.erase(0, pos + 1);
}
ggml_backend_rpc_buffer_type(servers.c_str());
return true;
}
if (arg == "--override-kv") {
CHECK_ARG
if (!string_parse_kv_override(argv[i], params.kv_overrides)) {
fprintf(stderr, "error: Invalid type for KV override: %s\n", argv[i]);
invalid_param = true;
return true;
}
return true;
}
if (arg == "--override-tensor" || arg == "-ot") {
CHECK_ARG
/*for (auto endpoint : params.rpc_servers.split)
{
}*/
if (!parse_buft_overrides(std::string{ argv[i] }, params.tensor_buft_overrides)) {
fprintf(stderr, "error: Invalid tensor buffer type override: %s\n", argv[i]);
invalid_param = true;
}
return true; return true;
} }
if (arg == "--no-mmap") { if (arg == "--no-mmap") {
@ -1242,7 +1211,23 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
sparams.grammar = json_schema_to_grammar(json::parse(argv[i])); sparams.grammar = json_schema_to_grammar(json::parse(argv[i]));
return true; return true;
} }
if (arg == "--override-kv") {
CHECK_ARG
if (!string_parse_kv_override(argv[i], params.kv_overrides)) {
fprintf(stderr, "error: Invalid type for KV override: %s\n", argv[i]);
invalid_param = true;
return true;
}
return true;
}
if (arg == "--override-tensor" || arg == "-ot") {
CHECK_ARG
if (!parse_buft_overrides(std::string{argv[i]}, params.tensor_buft_overrides)) {
fprintf(stderr, "error: Invalid tensor buffer type override: %s\n", argv[i]);
invalid_param = true;
}
return true;
}
if (arg == "--offload-policy" || arg == "-op") { if (arg == "--offload-policy" || arg == "-op") {
CHECK_ARG CHECK_ARG
auto p = string_split_pairs<int,int>(argv[i], ','); auto p = string_split_pairs<int,int>(argv[i], ',');

6
examples/rpc/CMakeLists.txt
View File

@ -1,4 +1,2 @@
set(TARGET rpc-server) add_executable(rpc-server rpc-server.cpp)
add_executable(${TARGET} rpc-server.cpp) target_link_libraries(rpc-server PRIVATE ggml llama)
target_link_libraries(${TARGET} PRIVATE ggml)
target_compile_features(${TARGET} PRIVATE cxx_std_17)

214
examples/rpc/rpc-server.cpp
View File

@ -5,166 +5,33 @@
#ifdef GGML_USE_METAL #ifdef GGML_USE_METAL
#include "ggml-metal.h" #include "ggml-metal.h"
#endif #endif
#ifdef GGML_USE_VULKAN
#include "ggml-vulkan.h"
#endif
#ifdef GGML_USE_SYCL
#include "ggml-sycl.h"
#endif
#include "ggml-rpc.h" #include "ggml-rpc.h"
#ifdef _WIN32 #ifdef _WIN32
# define DIRECTORY_SEPARATOR '\\'
# define NOMINMAX
# include <locale>
# include <windows.h> # include <windows.h>
# include <fcntl.h>
# include <io.h>
#else #else
# define DIRECTORY_SEPARATOR '/'
# include <unistd.h> # include <unistd.h>
# include <sys/stat.h>
#endif #endif
#include <string> #include <string>
#include <stdio.h> #include <stdio.h>
#include <algorithm>
#include <thread>
#include <fstream>
#include <filesystem>
#include <codecvt>
namespace fs = std::filesystem;
// NOTE: this is copied from common.cpp to avoid linking with libcommon
// returns true if successful, false otherwise
static bool fs_create_directory_with_parents(const std::string& path) {
#ifdef _WIN32
std::wstring_convert<std::codecvt_utf8<wchar_t>> converter;
std::wstring wpath = converter.from_bytes(path);
// if the path already exists, check whether it's a directory
const DWORD attributes = GetFileAttributesW(wpath.c_str());
if ((attributes != INVALID_FILE_ATTRIBUTES) && (attributes & FILE_ATTRIBUTE_DIRECTORY)) {
return true;
}
size_t pos_slash = 0;
// process path from front to back, procedurally creating directories
while ((pos_slash = path.find('\\', pos_slash)) != std::string::npos) {
const std::wstring subpath = wpath.substr(0, pos_slash);
const wchar_t* test = subpath.c_str();
const bool success = CreateDirectoryW(test, NULL);
if (!success) {
const DWORD error = GetLastError();
// if the path already exists, ensure that it's a directory
if (error == ERROR_ALREADY_EXISTS) {
const DWORD attributes = GetFileAttributesW(subpath.c_str());
if (attributes == INVALID_FILE_ATTRIBUTES || !(attributes & FILE_ATTRIBUTE_DIRECTORY)) {
return false;
}
}
else {
return false;
}
}
pos_slash += 1;
}
return true;
#else
// if the path already exists, check whether it's a directory
struct stat info;
if (stat(path.c_str(), &info) == 0) {
return S_ISDIR(info.st_mode);
}
size_t pos_slash = 1; // skip leading slashes for directory creation
// process path from front to back, procedurally creating directories
while ((pos_slash = path.find('/', pos_slash)) != std::string::npos) {
const std::string subpath = path.substr(0, pos_slash);
struct stat info;
// if the path already exists, ensure that it's a directory
if (stat(subpath.c_str(), &info) == 0) {
if (!S_ISDIR(info.st_mode)) {
return false;
}
}
else {
// create parent directories
const int ret = mkdir(subpath.c_str(), 0755);
if (ret != 0) {
return false;
}
}
pos_slash += 1;
}
return true;
#endif // _WIN32
}
// NOTE: this is copied from common.cpp to avoid linking with libcommon
static std::string fs_get_cache_directory() {
std::string cache_directory = "";
auto ensure_trailing_slash = [](std::string p) {
// Make sure to add trailing slash
if (p.back() != DIRECTORY_SEPARATOR) {
p += DIRECTORY_SEPARATOR;
}
return p;
};
if (getenv("LLAMA_CACHE")) {
cache_directory = std::getenv("LLAMA_CACHE");
}
else {
#if defined(__linux__) || defined(__FreeBSD__) || defined(_AIX)
if (std::getenv("XDG_CACHE_HOME")) {
cache_directory = std::getenv("XDG_CACHE_HOME");
}
else {
cache_directory = std::getenv("HOME") + std::string("/.cache/");
}
#elif defined(__APPLE__)
cache_directory = std::getenv("HOME") + std::string("/Library/Caches/");
#elif defined(_WIN32)
cache_directory = std::getenv("LOCALAPPDATA");
#else
# error Unknown architecture
#endif
cache_directory = ensure_trailing_slash(cache_directory);
cache_directory += "llama.cpp";
}
return ensure_trailing_slash(cache_directory);
}
struct rpc_server_params { struct rpc_server_params {
std::string host = "127.0.0.1"; std::string host = "127.0.0.1";
int port = 50052; int port = 50052;
size_t backend_mem = 0; size_t backend_mem = 0;
bool use_cache = false;
int n_threads = std::max(1U, std::thread::hardware_concurrency() / 2);
}; };
static void print_usage(int /*argc*/, char** argv, rpc_server_params params) { static void print_usage(int /*argc*/, char ** argv, rpc_server_params params) {
fprintf(stderr, "Usage: %s [options]\n\n", argv[0]); fprintf(stderr, "Usage: %s [options]\n\n", argv[0]);
fprintf(stderr, "options:\n"); fprintf(stderr, "options:\n");
fprintf(stderr, " -h, --help show this help message and exit\n"); fprintf(stderr, " -h, --help show this help message and exit\n");
fprintf(stderr, " -t, --threads number of threads for the CPU backend (default: %d)\n", params.n_threads); fprintf(stderr, " -H HOST, --host HOST host to bind to (default: %s)\n", params.host.c_str());
fprintf(stderr, " -H HOST, --host HOST host to bind to (default: %s)\n", params.host.c_str()); fprintf(stderr, " -p PORT, --port PORT port to bind to (default: %d)\n", params.port);
fprintf(stderr, " -p PORT, --port PORT port to bind to (default: %d)\n", params.port); fprintf(stderr, " -m MEM, --mem MEM backend memory size (in MB)\n");
fprintf(stderr, " -m MEM, --mem MEM backend memory size (in MB)\n");
fprintf(stderr, " -c, --cache enable local file cache\n");
fprintf(stderr, "\n"); fprintf(stderr, "\n");
} }
static bool rpc_server_params_parse(int argc, char** argv, rpc_server_params& params) { static bool rpc_server_params_parse(int argc, char ** argv, rpc_server_params & params) {
std::string arg; std::string arg;
for (int i = 1; i < argc; i++) { for (int i = 1; i < argc; i++) {
arg = argv[i]; arg = argv[i];
@ -173,18 +40,7 @@ static bool rpc_server_params_parse(int argc, char** argv, rpc_server_params& pa
return false; return false;
} }
params.host = argv[i]; params.host = argv[i];
} } else if (arg == "-p" || arg == "--port") {
else if (arg == "-t" || arg == "--threads") {
if (++i >= argc) {
return false;
}
params.n_threads = std::stoi(argv[i]);
if (params.n_threads <= 0) {
fprintf(stderr, "error: invalid number of threads: %d\n", params.n_threads);
return false;
}
}
else if (arg == "-p" || arg == "--port") {
if (++i >= argc) { if (++i >= argc) {
return false; return false;
} }
@ -192,21 +48,15 @@ static bool rpc_server_params_parse(int argc, char** argv, rpc_server_params& pa
if (params.port <= 0 || params.port > 65535) { if (params.port <= 0 || params.port > 65535) {
return false; return false;
} }
} } else if (arg == "-m" || arg == "--mem") {
else if (arg == "-c" || arg == "--cache") {
params.use_cache = true;
}
else if (arg == "-m" || arg == "--mem") {
if (++i >= argc) { if (++i >= argc) {
return false; return false;
} }
params.backend_mem = std::stoul(argv[i]) * 1024 * 1024; params.backend_mem = std::stoul(argv[i]) * 1024 * 1024;
} } else if (arg == "-h" || arg == "--help") {
else if (arg == "-h" || arg == "--help") {
print_usage(argc, argv, params); print_usage(argc, argv, params);
exit(0); exit(0);
} } else {
else {
fprintf(stderr, "error: unknown argument: %s\n", arg.c_str()); fprintf(stderr, "error: unknown argument: %s\n", arg.c_str());
print_usage(argc, argv, params); print_usage(argc, argv, params);
exit(0); exit(0);
@ -215,7 +65,7 @@ static bool rpc_server_params_parse(int argc, char** argv, rpc_server_params& pa
return true; return true;
} }
static ggml_backend_t create_backend(const rpc_server_params& params) { static ggml_backend_t create_backend() {
ggml_backend_t backend = NULL; ggml_backend_t backend = NULL;
#ifdef GGML_USE_CUDA #ifdef GGML_USE_CUDA
fprintf(stderr, "%s: using CUDA backend\n", __func__); fprintf(stderr, "%s: using CUDA backend\n", __func__);
@ -229,25 +79,12 @@ static ggml_backend_t create_backend(const rpc_server_params& params) {
if (!backend) { if (!backend) {
fprintf(stderr, "%s: ggml_backend_metal_init() failed\n", __func__); fprintf(stderr, "%s: ggml_backend_metal_init() failed\n", __func__);
} }
#elif GGML_USE_VULKAN
fprintf(stderr, "%s: using Vulkan backend\n", __func__);
backend = ggml_backend_vk_init(0); // init device 0
if (!backend) {
fprintf(stderr, "%s: ggml_backend_vulkan_init() failed\n", __func__);
}
#elif GGML_USE_SYCL
fprintf(stderr, "%s: using SYCL backend\n", __func__);
backend = ggml_backend_sycl_init(0); // init device 0
if (!backend) {
fprintf(stderr, "%s: ggml_backend_sycl_init() failed\n", __func__);
}
#endif #endif
// if there aren't GPU Backends fallback to CPU backend // if there aren't GPU Backends fallback to CPU backend
if (!backend) { if (!backend) {
fprintf(stderr, "%s: using CPU backend\n", __func__); fprintf(stderr, "%s: using CPU backend\n", __func__);
backend = ggml_backend_cpu_init(); backend = ggml_backend_cpu_init();
ggml_backend_cpu_set_n_threads(backend, params.n_threads);
} }
return backend; return backend;
} }
@ -255,10 +92,6 @@ static ggml_backend_t create_backend(const rpc_server_params& params) {
static void get_backend_memory(size_t * free_mem, size_t * total_mem) { static void get_backend_memory(size_t * free_mem, size_t * total_mem) {
#ifdef GGML_USE_CUDA #ifdef GGML_USE_CUDA
ggml_backend_cuda_get_device_memory(0, free_mem, total_mem); ggml_backend_cuda_get_device_memory(0, free_mem, total_mem);
#elif GGML_USE_VULKAN
ggml_backend_vk_get_device_memory(0, free_mem, total_mem);
#elif GGML_USE_SYCL
ggml_backend_sycl_get_device_memory(0, free_mem, total_mem);
#else #else
#ifdef _WIN32 #ifdef _WIN32
MEMORYSTATUSEX status; MEMORYSTATUSEX status;
@ -292,7 +125,7 @@ int main(int argc, char * argv[]) {
fprintf(stderr, "\n"); fprintf(stderr, "\n");
} }
ggml_backend_t backend = create_backend(params); ggml_backend_t backend = create_backend();
if (!backend) { if (!backend) {
fprintf(stderr, "Failed to create backend\n"); fprintf(stderr, "Failed to create backend\n");
return 1; return 1;
@ -302,28 +135,11 @@ int main(int argc, char * argv[]) {
if (params.backend_mem > 0) { if (params.backend_mem > 0) {
free_mem = params.backend_mem; free_mem = params.backend_mem;
total_mem = params.backend_mem; total_mem = params.backend_mem;
} } else {
else {
get_backend_memory(&free_mem, &total_mem); get_backend_memory(&free_mem, &total_mem);
} }
const char * cache_dir = nullptr; printf("Starting RPC server on %s, backend memory: %zu MB\n", endpoint.c_str(), free_mem / (1024 * 1024));
std::string cache_dir_str; start_rpc_server(backend, endpoint.c_str(), free_mem, total_mem);
if (params.use_cache) {
cache_dir_str = fs_get_cache_directory() + "rpc/";
if (!fs_create_directory_with_parents(cache_dir_str)) {
fprintf(stderr, "Failed to create cache directory: %s\n", cache_dir_str.c_str());
return 1;
}
cache_dir = cache_dir_str.c_str();
}
printf("Starting RPC server v%d.%d.%d\n",
RPC_PROTO_MAJOR_VERSION,
RPC_PROTO_MINOR_VERSION,
RPC_PROTO_PATCH_VERSION);
printf(" endpoint : %s\n", endpoint.c_str());
printf(" local cache : %s\n", cache_dir ? cache_dir : "n/a");
printf(" backend memory : %zu MB\n", free_mem / (1024 * 1024));
ggml_backend_rpc_start_server(backend, endpoint.c_str(), cache_dir, free_mem, total_mem);
ggml_backend_free(backend); ggml_backend_free(backend);
return 0; return 0;
} }

2
examples/server/server.cpp
View File

@ -12,8 +12,6 @@
#endif #endif
// increase max payload length to allow use of larger context size // increase max payload length to allow use of larger context size
#define CPPHTTPLIB_FORM_URL_ENCODED_PAYLOAD_MAX_LENGTH 1048576 #define CPPHTTPLIB_FORM_URL_ENCODED_PAYLOAD_MAX_LENGTH 1048576
// disable Nagle's algorithm
#define CPPHTTPLIB_TCP_NODELAY true
#include "httplib.h" #include "httplib.h"
// Change JSON_ASSERT from assert() to GGML_ASSERT: // Change JSON_ASSERT from assert() to GGML_ASSERT:
#define JSON_ASSERT GGML_ASSERT #define JSON_ASSERT GGML_ASSERT

1
examples/server/utils.hpp
View File

@ -6,6 +6,7 @@
// Change JSON_ASSERT from assert() to GGML_ASSERT: // Change JSON_ASSERT from assert() to GGML_ASSERT:
#define JSON_ASSERT GGML_ASSERT #define JSON_ASSERT GGML_ASSERT
#include "json.hpp" #include "json.hpp"
#include <string> #include <string>
#include <vector> #include <vector>
#include <sstream> #include <sstream>

7
ggml/include/ggml-rpc.h
View File

@ -7,9 +7,6 @@
extern "C" { extern "C" {
#endif #endif
#define RPC_PROTO_MAJOR_VERSION 2
#define RPC_PROTO_MINOR_VERSION 0
#define RPC_PROTO_PATCH_VERSION 1
#define GGML_RPC_MAX_SERVERS 16 #define GGML_RPC_MAX_SERVERS 16
// backend API // backend API
@ -20,9 +17,7 @@ GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const
GGML_API GGML_CALL void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total); GGML_API GGML_CALL void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total);
GGML_API GGML_CALL void ggml_backend_rpc_start_server(ggml_backend_t backend, const char * endpoint, GGML_API GGML_CALL void start_rpc_server(ggml_backend_t backend, const char * endpoint, size_t free_mem, size_t total_mem);
const char * cache_dir,
size_t free_mem, size_t total_mem);
#ifdef __cplusplus #ifdef __cplusplus
} }

12
ggml/src/ggml-backend.c
View File

@ -1,7 +1,6 @@
#include "ggml-backend-impl.h" #include "ggml-backend-impl.h"
#include "ggml-alloc.h" #include "ggml-alloc.h"
#include "ggml-impl.h" #include "ggml-impl.h"
#include "ggml-rpc.h"
#include <assert.h> #include <assert.h>
#include <limits.h> #include <limits.h>
@ -469,10 +468,6 @@ GGML_CALL static void ggml_backend_registry_init(void) {
extern GGML_CALL int ggml_backend_cann_reg_devices(void); extern GGML_CALL int ggml_backend_cann_reg_devices(void);
ggml_backend_cann_reg_devices(); ggml_backend_cann_reg_devices();
#endif #endif
#ifdef GGML_USE_RPC
extern GGML_CALL void ggml_backend_rpc_reg_devices(void);
ggml_backend_rpc_reg_devices();
#endif
} }
GGML_CALL void ggml_backend_register(const char * name, ggml_backend_init_fn init_fn, ggml_backend_buffer_type_t default_buffer_type, void * user_data) { GGML_CALL void ggml_backend_register(const char * name, ggml_backend_init_fn init_fn, ggml_backend_buffer_type_t default_buffer_type, void * user_data) {
@ -948,13 +943,6 @@ GGML_CALL static ggml_backend_t ggml_backend_reg_cpu_init(const char * params, v
GGML_UNUSED(user_data); GGML_UNUSED(user_data);
} }
GGML_CALL static ggml_backend_t ggml_backend_reg_rpc_init(const char* params, void* user_data) {
return ggml_backend_rpc_init((const char*)user_data);
GGML_UNUSED(params);
GGML_UNUSED(user_data);
}
// multi-buffer buffer // multi-buffer buffer
struct ggml_backend_multi_buffer_context { struct ggml_backend_multi_buffer_context {

1291
ggml/src/ggml-rpc.cpp
View File

File diff suppressed because it is too large Load Diff

2
ggml/src/ggml.c
View File

@ -4935,7 +4935,7 @@ static struct ggml_object * ggml_new_object(struct ggml_context * ctx, enum ggml
if (cur_end + size_needed + GGML_OBJECT_SIZE > ctx->mem_size) { if (cur_end + size_needed + GGML_OBJECT_SIZE > ctx->mem_size) {
GGML_PRINT("%s: not enough space in the context's memory pool (needed %zu, available %zu)\n", GGML_PRINT("%s: not enough space in the context's memory pool (needed %zu, available %zu)\n",
__func__, cur_end + size_needed + GGML_OBJECT_SIZE, ctx->mem_size); __func__, cur_end + size_needed, ctx->mem_size);
assert(false); assert(false);
return NULL; return NULL;
} }

179
src/unicode.cpp
View File

@ -18,7 +18,6 @@
#include <vector> #include <vector>
#include <locale> #include <locale>
#include <codecvt> #include <codecvt>
#include <iostream>
size_t unicode_len_utf8(char src) { size_t unicode_len_utf8(char src) {
const size_t lookup[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 4 }; const size_t lookup[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 4 };
@ -26,7 +25,7 @@ size_t unicode_len_utf8(char src) {
return lookup[highbits]; return lookup[highbits];
} }
static std::string unicode_cpts_to_utf8(const std::vector<uint32_t>& cps) { static std::string unicode_cpts_to_utf8(const std::vector<uint32_t> & cps) {
std::string result; std::string result;
for (size_t i = 0; i < cps.size(); ++i) { for (size_t i = 0; i < cps.size(); ++i) {
result.append(unicode_cpt_to_utf8(cps[i])); result.append(unicode_cpt_to_utf8(cps[i]));
@ -34,7 +33,7 @@ static std::string unicode_cpts_to_utf8(const std::vector<uint32_t>& cps) {
return result; return result;
} }
uint32_t unicode_cpt_from_utf8(const std::string& utf8, size_t& offset) { uint32_t unicode_cpt_from_utf8(const std::string & utf8, size_t & offset) {
assert(offset < utf8.size()); assert(offset < utf8.size());
if (!(utf8[offset + 0] & 0x80)) { if (!(utf8[offset + 0] & 0x80)) {
auto result = utf8[offset + 0]; auto result = utf8[offset + 0];
@ -45,7 +44,7 @@ uint32_t unicode_cpt_from_utf8(const std::string& utf8, size_t& offset) {
throw std::invalid_argument("invalid character"); throw std::invalid_argument("invalid character");
} }
if (!(utf8[offset + 0] & 0x20)) { if (!(utf8[offset + 0] & 0x20)) {
if (offset + 1 >= utf8.size() || !((utf8[offset + 1] & 0xc0) == 0x80)) { if (offset + 1 >= utf8.size() || ! ((utf8[offset + 1] & 0xc0) == 0x80)) {
throw std::invalid_argument("invalid character"); throw std::invalid_argument("invalid character");
} }
auto result = ((utf8[offset + 0] & 0x1f) << 6) | (utf8[offset + 1] & 0x3f); auto result = ((utf8[offset + 0] & 0x1f) << 6) | (utf8[offset + 1] & 0x3f);
@ -53,7 +52,7 @@ uint32_t unicode_cpt_from_utf8(const std::string& utf8, size_t& offset) {
return result; return result;
} }
if (!(utf8[offset + 0] & 0x10)) { if (!(utf8[offset + 0] & 0x10)) {
if (offset + 2 >= utf8.size() || !((utf8[offset + 1] & 0xc0) == 0x80) || !((utf8[offset + 2] & 0xc0) == 0x80)) { if (offset + 2 >= utf8.size() || ! ((utf8[offset + 1] & 0xc0) == 0x80) || ! ((utf8[offset + 2] & 0xc0) == 0x80)) {
throw std::invalid_argument("invalid character"); throw std::invalid_argument("invalid character");
} }
auto result = ((utf8[offset + 0] & 0x0f) << 12) | ((utf8[offset + 1] & 0x3f) << 6) | (utf8[offset + 2] & 0x3f); auto result = ((utf8[offset + 0] & 0x0f) << 12) | ((utf8[offset + 1] & 0x3f) << 6) | (utf8[offset + 2] & 0x3f);
@ -61,7 +60,7 @@ uint32_t unicode_cpt_from_utf8(const std::string& utf8, size_t& offset) {
return result; return result;
} }
if (!(utf8[offset + 0] & 0x08)) { if (!(utf8[offset + 0] & 0x08)) {
if (offset + 3 >= utf8.size() || !((utf8[offset + 1] & 0xc0) == 0x80) || !((utf8[offset + 2] & 0xc0) == 0x80) || !((utf8[offset + 3] & 0xc0) == 0x80)) { if (offset + 3 >= utf8.size() || ! ((utf8[offset + 1] & 0xc0) == 0x80) || ! ((utf8[offset + 2] & 0xc0) == 0x80) || !((utf8[offset + 3] & 0xc0) == 0x80)) {
throw std::invalid_argument("invalid character"); throw std::invalid_argument("invalid character");
} }
auto result = ((utf8[offset + 0] & 0x07) << 18) | ((utf8[offset + 1] & 0x3f) << 12) | ((utf8[offset + 2] & 0x3f) << 6) | (utf8[offset + 3] & 0x3f); auto result = ((utf8[offset + 0] & 0x07) << 18) | ((utf8[offset + 1] & 0x3f) << 12) | ((utf8[offset + 2] & 0x3f) << 6) | (utf8[offset + 3] & 0x3f);
@ -123,10 +122,10 @@ uint32_t unicode_cpt_from_utf8(const std::string& utf8, size_t& offset) {
static std::vector<codepoint_flags> unicode_cpt_flags_array() { static std::vector<codepoint_flags> unicode_cpt_flags_array() {
std::vector<codepoint_flags> cpt_flags(MAX_CODEPOINTS, codepoint_flags::UNDEFINED); std::vector<codepoint_flags> cpt_flags(MAX_CODEPOINTS, codepoint_flags::UNDEFINED);
assert(unicode_ranges_flags.front().first == 0); assert (unicode_ranges_flags.front().first == 0);
assert(unicode_ranges_flags.back().first == MAX_CODEPOINTS); assert (unicode_ranges_flags.back().first == MAX_CODEPOINTS);
for (size_t i = 1; i < unicode_ranges_flags.size(); ++i) { for (size_t i = 1; i < unicode_ranges_flags.size(); ++i) {
const auto range_ini = unicode_ranges_flags[i - 1]; // codepoint_ini, flags const auto range_ini = unicode_ranges_flags[i-1]; // codepoint_ini, flags
const auto range_end = unicode_ranges_flags[i]; // codepoint_end, flags const auto range_end = unicode_ranges_flags[i]; // codepoint_end, flags
for (uint32_t cpt = range_ini.first; cpt < range_end.first; ++cpt) { for (uint32_t cpt = range_ini.first; cpt < range_end.first; ++cpt) {
cpt_flags[cpt] = range_ini.second; cpt_flags[cpt] = range_ini.second;
@ -145,7 +144,7 @@ static std::vector<codepoint_flags> unicode_cpt_flags_array() {
cpt_flags[p.second].is_uppercase = true; cpt_flags[p.second].is_uppercase = true;
} }
for (auto& range : unicode_ranges_nfd) { // start, last, nfd for (auto &range : unicode_ranges_nfd) { // start, last, nfd
cpt_flags[range.nfd].is_nfd = true; cpt_flags[range.nfd].is_nfd = true;
} }
@ -200,55 +199,22 @@ static std::unordered_map<std::string, uint8_t> unicode_utf8_to_byte_map() {
return map; return map;
} }
static inline bool is_valid_utf8(const std::string& str) { static inline std::wstring unicode_wstring_from_utf8(const std::string & s) {
int remaining_bytes = 0; // 当前多字节字符剩余的字节数
for (unsigned char c : str) {
if (remaining_bytes == 0) {
if ((c & 0x80) == 0x00) continue; // 1字节字符
else if ((c & 0xE0) == 0xC0) remaining_bytes = 1; // 2字节
else if ((c & 0xF0) == 0xE0) remaining_bytes = 2; // 3字节
else if ((c & 0xF8) == 0xF0) remaining_bytes = 3; // 4字节
else return false; // 非法起始字节
}
else {
// 检查后续字节是否为10xxxxxx
if ((c & 0xC0) != 0x80)
{
return false;
}
remaining_bytes--;
}
}
return (remaining_bytes == 0); // 确保多字节字符完整
}
static inline std::wstring unicode_wstring_from_utf8(const std::string& s) {
#if defined(__clang__)
// disable C++17 deprecation warning for std::codecvt_utf8
# pragma clang diagnostic push
# pragma clang diagnostic ignored "-Wdeprecated-declarations"
#endif
bool isvalid = is_valid_utf8(s);
std::wstring_convert<std::codecvt_utf8<wchar_t>> conv; std::wstring_convert<std::codecvt_utf8<wchar_t>> conv;
#if defined(__clang__)
# pragma clang diagnostic pop
#endif
return conv.from_bytes(s); return conv.from_bytes(s);
} }
static std::vector<std::string> unicode_byte_encoding_process(const std::vector<std::string>& bpe_words) { static std::vector<std::string> unicode_byte_encoding_process(const std::vector<std::string> & bpe_words) {
std::vector<std::string> bpe_encoded_words; std::vector<std::string> bpe_encoded_words;
for (const auto& word : bpe_words) { for (const auto & word : bpe_words) {
std::string text_utf; std::string text_utf;
auto utf_word = unicode_cpts_from_utf8(word); auto utf_word = unicode_cpts_from_utf8(word);
for (size_t i = 0; i < utf_word.size(); ++i) { for (size_t i = 0; i < utf_word.size(); ++i) {
text_utf += unicode_cpt_to_utf8(utf_word[i]); text_utf += unicode_cpt_to_utf8(utf_word[i]);
} }
std::string encoded_token; std::string encoded_token;
for (char& c : text_utf) { for (char & c : text_utf) {
encoded_token += unicode_byte_to_utf8(c); encoded_token += unicode_byte_to_utf8(c);
} }
bpe_encoded_words.emplace_back(encoded_token); bpe_encoded_words.emplace_back(encoded_token);
@ -257,7 +223,7 @@ static std::vector<std::string> unicode_byte_encoding_process(const std::vector<
} }
// GPT2 system regex: 's|'t|'re|'ve|'m|'ll|'d| ?\p{L}+| ?\p{N}+| ?[^\s\p{L}\p{N}]+|\s+(?!\S)|\s+ // GPT2 system regex: 's|'t|'re|'ve|'m|'ll|'d| ?\p{L}+| ?\p{N}+| ?[^\s\p{L}\p{N}]+|\s+(?!\S)|\s+
static std::vector<size_t> unicode_regex_split_custom_gpt2(const std::string& text, const std::vector<size_t>& offsets) { static std::vector<size_t> unicode_regex_split_custom_gpt2(const std::string & text, const std::vector<size_t> & offsets) {
std::vector<size_t> bpe_offsets; // store the offset of each word std::vector<size_t> bpe_offsets; // store the offset of each word
bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size
@ -271,16 +237,16 @@ static std::vector<size_t> unicode_regex_split_custom_gpt2(const std::string& te
start = offset_end; start = offset_end;
static const uint32_t OUT_OF_RANGE = 0xFFFFFFFF; static const uint32_t OUT_OF_RANGE = 0xFFFFFFFF;
auto _get_cpt = [&](const size_t pos) -> uint32_t { auto _get_cpt = [&] (const size_t pos) -> uint32_t {
return (offset_ini <= pos && pos < offset_end) ? cpts[pos] : OUT_OF_RANGE; return (offset_ini <= pos && pos < offset_end) ? cpts[pos] : OUT_OF_RANGE;
}; };
auto _get_flags = [&](const size_t pos) -> codepoint_flags { auto _get_flags = [&] (const size_t pos) -> codepoint_flags {
return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags(cpts[pos]) : codepoint_flags{}; return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags(cpts[pos]) : codepoint_flags{};
}; };
size_t _prev_end = offset_ini; size_t _prev_end = offset_ini;
auto _add_token = [&](const size_t end) -> size_t { auto _add_token = [&] (const size_t end) -> size_t {
assert(_prev_end <= end && end <= offset_end); assert(_prev_end <= end && end <= offset_end);
size_t len = end - _prev_end; size_t len = end - _prev_end;
if (len > 0) { if (len > 0) {
@ -296,29 +262,29 @@ static std::vector<size_t> unicode_regex_split_custom_gpt2(const std::string& te
return len; return len;
}; };
for (size_t pos = offset_ini; pos < offset_end; /*pos++*/) { for (size_t pos = offset_ini; pos < offset_end; /*pos++*/ ) {
const uint32_t cpt = _get_cpt(pos); const uint32_t cpt = _get_cpt(pos);
const auto flags = _get_flags(pos); const auto flags = _get_flags(pos);
// regex: 's|'t|'re|'ve|'m|'ll|'d // regex: 's|'t|'re|'ve|'m|'ll|'d
if (cpt == '\'' && pos + 1 < offset_end) { if (cpt == '\'' && pos+1 < offset_end) {
uint32_t cpt_next = _get_cpt(pos + 1); uint32_t cpt_next = _get_cpt(pos+1);
if (cpt_next == 's' || cpt_next == 't' || cpt_next == 'm' || cpt_next == 'd') { if (cpt_next == 's' || cpt_next == 't' || cpt_next == 'm' || cpt_next == 'd') {
pos += _add_token(pos + 2); pos += _add_token(pos+2);
continue; continue;
} }
if (pos + 2 < offset_end) { if (pos+2 < offset_end) {
uint32_t cpt_next_next = _get_cpt(pos + 2); uint32_t cpt_next_next = _get_cpt(pos+2);
if ((cpt_next == 'r' && cpt_next_next == 'e') || if ((cpt_next == 'r' && cpt_next_next == 'e') ||
(cpt_next == 'v' && cpt_next_next == 'e') || (cpt_next == 'v' && cpt_next_next == 'e') ||
(cpt_next == 'l' && cpt_next_next == 'l')) { (cpt_next == 'l' && cpt_next_next == 'l')) {
pos += _add_token(pos + 3); pos += _add_token(pos+3);
continue; continue;
} }
} }
} }
auto flags2 = (cpt == ' ' ? _get_flags(pos + 1) : flags); auto flags2 = (cpt == ' ' ? _get_flags(pos+1) : flags);
// regex: <space>?\p{L}+ // regex: <space>?\p{L}+
if (flags2.is_letter) { if (flags2.is_letter) {
pos += (cpt == ' '); pos += (cpt == ' ');
@ -348,12 +314,12 @@ static std::vector<size_t> unicode_regex_split_custom_gpt2(const std::string& te
} }
size_t num_whitespaces = 0; size_t num_whitespaces = 0;
while (_get_flags(pos + num_whitespaces).is_whitespace) { while (_get_flags(pos+num_whitespaces).is_whitespace) {
num_whitespaces++; num_whitespaces++;
} }
// regex: \s+(?!\S) // regex: \s+(?!\S)
if (num_whitespaces > 1 && _get_cpt(pos + num_whitespaces) != OUT_OF_RANGE) { if (num_whitespaces > 1 && _get_cpt(pos+num_whitespaces) != OUT_OF_RANGE) {
pos += num_whitespaces - 1; pos += num_whitespaces - 1;
_add_token(pos); _add_token(pos);
continue; continue;
@ -375,7 +341,7 @@ static std::vector<size_t> unicode_regex_split_custom_gpt2(const std::string& te
} }
// LLAMA3 system regex: "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}{1,3}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+" // LLAMA3 system regex: "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}{1,3}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+"
static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string& text, const std::vector<size_t>& offsets) { static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string & text, const std::vector<size_t> & offsets) {
std::vector<size_t> bpe_offsets; // store the offset of each word std::vector<size_t> bpe_offsets; // store the offset of each word
bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size
@ -389,16 +355,16 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
start = offset_end; start = offset_end;
static const uint32_t OUT_OF_RANGE = 0xFFFFFFFF; static const uint32_t OUT_OF_RANGE = 0xFFFFFFFF;
auto _get_cpt = [&](const size_t pos) -> uint32_t { auto _get_cpt = [&] (const size_t pos) -> uint32_t {
return (offset_ini <= pos && pos < offset_end) ? cpts[pos] : OUT_OF_RANGE; return (offset_ini <= pos && pos < offset_end) ? cpts[pos] : OUT_OF_RANGE;
}; };
auto _get_flags = [&](const size_t pos) -> codepoint_flags { auto _get_flags = [&] (const size_t pos) -> codepoint_flags {
return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags(cpts[pos]) : codepoint_flags{}; return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags(cpts[pos]) : codepoint_flags{};
}; };
size_t _prev_end = offset_ini; size_t _prev_end = offset_ini;
auto _add_token = [&](const size_t end) -> size_t { auto _add_token = [&] (const size_t end) -> size_t {
assert(_prev_end <= end && end <= offset_end); assert(_prev_end <= end && end <= offset_end);
size_t len = end - _prev_end; size_t len = end - _prev_end;
if (len > 0) { if (len > 0) {
@ -414,23 +380,23 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
return len; return len;
}; };
for (size_t pos = offset_ini; pos < offset_end; /*pos++*/) { for (size_t pos = offset_ini; pos < offset_end; /*pos++*/ ) {
const uint32_t cpt = _get_cpt(pos); const uint32_t cpt = _get_cpt(pos);
const auto flags = _get_flags(pos); const auto flags = _get_flags(pos);
// regex: (?i:'s|'t|'re|'ve|'m|'ll|'d) // case insensitive // regex: (?i:'s|'t|'re|'ve|'m|'ll|'d) // case insensitive
if (cpt == '\'' && pos + 1 < offset_end) { if (cpt == '\'' && pos+1 < offset_end) {
uint32_t cpt_next = unicode_tolower(_get_cpt(pos + 1)); uint32_t cpt_next = unicode_tolower(_get_cpt(pos+1));
if (cpt_next == 's' || cpt_next == 't' || cpt_next == 'm' || cpt_next == 'd') { if (cpt_next == 's' || cpt_next == 't' || cpt_next == 'm' || cpt_next == 'd') {
pos += _add_token(pos + 2); pos += _add_token(pos+2);
continue; continue;
} }
if (pos + 2 < offset_end) { if (pos+2 < offset_end) {
uint32_t cpt_next_next = unicode_tolower(_get_cpt(pos + 2)); uint32_t cpt_next_next = unicode_tolower(_get_cpt(pos+2));
if ((cpt_next == 'r' && cpt_next_next == 'e') || if ((cpt_next == 'r' && cpt_next_next == 'e') ||
(cpt_next == 'v' && cpt_next_next == 'e') || (cpt_next == 'v' && cpt_next_next == 'e') ||
(cpt_next == 'l' && cpt_next_next == 'l')) { (cpt_next == 'l' && cpt_next_next == 'l')) {
pos += _add_token(pos + 3); pos += _add_token(pos+3);
continue; continue;
} }
} }
@ -438,7 +404,7 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
// regex: [^\r\n\p{L}\p{N}]?\p{L}+ // regex: [^\r\n\p{L}\p{N}]?\p{L}+
if (!(cpt == '\r' || cpt == '\n' || flags.is_number)) { if (!(cpt == '\r' || cpt == '\n' || flags.is_number)) {
if (flags.is_letter || _get_flags(pos + 1).is_letter) { // one or more letters if (flags.is_letter || _get_flags(pos+1).is_letter) { // one or more letters
pos++; pos++;
while (_get_flags(pos).is_letter) { while (_get_flags(pos).is_letter) {
pos++; pos++;
@ -452,7 +418,7 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
if (flags.is_number) { if (flags.is_number) {
size_t ini = pos; size_t ini = pos;
while (_get_flags(pos).is_number) { while (_get_flags(pos).is_number) {
if (++pos - ini >= 3) { if (++pos - ini >= 3 ) {
_add_token(pos); _add_token(pos);
ini = pos; ini = pos;
} }
@ -462,7 +428,7 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
} }
// regex: <space>?[^\s\p{L}\p{N}]+[\r\n]* // regex: <space>?[^\s\p{L}\p{N}]+[\r\n]*
auto flags2 = (cpt == ' ' ? _get_flags(pos + 1) : flags); auto flags2 = (cpt == ' ' ? _get_flags(pos+1) : flags);
if (!(flags2.is_whitespace | flags2.is_letter | flags2.is_number) && flags.as_uint()) { if (!(flags2.is_whitespace | flags2.is_letter | flags2.is_number) && flags.as_uint()) {
pos += (cpt == ' '); pos += (cpt == ' ');
while (!(flags2.is_whitespace | flags2.is_letter | flags2.is_number) && flags2.as_uint()) { while (!(flags2.is_whitespace | flags2.is_letter | flags2.is_number) && flags2.as_uint()) {
@ -478,8 +444,8 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
size_t num_whitespaces = 0; size_t num_whitespaces = 0;
size_t last_end_r_or_n = 0; size_t last_end_r_or_n = 0;
while (_get_flags(pos + num_whitespaces).is_whitespace) { while (_get_flags(pos+num_whitespaces).is_whitespace) {
uint32_t cpt2 = _get_cpt(pos + num_whitespaces); uint32_t cpt2 = _get_cpt(pos+num_whitespaces);
if (cpt2 == '\r' || cpt2 == '\n') { if (cpt2 == '\r' || cpt2 == '\n') {
last_end_r_or_n = pos + num_whitespaces + 1; last_end_r_or_n = pos + num_whitespaces + 1;
} }
@ -494,7 +460,7 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
} }
// regex: \s+(?!\S) // regex: \s+(?!\S)
if (num_whitespaces > 1 && _get_cpt(pos + num_whitespaces) != OUT_OF_RANGE) { if (num_whitespaces > 1 && _get_cpt(pos+num_whitespaces) != OUT_OF_RANGE) {
pos += num_whitespaces - 1; pos += num_whitespaces - 1;
_add_token(pos); _add_token(pos);
continue; continue;
@ -516,7 +482,7 @@ static std::vector<size_t> unicode_regex_split_custom_llama3(const std::string&
} }
// use std::wregex to split the text // use std::wregex to split the text
static std::vector<size_t> unicode_regex_split_stl(const std::wstring& wtext, const std::wstring& regex_expr, const std::vector<size_t>& offsets) { static std::vector<size_t> unicode_regex_split_stl(const std::wstring & wtext, const std::wstring & regex_expr, const std::vector<size_t> & offsets) {
std::wregex expr(regex_expr); std::wregex expr(regex_expr);
std::vector<size_t> bpe_offsets; // store the offset of each word std::vector<size_t> bpe_offsets; // store the offset of each word
bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size
@ -536,7 +502,7 @@ static std::vector<size_t> unicode_regex_split_stl(const std::wstring& wtext, co
++it; ++it;
} }
if (start_idx < (int64_t)offset) { if (start_idx < (int64_t) offset) {
bpe_offsets.emplace_back(offset - start_idx); bpe_offsets.emplace_back(offset - start_idx);
} }
start += offset; start += offset;
@ -546,7 +512,7 @@ static std::vector<size_t> unicode_regex_split_stl(const std::wstring& wtext, co
} }
// use std::regex to split the text // use std::regex to split the text
static std::vector<size_t> unicode_regex_split_stl(const std::string& text, const std::string& regex_expr, const std::vector<size_t>& offsets) { static std::vector<size_t> unicode_regex_split_stl(const std::string & text, const std::string & regex_expr, const std::vector<size_t> & offsets) {
std::regex expr(regex_expr); std::regex expr(regex_expr);
std::vector<size_t> bpe_offsets; // store the offset of each word std::vector<size_t> bpe_offsets; // store the offset of each word
bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size
@ -566,7 +532,7 @@ static std::vector<size_t> unicode_regex_split_stl(const std::string& text, cons
++it; ++it;
} }
if (start_idx < (int64_t)offset) { if (start_idx < (int64_t) offset) {
bpe_offsets.emplace_back(offset - start_idx); bpe_offsets.emplace_back(offset - start_idx);
} }
start += offset; start += offset;
@ -575,15 +541,14 @@ static std::vector<size_t> unicode_regex_split_stl(const std::string& text, cons
return bpe_offsets; return bpe_offsets;
} }
static std::vector<size_t> unicode_regex_split_custom(const std::string& text, const std::string& regex_expr, const std::vector<size_t>& offsets) { static std::vector<size_t> unicode_regex_split_custom(const std::string & text, const std::string & regex_expr, const std::vector<size_t> & offsets) {
std::vector<size_t> bpe_offsets; std::vector<size_t> bpe_offsets;
if (regex_expr == "'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)") { if (regex_expr == "'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)") {
bpe_offsets = unicode_regex_split_custom_gpt2(text, offsets); bpe_offsets = unicode_regex_split_custom_gpt2(text, offsets);
} } else if (
else if ( regex_expr == "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+" ||
regex_expr == "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+" || regex_expr == "(?:'[sS]|'[tT]|'[rR][eE]|'[vV][eE]|'[mM]|'[lL][lL]|'[dD])|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+") {
regex_expr == "(?:'[sS]|'[tT]|'[rR][eE]|'[vV][eE]|'[mM]|'[lL][lL]|'[dD])|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+") {
bpe_offsets = unicode_regex_split_custom_llama3(text, offsets); bpe_offsets = unicode_regex_split_custom_llama3(text, offsets);
} }
@ -624,8 +589,8 @@ std::string unicode_cpt_to_utf8(uint32_t cp) {
throw std::invalid_argument("invalid codepoint"); throw std::invalid_argument("invalid codepoint");
} }
std::vector<uint32_t> unicode_cpts_normalize_nfd(const std::vector<uint32_t>& cpts) { std::vector<uint32_t> unicode_cpts_normalize_nfd(const std::vector<uint32_t> & cpts) {
auto comp = [](const uint32_t cpt, const range_nfd& range) { auto comp = [] (const uint32_t cpt, const range_nfd & range) {
return cpt < range.first; return cpt < range.first;
}; };
std::vector<uint32_t> result(cpts.size()); std::vector<uint32_t> result(cpts.size());
@ -637,7 +602,7 @@ std::vector<uint32_t> unicode_cpts_normalize_nfd(const std::vector<uint32_t>& cp
return result; return result;
} }
std::vector<uint32_t> unicode_cpts_from_utf8(const std::string& utf8) { std::vector<uint32_t> unicode_cpts_from_utf8(const std::string & utf8) {
std::vector<uint32_t> result; std::vector<uint32_t> result;
result.reserve(utf8.size()); result.reserve(utf8.size());
size_t offset = 0; size_t offset = 0;
@ -653,7 +618,7 @@ codepoint_flags unicode_cpt_flags(const uint32_t cp) {
return cp < cpt_flags.size() ? cpt_flags[cp] : undef; return cp < cpt_flags.size() ? cpt_flags[cp] : undef;
} }
codepoint_flags unicode_cpt_flags(const std::string& utf8) { codepoint_flags unicode_cpt_flags(const std::string & utf8) {
static const codepoint_flags undef(codepoint_flags::UNDEFINED); static const codepoint_flags undef(codepoint_flags::UNDEFINED);
if (utf8.empty()) { if (utf8.empty()) {
return undef; // undefined return undef; // undefined
@ -667,7 +632,7 @@ std::string unicode_byte_to_utf8(uint8_t byte) {
return map.at(byte); return map.at(byte);
} }
uint8_t unicode_utf8_to_byte(const std::string& utf8) { uint8_t unicode_utf8_to_byte(const std::string & utf8) {
static std::unordered_map<std::string, uint8_t> map = unicode_utf8_to_byte_map(); static std::unordered_map<std::string, uint8_t> map = unicode_utf8_to_byte_map();
return map.at(utf8); return map.at(utf8);
} }
@ -677,7 +642,7 @@ uint32_t unicode_tolower(uint32_t cp) {
return it == unicode_map_lowercase.end() ? cp : it->second; return it == unicode_map_lowercase.end() ? cp : it->second;
} }
std::vector<std::string> unicode_regex_split(const std::string& text, const std::vector<std::string>& regex_exprs) { std::vector<std::string> unicode_regex_split(const std::string & text, const std::vector<std::string> & regex_exprs) {
// unicode categories // unicode categories
static const std::map<std::string, int> k_ucat_enum = { static const std::map<std::string, int> k_ucat_enum = {
{ "\\p{N}", codepoint_flags::NUMBER }, { "\\p{N}", codepoint_flags::NUMBER },
@ -706,9 +671,9 @@ std::vector<std::string> unicode_regex_split(const std::string& text, const std:
// compute collapsed codepoints only if needed by at least one regex // compute collapsed codepoints only if needed by at least one regex
bool need_collapse = false; bool need_collapse = false;
for (auto& regex_expr : regex_exprs) { for (auto & regex_expr : regex_exprs) {
// search for unicode categories // search for unicode categories
for (const auto& ucat : k_ucat_enum) { for (const auto & ucat : k_ucat_enum) {
if (std::string::npos != regex_expr.find(ucat.first)) { if (std::string::npos != regex_expr.find(ucat.first)) {
need_collapse = true; need_collapse = true;
break; break;
@ -737,20 +702,18 @@ std::vector<std::string> unicode_regex_split(const std::string& text, const std:
if (flags.is_whitespace) { if (flags.is_whitespace) {
//NOTE: C++ std::regex \s does not mach 0x85, Rust and Python regex does. //NOTE: C++ std::regex \s does not mach 0x85, Rust and Python regex does.
//text_collapsed[i] = (char) 0x85; // <Next Line> as whitespace fallback //text_collapsed[i] = (char) 0x85; // <Next Line> as whitespace fallback
text_collapsed[i] = (char)0x0B; // <vertical tab> as whitespace fallback text_collapsed[i] = (char) 0x0B; // <vertical tab> as whitespace fallback
} } else if (k_ucat_cpt.find(flags.category_flag()) != k_ucat_cpt.end()) {
else if (k_ucat_cpt.find(flags.category_flag()) != k_ucat_cpt.end()) {
text_collapsed[i] = k_ucat_cpt.at(flags.category_flag()); text_collapsed[i] = k_ucat_cpt.at(flags.category_flag());
} } else {
else { text_collapsed[i] = (char) 0xD0; // fallback
text_collapsed[i] = (char)0xD0; // fallback
} }
} }
} }
std::vector<size_t> bpe_offsets = { cpts.size() }; std::vector<size_t> bpe_offsets = { cpts.size() };
for (auto& regex_expr : regex_exprs) { for (auto & regex_expr : regex_exprs) {
// first, see if we have an efficient custom regex implementation // first, see if we have an efficient custom regex implementation
auto tmp = unicode_regex_split_custom(text, regex_expr, bpe_offsets); auto tmp = unicode_regex_split_custom(text, regex_expr, bpe_offsets);
@ -764,7 +727,7 @@ std::vector<std::string> unicode_regex_split(const std::string& text, const std:
// if a unicode category is used in the regex, we use the collapsed text and replace the unicode category // if a unicode category is used in the regex, we use the collapsed text and replace the unicode category
// with the corresponding collapsed representation // with the corresponding collapsed representation
bool use_collapsed = false; bool use_collapsed = false;
for (auto& ucat : k_ucat_enum) { for (auto & ucat : k_ucat_enum) {
if (std::string::npos != regex_expr.find(ucat.first)) { if (std::string::npos != regex_expr.find(ucat.first)) {
use_collapsed = true; use_collapsed = true;
break; break;
@ -823,8 +786,7 @@ std::vector<std::string> unicode_regex_split(const std::string& text, const std:
//printf("text_collapsed: %s\n", text_collapsed.c_str()); //printf("text_collapsed: %s\n", text_collapsed.c_str());
//printf("regex_expr_collapsed: %s\n", regex_expr_collapsed.c_str()); //printf("regex_expr_collapsed: %s\n", regex_expr_collapsed.c_str());
bpe_offsets = unicode_regex_split_stl(text_collapsed, regex_expr_collapsed, bpe_offsets); bpe_offsets = unicode_regex_split_stl(text_collapsed, regex_expr_collapsed, bpe_offsets);
} } else {
else {
// no unicode category used, we can use std::wregex directly // no unicode category used, we can use std::wregex directly
const std::wstring wregex_expr = unicode_wstring_from_utf8(regex_expr); const std::wstring wregex_expr = unicode_wstring_from_utf8(regex_expr);
@ -840,8 +802,7 @@ std::vector<std::string> unicode_regex_split(const std::string& text, const std:
//printf("regex_expr: %s\n", regex_expr.c_str()); //printf("regex_expr: %s\n", regex_expr.c_str());
bpe_offsets = unicode_regex_split_stl(wtext, wregex_expr, bpe_offsets); bpe_offsets = unicode_regex_split_stl(wtext, wregex_expr, bpe_offsets);
} }
} } catch (std::regex_error & e) {
catch (std::regex_error& e) {
fprintf(stderr, "Failed to process regex: '%s'\n", regex_expr.c_str()); fprintf(stderr, "Failed to process regex: '%s'\n", regex_expr.c_str());
fprintf(stderr, "Regex error: %s\n", e.what()); fprintf(stderr, "Regex error: %s\n", e.what());
throw std::runtime_error("Failed to process regex"); throw std::runtime_error("Failed to process regex");
@ -852,7 +813,7 @@ std::vector<std::string> unicode_regex_split(const std::string& text, const std:
bpe_words.reserve(bpe_offsets.size()); // reserve memory for the approximate size bpe_words.reserve(bpe_offsets.size()); // reserve memory for the approximate size
size_t start = 0; size_t start = 0;
for (size_t& offset : bpe_offsets) { for (size_t & offset : bpe_offsets) {
bpe_words.emplace_back(); bpe_words.emplace_back();
for (size_t i = start; i < start + offset; ++i) { for (size_t i = start; i < start + offset; ++i) {
bpe_words.back() += unicode_cpt_to_utf8(cpts[i]); bpe_words.back() += unicode_cpt_to_utf8(cpts[i]);