Remove libarchive, implement untar manually. Use vosk-browser models

This commit is contained in:
msqr1
2024-06-02 02:53:30 -07:00
parent 43048b0685
commit 410bff9be1
14 changed files with 122 additions and 98 deletions

View File

@@ -1,49 +1,27 @@
#include "genericModel.h"
genericModel::genericModel(int index, bool normalMdl, std::string storepath, std::string id) : normalMdl{normalMdl}, index{index}, storepath{std::move(storepath)}, id{std::move(id)}, entry{archive_entry_new()} {}
genericModel::genericModel(int index, bool normalMdl, std::string storepath, std::string id) : normalMdl{normalMdl}, index{index}, storepath{std::move(storepath)}, id{std::move(id)} {}
void genericModel::extractAndLoad(int tarStart, int tarSize) {
static fs::path path{};
static int fd{};
void* tar{reinterpret_cast<void*>(tarStart)};
func = [this, tar, tarSize](){
archive* src {archive_read_new()};
archive_read_support_format_tar(src);
archive_read_open_memory(src, tar, tarSize);
if(archive_errno(src) != 0) {
free(tar);
fireEv(index, "Unable to open tar in WASM memory");
return;
}
while(1) {
int headerRes {archive_read_next_header2(src, entry)};
if(headerRes == ARCHIVE_EOF) break;
if(headerRes < ARCHIVE_OK) {
free(tar);
fireEv(index, archive_error_string(src));
return;
}
path = archive_entry_pathname(entry);
path = storepath + path.string().substr(path.string().find("/"));
if(!path.has_extension()) {
fs::create_directory(path);
continue;
}
fd = open(path.c_str(), O_CREAT | O_WRONLY | O_TRUNC, 0777);
if(fd == -1) {
free(tar);
fireEv(index, "Unable to create model files");
return;
}
archive_read_data_into_fd(src, fd);
close(fd);
if(archive_errno(src) != 0) {
free(tar);
fireEv(index, "Cannot write into model files");
return;
}
}
func = [this, tar = reinterpret_cast<unsigned char*>(tarStart), tarSize]() {
int res{untar(tar, tarSize, storepath)};
free(tar);
archive_read_free(src);
switch(res) {
case 1:
fireEv(index, "Untar: Incorrect tar format, must be USTAR");
return;
case 2:
fireEv(index, "Untar: Not a directory or regular file");
return;
case 3:
fireEv(index, "Untar: Unable to open file for write");
return;
case 4:
fireEv(index, "Untar: Unable to write file");
return;
case 5:
fireEv(index, "Untar: Unable to close file after write");
return;
};
if(normalMdl) mdl = vosk_model_new(storepath.c_str());
else mdl = vosk_spk_model_new(storepath.c_str());
if(normalMdl ? std::get<VoskModel*>(mdl) != nullptr : std::get<VoskSpkModel*>(mdl) != nullptr) fireEv(index, "0");
@@ -61,7 +39,6 @@ int genericModel::findWord(std::string word) {
return vosk_model_find_word(std::get<VoskModel*>(mdl), word.c_str());
}
genericModel::~genericModel() {
archive_entry_free(entry);
if(normalMdl) vosk_model_free(std::get<VoskModel*>(mdl));
else vosk_spk_model_free(std::get<VoskSpkModel*>(mdl));
}

View File

@@ -1,15 +1,7 @@
#pragma once
#include "link.h"
#include <filesystem>
#include <functional>
#include <variant>
#include <fcntl.h>
#include "util.h"
#include <vosk_api.h>
#include <archive.h>
#include <archive_entry.h>
namespace fs = std::filesystem;
extern void free(void*);
struct genericModel {
@@ -21,7 +13,6 @@ struct genericModel {
std::string id;
std::variant<VoskModel*, VoskSpkModel*> mdl;
std::function<void()> func;
archive_entry* entry;
void extractAndLoad(int tarStart, int tarSize);
int findWord(std::string word);
genericModel(int index, bool normalMdl, std::string storepath, std::string id);

View File

@@ -1,7 +0,0 @@
#include "link.h"
void fireEv(int index, const char* content, const char* type) {
MAIN_THREAD_EM_ASM({
objs[$0].dispatchEvent(new CustomEvent($2 === 0 ? "0" : UTF8ToString($2), { "detail" : UTF8ToString($1) }));
}, index, content, type);
}

View File

@@ -1,6 +0,0 @@
#pragma once
#include <thread>
#include <emscripten/em_asm.h>
void fireEv(int index, const char* content, const char* type = nullptr);

View File

@@ -28,25 +28,15 @@ if [ $EMSDK = ../emsdk ] && [ ! -d $EMSDK ]; then
./emsdk activate 3.1.59
fi
. $(realpath $EMSDK)/emsdk_env.sh &&
export PATH=:$PATH:$(realpath $EMSDK)/upstream/bin
export PATH=:$PATH:$(realpath $EMSDK)/upstream/bin &&
cd ..
cd .. &&
SRC=$(realpath src)
KALDI=$(realpath kaldi)
VOSK=$(realpath vosk)
OPENFST=$(realpath openfst)
LIBARCHIVE=$(realpath libarchive)
CLAPACK_WASM=$(realpath clapack-wasm)
if [ ! -d $LIBARCHIVE ]; then
rm -rf /tmp/libarchive &&
git clone -b v3.7.4 --depth=1 https://github.com/libarchive/libarchive /tmp/libarchive &&
cd /tmp/libarchive &&
build/autogen.sh &&
CFLAGS="-O3 -flto -msimd128 -mreference-types -mnontrapping-fptoint -mextended-const -msign-ext -mmutable-globals" LDFLAGS="-O3 -flto" emconfigure ./configure --prefix=$LIBARCHIVE --without-lz4 --without-lzma --without-zlib --without-bz2lib --without-xml2 --without-expat --without-cng --without-openssl --without-libb2 --without-zstd --disable-bsdunzip --disable-xattr --disable-acl --disable-bsdcpio --disable-bsdcat --disable-rpath --disable-maintainer-mode --disable-dependency-tracking --enable-static --disable-shared &&
emmake make -j$JOBS install
fi
if [ ! -d $OPENFST ]; then
rm -rf /tmp/openfst &&
git clone --depth=1 https://github.com/alphacep/openfst /tmp/openfst &&
@@ -82,13 +72,13 @@ if [ ! -d $VOSK ]; then
fi
cd $SRC &&
em++ link.cc genericModel.cc recognizer.cc bindings.cc -O3 -sWASMFS -sWASM_BIGINT -sSINGLE_FILE -sMODULARIZE -sEMBIND_STD_STRING_IS_UTF8 -sPTHREAD_POOL_DELAY_LOAD -sTEXTDECODER=2 -sPTHREAD_POOL_SIZE_STRICT=2 -sINITIAL_MEMORY=$MAX_MEMORY -sPTHREAD_POOL_SIZE=$MAX_THREADS -sPOLYFILL=0 -sEXIT_RUNTIME=0 -sINVOKE_RUN=0 -sSUPPORT_LONGJMP=0 -sEXPORTED_FUNCTIONS=_malloc -sEXPORT_NAME=loadVosklet -sMALLOC=emmalloc -sEXPORTED_RUNTIME_METHODS=UTF8ToString,stringToUTF8OnStack -sENVIRONMENT=web,worker -I. -I$LIBARCHIVE/include -I$VOSK/src -L$LIBARCHIVE/lib -larchive -L$KALDI/src -l:online2/kaldi-online2.a -l:decoder/kaldi-decoder.a -l:ivector/kaldi-ivector.a -l:gmm/kaldi-gmm.a -l:tree/kaldi-tree.a -l:feat/kaldi-feat.a -l:cudamatrix/kaldi-cudamatrix.a -l:lat/kaldi-lat.a -l:lm/kaldi-lm.a -l:rnnlm/kaldi-rnnlm.a -l:hmm/kaldi-hmm.a -l:nnet3/kaldi-nnet3.a -l:transform/kaldi-transform.a -l:matrix/kaldi-matrix.a -l:fstext/kaldi-fstext.a -l:util/kaldi-util.a -l:base/kaldi-base.a -L$OPENFST/lib -l:libfst.a -l:libfstngram.a -L$CLAPACK_WASM -l:CBLAS/lib/cblas.a -l:CLAPACK-3.2.1/lapack.a -l:CLAPACK-3.2.1/libcblaswr.a -l:f2c_BLAS-3.8.0/blas.a -l:libf2c/libf2c.a -L$VOSK/src -l:vosk.a -lembind -pthread -flto -msimd128 -mreference-types -mnontrapping-fptoint -mextended-const -msign-ext -mmutable-globals --pre-js wrapper.js -o ../Vosklet.js &&
em++ util.cc genericModel.cc recognizer.cc bindings.cc -O3 -sWASMFS -sWASM_BIGINT -sSINGLE_FILE -sMODULARIZE -sEMBIND_STD_STRING_IS_UTF8 -sPTHREAD_POOL_DELAY_LOAD -sTEXTDECODER=2 -sPTHREAD_POOL_SIZE_STRICT=2 -sINITIAL_MEMORY=$MAX_MEMORY -sPTHREAD_POOL_SIZE=$MAX_THREADS -sPOLYFILL=0 -sEXIT_RUNTIME=0 -sINVOKE_RUN=0 -sSUPPORT_LONGJMP=0 -sEXPORTED_FUNCTIONS=_malloc -sEXPORT_NAME=loadVosklet -sMALLOC=emmalloc -sEXPORTED_RUNTIME_METHODS=UTF8ToString,stringToUTF8OnStack -sENVIRONMENT=web,worker -I. -I"$VOSK"/src -L$KALDI/src -l:online2/kaldi-online2.a -l:decoder/kaldi-decoder.a -l:ivector/kaldi-ivector.a -l:gmm/kaldi-gmm.a -l:tree/kaldi-tree.a -l:feat/kaldi-feat.a -l:cudamatrix/kaldi-cudamatrix.a -l:lat/kaldi-lat.a -l:lm/kaldi-lm.a -l:rnnlm/kaldi-rnnlm.a -l:hmm/kaldi-hmm.a -l:nnet3/kaldi-nnet3.a -l:transform/kaldi-transform.a -l:matrix/kaldi-matrix.a -l:fstext/kaldi-fstext.a -l:util/kaldi-util.a -l:base/kaldi-base.a -L"$OPENFST"/lib -l:libfst.a -l:libfstngram.a -L"$CLAPACK_WASM" -l:CBLAS/lib/cblas.a -l:CLAPACK-3.2.1/lapack.a -l:CLAPACK-3.2.1/libcblaswr.a -l:f2c_BLAS-3.8.0/blas.a -l:libf2c/libf2c.a -L$VOSK/src -l:vosk.a -lembind -pthread -flto -msimd128 -mreference-types -mnontrapping-fptoint -mextended-const -msign-ext -mmutable-globals --pre-js wrapper.js -o ../Vosklet.js &&
cd .. &&
rm -f Vosklet.worker.js
cp Vosklet.js examples/Vosklet.js &&
# Can't serve files from raw.githubusercontent with Content-Encoding: gzip header so the browser won't decompress automatically. Change to manually decompressing instead.
# Can't serve files from raw.githubusercontent with Content-Encoding: gzip header so the browser won't decompress automatically. Manually decompressing instead.
sed -i 's/res.body/new Response(res.body.pipeThrough(new DecompressionStream("gzip"))).body/' examples/Vosklet.js &&
rm -rf /tmp/libarchive

47
src/util.cc Normal file
View File

@@ -0,0 +1,47 @@
#include "util.h"
void fireEv(int index, const char* content, const char* type) {
MAIN_THREAD_EM_ASM({
objs[$0].dispatchEvent(new CustomEvent($2 === 0 ? "0" : UTF8ToString($2), { "detail" : UTF8ToString($1) }));
}, index, content, type);
}
int untar(unsigned char* tar, int tarSize, const std::string& storepath) {
if(memcmp(tar + 257, "ustar", 5)) return 1;
unsigned char* ptr = tar;
size_t size{0};
std::string path{};
path.reserve(100); // Max length
unsigned char* end = tar + tarSize;
while(ptr <= end) {
if(ptr[156] != '5' && ptr[156] != 0 &&
ptr[156] != '0') {
return 2;
}
path.clear();
path += reinterpret_cast<char*>(ptr + 345);
path += reinterpret_cast<char*>(ptr);
ptr += 124;
for(int i{0}; i < 11; i++) {
size *= 8;
size += *ptr - 48;
ptr++;
}
ptr += 377;
size_t firstSlash = path.find_first_of("/");
if(firstSlash == std::string::npos) {
if(size != 0) ptr += size + 512 - size % 512;
continue;
}
path = storepath + path.substr(firstSlash);
if(size == 0) fs::create_directory(path);
else {
int fd {open(path.c_str(), O_CREAT | O_WRONLY | O_TRUNC, 0777)};
if(fd == -1) return 3;
int res = write(fd, ptr, size);
if(res == -1) return 4;
if(close(fd) == -1) return 5;
ptr += size + 512 - size % 512;
}
}
return 0;
}

15
src/util.h Normal file
View File

@@ -0,0 +1,15 @@
#pragma once
#include <thread>
#include <filesystem>
#include <functional>
#include <variant>
#include <unistd.h>
#include <fcntl.h>
#include <emscripten/em_asm.h>
#include <emscripten/console.h>
namespace fs = std::filesystem;
void fireEv(int index, const char* content, const char* type = nullptr);
int untar(unsigned char* tar, int tarSize, const std::string& storepath);