Non trapping float to int conversion

This commit is contained in:
msqr1
2024-03-24 19:38:54 -07:00
parent 8bb9733cd1
commit 1d4bb738a6
15 changed files with 2765 additions and 238 deletions

File diff suppressed because one or more lines are too long

View File

@@ -4,7 +4,7 @@
<script src="Vosklet.js"></script> <script src="Vosklet.js"></script>
<script> <script>
window.onload = async () => { window.onload = async () => {
window.ctx = new AudioContext() /*window.ctx = new AudioContext()
let micNode = ctx.createMediaStreamSource(await navigator.mediaDevices.getUserMedia({ let micNode = ctx.createMediaStreamSource(await navigator.mediaDevices.getUserMedia({
video: false, video: false,
audio: { audio: {
@@ -13,10 +13,10 @@
channelCount: 1, channelCount: 1,
sampleRate: 16000 sampleRate: 16000
}, },
})) }))*/
window.module = await loadVosklet() window.module = await loadVosklet()
window.model = await module.makeModel("../usage/en-model.tgz","model","ID") window.model = await module.makeModel("../usage/en-model.tgz","model","ID")
window.recognizer = await module.makeRecognizer(model, ctx.sampleRate) window.recognizer = await module.makeRecognizer(model, 48000)
recognizer.addEventListener("result", e => { recognizer.addEventListener("result", e => {
console.log("Result: ", e.detail) console.log("Result: ", e.detail)
}) })

View File

@@ -31,9 +31,9 @@ cd $SRC &&
MODE=2 && # 0: Ultra debug info, 1: Extremely optimized release, else custom MODE=2 && # 0: Ultra debug info, 1: Extremely optimized release, else custom
echo "Mode = $MODE" && echo "Mode = $MODE" &&
if [ $MODE = 0 ]; then if [ $MODE = 0 ]; then
em++ -O0 proxier.cc genericModel.cc recognizer.cc bindings.cc -sWASMFS -sWASM_BIGINT -sSINGLE_FILE -sMODULARIZE -sEMBIND_STD_STRING_IS_UTF8 -sPTHREAD_POOL_DELAY_LOAD -sRUNTIME_DEBUG -sSTACK_OVERFLOW_CHECK=2 -sTEXTDECODER=2 -sPTHREAD_POOL_SIZE_STRICT=2 -sASSERTIONS=2 -sINITIAL_MEMORY=$MAX_MEMORY -sPTHREAD_POOL_SIZE=$MAX_THREADS -sDISABLE_EXCEPTION_CATCHING=0 -sEXIT_RUNTIME=0 -sINVOKE_RUN=0 -sPOLYFILL=0 -sEXPORTED_FUNCTIONS=_malloc -sEXPORT_NAME=loadVosklet -sMALLOC=emmalloc -sEXPORTED_RUNTIME_METHODS=UTF8ToString,stringToUTF8OnStack -sENVIRONMENT=web,worker -I. -I$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 --embind-emit-tsd Vosklet.d.ts -fsanitize=undefined -fsanitize=address -fsanitize=leak -g3 --pre-js pre.js -o ../devel/Vosklet.js em++ -O0 link.cc genericModel.cc recognizer.cc bindings.cc -sWASMFS -sWASM_BIGINT -sSINGLE_FILE -sMODULARIZE -sEMBIND_STD_STRING_IS_UTF8 -sPTHREAD_POOL_DELAY_LOAD -sRUNTIME_DEBUG -sSTACK_OVERFLOW_CHECK=2 -sTEXTDECODER=2 -sPTHREAD_POOL_SIZE_STRICT=2 -sASSERTIONS=2 -sINITIAL_MEMORY=$MAX_MEMORY -sPTHREAD_POOL_SIZE=$MAX_THREADS -sDISABLE_EXCEPTION_CATCHING=0 -sEXIT_RUNTIME=0 -sINVOKE_RUN=0 -sPOLYFILL=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 --embind-emit-tsd Vosklet.d.ts -fsanitize=undefined -fsanitize=address -fsanitize=leak -msimd128 -mnontrapping-fptoint -g3 --pre-js pre.js -o ../devel/Vosklet.js
elif [ $MODE = 1 ]; then elif [ $MODE = 1 ]; then
em++ -O3 proxier.cc genericModel.cc recognizer.cc bindings.cc -sWASMFS -sWASM_BIGINT -sSINGLE_FILE -sMODULARIZE -sEMBIND_STD_STRING_IS_UTF8 -sPTHREAD_POOL_DELAY_LOAD -sTEXTDECODER=2 -sPTHREAD_POOL_SIZE_STRICT=2 -sINITIAL_MEMORY=$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 --pre-js pre.js -o ../devel/Vosklet.js em++ -O3 link.cc genericModel.cc recognizer.cc bindings.cc -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$VOSK/src -l:vosk.a -lembind -pthread -flto -msimd128 -mnontrapping-fptoint --pre-js pre.js -o ../devel/Vosklet.js
else else
em++ -O3 proxier.cc genericModel.cc recognizer.cc bindings.cc -sWASMFS -sWASM_BIGINT -sSINGLE_FILE -sMODULARIZE -sEMBIND_STD_STRING_IS_UTF8 -sPTHREAD_POOL_DELAY_LOAD -sTEXTDECODER=2 -sPTHREAD_POOL_SIZE_STRICT=2 -sINITIAL_MEMORY=$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 -g3 --pre-js pre.js -o ../devel/Vosklet.js em++ -O3 link.cc genericModel.cc recognizer.cc bindings.cc -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 -lopenal -pthread -flto -msimd128 -mnontrapping-fptoint -g3 --pre-js pre.js -o ../devel/Vosklet.js
fi fi

View File

@@ -20,7 +20,7 @@ CLAPACK_WASM:=$(realpath clapack-wasm)
Vosklet.js: | vosk libarchive Vosklet.js: | vosk libarchive
cd $(SRC) && \ cd $(SRC) && \
em++ -O3 proxier.cc genericModel.cc recognizer.cc bindings.cc -sWASMFS \ -sWASM_BIGINT -sSINGLE_FILE -sGZIP_EMBEDDINGS -sMODULARIZE -sEMBIND_STD_STRING_IS_UTF8 -sPTHREAD_POOL_DELAY_LOAD -sTEXTDECODER=2 -sPTHREAD_POOL_SIZE_STRICT=2 -sINITIAL_MEMORY=$(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 -lopfs.js -lembind -pthread -flto -msimd128 --pre-js pre.js -o ../Vosklet.js && \ em++ -O3 link.cc genericModel.cc recognizer.cc bindings.cc -sWASMFS \ -sWASM_BIGINT -sSINGLE_FILE -sGZIP_EMBEDDINGS -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 --pre-js pre.js -o ../Vosklet.js && \
cd .. && \ cd .. && \
rm -f Vosklet.worker.js && \ rm -f Vosklet.worker.js && \

View File

@@ -21,11 +21,13 @@ EMSCRIPTEN_BINDINGS() {
.constructor<int, float, genericModel*>(allow_raw_pointers()) .constructor<int, float, genericModel*>(allow_raw_pointers())
.constructor<int, float, genericModel*, genericModel*>(allow_raw_pointers()) .constructor<int, float, genericModel*, genericModel*>(allow_raw_pointers())
.constructor<int, float, genericModel*, std::string, int>(allow_raw_pointers()) .constructor<int, float, genericModel*, std::string, int>(allow_raw_pointers())
.function("reset", &recognizer::reset, allow_raw_pointers())
.function("setEndpointerMode", &recognizer::setEndpointerMode, allow_raw_pointers())
.function("setEndpointerDelays", &recognizer::setEndpointerDelays, allow_raw_pointers())
.function("setWords", &recognizer::setWords, allow_raw_pointers()) .function("setWords", &recognizer::setWords, allow_raw_pointers())
.function("setPartialWords", &recognizer::setPartialWords, allow_raw_pointers()) .function("setPartialWords", &recognizer::setPartialWords, allow_raw_pointers())
.function("setGrm", &recognizer::setGrm, allow_raw_pointers()) .function("setGrm", &recognizer::setGrm, allow_raw_pointers())
.function("setNLSML", &recognizer::setNLSML, allow_raw_pointers()) .function("setNLSML", &recognizer::setNLSML, allow_raw_pointers())
.function("setSpkModel", &recognizer::setSpkModel, allow_raw_pointers()) .function("setSpkModel", &recognizer::setSpkModel, allow_raw_pointers())
.function("setMaxAlternatives", &recognizer::setMaxAlternatives, allow_raw_pointers()) .function("setMaxAlternatives", &recognizer::setMaxAlternatives, allow_raw_pointers());
.function("acceptWaveForm", &recognizer::acceptWaveForm, allow_raw_pointers());
}; };

View File

@@ -1,11 +1,13 @@
#include "genericModel.h" #include "genericModel.h"
genericModel::genericModel(int index, bool normalMdl, std::string storepath, std::string id) : index(index), normalMdl(normalMdl), storepath(std::move(storepath)), id(std::move(id)), entry(archive_entry_new()) {} genericModel::genericModel(int index, bool normalMdl, std::string storepath, std::string id) : index{index}, normalMdl{normalMdl}, storepath{std::move(storepath)}, id{std::move(id)}, entry{archive_entry_new()} {
blocker.lock();
}
void genericModel::extractAndLoad(int tarStart, int tarSize) { void genericModel::extractAndLoad(int tarStart, int tarSize) {
emscripten_console_log("Untaring");
static fs::path path{}; static fs::path path{};
static int fd{}; static int fd{};
thrd.addTask([this, tarStart, tarSize](){ func = [this, tarStart, tarSize](){
emscripten_console_log("Untaring...");
archive* src {archive_read_new()}; archive* src {archive_read_new()};
archive_read_support_format_tar(src); archive_read_support_format_tar(src);
archive_read_open_memory(src, reinterpret_cast<void*>(tarStart), tarSize); archive_read_open_memory(src, reinterpret_cast<void*>(tarStart), tarSize);
@@ -24,7 +26,6 @@ void genericModel::extractAndLoad(int tarStart, int tarSize) {
} }
path = archive_entry_pathname(entry); path = archive_entry_pathname(entry);
path = storepath + path.generic_string().substr(path.generic_string().find("/")); path = storepath + path.generic_string().substr(path.generic_string().find("/"));
emscripten_console_logf("Creating file or directory %s", path.c_str());
if(!path.has_extension()) { if(!path.has_extension()) {
fs::create_directory(path); fs::create_directory(path);
continue; continue;
@@ -51,7 +52,13 @@ void genericModel::extractAndLoad(int tarStart, int tarSize) {
emscripten_console_log("Loading finished!"); emscripten_console_log("Loading finished!");
if(normalMdl ? std::get<0>(mdl) == nullptr : std::get<1>(mdl) == nullptr) fireEv(index, "Unable to load model for recognition"); if(normalMdl ? std::get<0>(mdl) == nullptr : std::get<1>(mdl) == nullptr) fireEv(index, "Unable to load model for recognition");
else fireEv(index, "0"); else fireEv(index, "0");
}); };
std::thread t{[this](){
func();
blocker.lock();
func();
}};
t.detach();
} }
genericModel::~genericModel() { genericModel::~genericModel() {
archive_entry_free(entry); archive_entry_free(entry);

View File

@@ -1,22 +1,27 @@
#pragma once #pragma once
#include "jsLink.h" #include "link.h"
#include <string>
#include <filesystem> #include <filesystem>
#include <fstream>
#include <variant> #include <variant>
#include <thread>
#include <fcntl.h> #include <fcntl.h>
#include <emscripten/console.h>
#include <vosk_api.h> #include <vosk_api.h>
#include <archive.h> #include <archive.h>
#include <archive_entry.h> #include <archive_entry.h>
namespace fs = std::filesystem; namespace fs = std::filesystem;
struct genericModel { struct genericModel {
bool normalMdl; bool normalMdl;
bool recognizerUsedThrd{}; bool thrdConsumed{};
int index; int index;
std::string storepath; std::string storepath;
std::string id; std::string id;
std::variant<VoskModel*, VoskSpkModel*> mdl; std::variant<VoskModel*, VoskSpkModel*> mdl;
std::function<void()> func;
std::mutex blocker{};
archive_entry* entry; archive_entry* entry;
void extractAndLoad(int tarStart, int tarSize); void extractAndLoad(int tarStart, int tarSize);
genericModel(int index, bool normalMdl, std::string storepath, std::string id); genericModel(int index, bool normalMdl, std::string storepath, std::string id);

View File

@@ -1,15 +0,0 @@
#include <jsLink.h>
pthread_t dstThrd{pthread_self()};
ProxyingQueue glbQ{};
void fireEv(int index, const char* content, const char* type) {
auto proxy{[index, content, type](){
emscripten_console_logf("Firing event, index: %d, content: %s, type: %s", index, content, type);
EM_ASM({
console.log(objs[$0]);
objs[$0].dispatchEvent(new CustomEvent($2 === 0 ? "0" : UTF8ToString($2), {"detail" : UTF8ToString($1)}));
}, index, content, type);
}};
if(dstThrd == pthread_self()) proxy();
else glbQ.proxySync(dstThrd, proxy);
}

22
src/link.cc Normal file
View File

@@ -0,0 +1,22 @@
#include "link.h"
pthread_t dstThrd{pthread_self()};
ProxyingQueue glbQ{};
void fireEv(int index, const char* content, const char* type) {
auto proxy{[index, content, type](){
EM_ASM({
objs[$0].dispatchEvent(new CustomEvent($2 === 0 ? "0" : UTF8ToString($2), { "detail" : UTF8ToString($1) }));
}, index, content, type);
}};
if(dstThrd == pthread_self()) proxy();
else glbQ.proxySync(dstThrd, proxy);
}
void fireEv(int index, char* state, float* dataBuf) {
auto proxy{[index, state, dataBuf](){
EM_ASM({
objs[$0].dispatchEvent(new CustomEvent("0", { "detail" : $1 + "," + $2}));
}, index, state, dataBuf);
}};
if(dstThrd == pthread_self()) proxy();
else glbQ.proxySync(dstThrd, proxy);
}

View File

@@ -1,12 +1,8 @@
#pragma once #pragma once
#include <thread>
#include <mutex>
#include <queue>
#include <emscripten/wasmfs.h>
#include <emscripten/console.h>
#include <emscripten/em_asm.h> #include <emscripten/em_asm.h>
#include <emscripten/proxying.h> #include <emscripten/proxying.h>
using namespace emscripten; using namespace emscripten;
void fireEv(int index, const char* content, const char* type = nullptr);
void fireEv(int index, const char* content, const char* type = nullptr); // Normal
void fireEv(int index, char* state, float* dataBuf); // For recognizer's success initialization

View File

@@ -93,14 +93,14 @@ class Recognizer extends EventTarget {
let rec = new Recognizer() let rec = new Recognizer()
let result = new Promise((resolve, reject) => { let result = new Promise((resolve, reject) => {
rec.addEventListener("0", ev => { rec.addEventListener("0", ev => {
if(ev.detail.indexOf(",") === -1) { if(ev.detail.indexOf(",") !== -1) {
let loadInfo = ev.detail.split(",") let loadInfo = ev.detail.split(",")
rec.dataBuf = Module.HEAPF32.subarray(parseInt(loadInfo[0]), parseInt(loadInfo[0]) + 128) rec.state = Module.HEAP8.subarray(parseInt(loadInfo[0]), parseInt(loadInfo[0]) + 1) // State is an array with 1 element, there is no other way to get a reference to a single element
rec.state = Module.HEAP8.subarray(parseInt(loadInfo[1]), parseInt(loadInfo[1]) + 1) // State is an array with 1 element, there is no other way to get a reference to a single element rec.dataBuf = Module.HEAPF32.subarray(parseInt(loadInfo[1]), parseInt(loadInfo[1]) + 128)
return resolve(rec) return resolve(rec)
} }
rec.delete() rec.delete()
reject(rec) reject(ev.detail)
}, { once : true }) }, { once : true })
}) })
switch(mode) { switch(mode) {

View File

@@ -1,24 +1,22 @@
#include "recognizer.h" #include "recognizer.h"
recognizer::recognizer(int index, float sampleRate, genericModel* model) : index(index) { recognizer::recognizer(int index, float sampleRate, genericModel* model) : index{index}, rec{vosk_recognizer_new(std::get<0>(model->mdl),sampleRate)} {
rec = vosk_recognizer_new(std::get<0>(model->mdl),sampleRate);
finishConstruction(model); finishConstruction(model);
} }
recognizer::recognizer(int index, float sampleRate, genericModel* model, genericModel* spkMdl) : index(index) { recognizer::recognizer(int index, float sampleRate, genericModel* model, genericModel* spkModel) : index(index), rec{vosk_recognizer_new_spk(std::get<0>(model->mdl), sampleRate, std::get<1>(spkModel->mdl))} {
rec = vosk_recognizer_new_spk(std::get<0>(model->mdl), sampleRate, std::get<1>(spkMdl->mdl)); finishConstruction(model, spkModel);
finishConstruction(model, spkMdl);
} }
recognizer::recognizer(int index, float sampleRate, genericModel* model, const std::string& grm, int dummy) : index(index) { recognizer::recognizer(int index, float sampleRate, genericModel* model, const std::string& grm, int dummy) : index{index}, rec{vosk_recognizer_new_grm(std::get<0>(model->mdl), sampleRate, grm.c_str())} {
rec = vosk_recognizer_new_grm(std::get<0>(model->mdl), sampleRate, grm.c_str());
finishConstruction(model); finishConstruction(model);
} }
recognizer::~recognizer() { recognizer::~recognizer() {
done.test_and_set(std::memory_order_relaxed); done.test_and_set(std::memory_order_relaxed);
done.notify_one(); done.notify_one();
controller.test_and_set(std::memory_order_relaxed);
controller.notify_one();
vosk_recognizer_free(rec); vosk_recognizer_free(rec);
} }
void recognizer::reset() {
vosk_recognizer_reset(rec);
}
void recognizer::finishConstruction(genericModel* model, genericModel* spkModel) { void recognizer::finishConstruction(genericModel* model, genericModel* spkModel) {
if(rec == nullptr) { if(rec == nullptr) {
fireEv(index, "Unable to initialize recognizer"); fireEv(index, "Unable to initialize recognizer");
@@ -26,34 +24,36 @@ void recognizer::finishConstruction(genericModel* model, genericModel* spkModel)
} }
auto main {[this](){ auto main {[this](){
emscripten_console_log("Recognizer loaded!"); emscripten_console_log("Recognizer loaded!");
std::string loadInfo{}; fireEv(index, &state, dataBuf);
loadInfo += reinterpret_cast<int>(dataBuf); int sample{};
loadInfo += ","; char buffer[22480];
loadInfo += reinterpret_cast<int>(&state); ALCdevice* mic{alcCaptureOpenDevice("Emscripten OpenAL capture", sampleRate, AL_FORMAT_MONO16, 22480)};
fireEv(index, loadInfo.c_str()); alcCaptureStart(mic);
while(!done.test(std::memory_order_relaxed)) { while(!done.test(std::memory_order_relaxed)) {
controller.wait(!done.test(std::memory_order_relaxed), std::memory_order_relaxed); alcGetIntegerv(mic, ALC_CAPTURE_SAMPLES, 4, &sample);
controller.clear(std::memory_order_relaxed); alcCaptureSamples(mic, buffer, sample);
switch(vosk_recognizer_accept_waveform_f(rec, dataBuf, 512)) { switch(vosk_recognizer_accept_waveform(rec, buffer, 22480)) {
case 0: case 0:
fireEv(index, vosk_recognizer_result(rec), "result"); fireEv(index, vosk_recognizer_result(rec), "result");
break; break;
case 1: case 1:
fireEv(index, vosk_recognizer_partial_result(rec), "partialResult"); fireEv(index, vosk_recognizer_partial_result(rec), "partialResult");
} }
state = 0; }
} alcCaptureCloseDevice(mic);
}}; }};
if(!model->recognizerUsedThrd) { if(!model->thrdConsumed) {
model->recognizerUsedThrd = true; model->thrdConsumed = true;
model->thrd.addTask(main); model->func = main;
emscripten_console_log("Adding task to model thread..."); model->blocker.unlock();
emscripten_console_log("Using model's thread");
return; return;
} }
if(spkModel != nullptr && !spkModel->recognizerUsedThrd) { if(spkModel != nullptr && !spkModel->thrdConsumed) {
spkModel->recognizerUsedThrd = true; spkModel->thrdConsumed = true;
spkModel->thrd.addTask(main); spkModel->func = main;
emscripten_console_log("Adding task to speaker model thread..."); spkModel->blocker.unlock();
emscripten_console_log("Using speaker model's thread");
return; return;
} }
emscripten_console_log("New recognizer thread"); emscripten_console_log("New recognizer thread");
@@ -66,10 +66,6 @@ void recognizer::setEndpointerMode(VoskEndpointerMode mode) {
void recognizer::setEndpointerDelays(float tStartMax, float tEnd, float tMax) { void recognizer::setEndpointerDelays(float tStartMax, float tEnd, float tMax) {
vosk_recognizer_set_endpointer_delays(rec, tStartMax, tEnd, tMax); vosk_recognizer_set_endpointer_delays(rec, tStartMax, tEnd, tMax);
} }
void recognizer::acceptWaveForm() {
controller.test_and_set(std::memory_order_relaxed);
controller.notify_one();
}
void recognizer::setGrm(const std::string& grm) { void recognizer::setGrm(const std::string& grm) {
vosk_recognizer_set_grm(rec, grm.c_str()); vosk_recognizer_set_grm(rec, grm.c_str());
} }

View File

@@ -1,17 +1,21 @@
#pragma once #pragma once
#include "genericModel.h" #include "genericModel.h"
#include <AL/al.h>
#include <AL/alc.h>
struct recognizer { struct recognizer {
std::atomic_flag done{}; std::atomic_flag done{};
std::atomic_flag controller{};
float dataBuf[128];
char state; // 0: Copying data from JS, 1: Processing from C++ char state; // 0: Copying data from JS, 1: Processing from C++
float dataBuf[128];
float sampleRate;
int index; int index;
VoskRecognizer* rec; VoskRecognizer* rec;
recognizer(int index, float sampleRate, genericModel* model); recognizer(int index, float sampleRate, genericModel* model);
recognizer(int index, float sampleRate, genericModel* model, genericModel* spkModel); recognizer(int index, float sampleRate, genericModel* model, genericModel* spkModel);
recognizer(int index, float sampleRate, genericModel* model, const std::string& grm, int dummy); recognizer(int index, float sampleRate, genericModel* model, const std::string& grm, int dummy);
~recognizer(); ~recognizer();
void reset();
void finishConstruction(genericModel* model, genericModel* spkModel = nullptr); void finishConstruction(genericModel* model, genericModel* spkModel = nullptr);
void setEndpointerMode(VoskEndpointerMode mode); void setEndpointerMode(VoskEndpointerMode mode);
void setEndpointerDelays(float tStartMax, float tEnd, float tMax); void setEndpointerDelays(float tStartMax, float tEnd, float tMax);

View File

@@ -1,135 +0,0 @@
// Taken from https://github.com/libarchive/libarchive/blob/master/contrib/untar.c
// Stripped all the way down for smallest size
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <filesystem>
#include <emscripten/console.h>
namespace fs = std::filesystem;
#define TO_STRING(x) #x
#define STR(x) TO_STRING(x)
/* Parse an octal number, ignoring leading and trailing nonsense. */
int parseoct(const char *p, size_t n)
{
int i = 0;
while ((*p < '0' || *p > '7') && n > 0) {
++p;
--n;
}
while (*p >= '0' && *p <= '7' && n > 0) {
i *= 8;
i += *p - '0';
++p;
--n;
}
return i;
}
/* Returns true if this is 512 zero bytes. */
bool is_end_of_archive(const char *p)
{
int n;
for (n = 511; n >= 0; --n)
if (p[n] != '\0')
return false;
return true;
}
/* Create a file, including parent directory as necessary. */
FILE * create_file(char *pathname) {
FILE *f;
if (f == NULL) {
/* Try creating parent dir and then creating file. */
char *p = strrchr(pathname, '/');
if (p != NULL) {
*p = '\0';
fs::create_directories(pathname);
*p = '/';
f = fopen(pathname, "wb+");
}
}
return f;
}
/* Verify the tar checksum. */
int verify_checksum(const char *p)
{
int n, u = 0;
for (n = 0; n < 512; ++n) {
if (n < 148 || n > 155)
/* Standard tar checksum adds unsigned bytes. */
u += ((unsigned char *)p)[n];
else
u += 0x20;
}
return (u == parseoct(p + 148, 8));
}
/* Extract a tar archive. */
bool untar(FILE *a, const char *path)
{
char buff[512];
FILE *f = NULL;
size_t bytes_read;
off_t filesize;
emscripten_console_logf("Extracting from %s", path);
for (;;) {
bytes_read = fread(buff, 1, 512, a);
if (bytes_read < 512) {
emscripten_console_logf("Short read on %s", path);
return false;
}
if (is_end_of_archive(buff)) {
emscripten_console_logf("End of %s", path);
return true;
}
if (!verify_checksum(buff)) {
emscripten_console_log("Checksum failure");
return false;
}
filesize = parseoct(buff + 124, 12);
switch (buff[156]) {
// Directory
case '5':
emscripten_console_logf(" Extracting dir %s", buff);
fs::create_directories(buff);
filesize = 0;
break;
// Normal file
case '0':
case '\0':
emscripten_console_logf(" Extracting file %s", buff);
f = create_file(buff);
}
while (filesize > 0) {
bytes_read = fread(buff, 1, 512, a);
if (bytes_read < 512) {
emscripten_console_logf("Short read on %s", path);
return false;
}
if (filesize < 512)
bytes_read = filesize;
if (f != NULL) {
if (fwrite(buff, 1, bytes_read, f)
!= bytes_read)
{
emscripten_console_log("Failed write");
fclose(f);
f = NULL;
}
}
filesize -= (off_t)bytes_read;
}
if (f != NULL) {
fclose(f);
f = NULL;
}
}
return true;
}

View File

@@ -34,8 +34,8 @@
| Event | Description | | Event | Description |
|---|---| |---|---|
| ```partialResult``` | There is a partial recognition result, check the event's "details" property | | ```partialResult``` | There is a partial recognition result, check the event's *detail* property |
| ```result``` | There is a full recognition result, check the event's "details" property | | ```result``` | There is a full recognition result, check the event's *detail* property |
# User agent notes # User agent notes
## SharedArrayBuffer ## SharedArrayBuffer