Merge with master

.github/scripts/node-addon/README-optional.md

@@ -0,0 +1,3 @@
+# Introduction
+
+Please see [sherpa-onnx-node](https://www.npmjs.com/package/sherpa-onnx-node)

.github/scripts/node-addon/README.md

@@ -0,0 +1,11 @@
+# Introduction
+
+Please see
+https://github.com/k2-fsa/sherpa-onnx/blob/master/nodejs-addon-examples/README.md
+for usages.
+
+
+||Method|Support multiple threads|Minimum required node version|
+|---|---|---|---|
+|this package| https://github.com/nodejs/node-addon-api | Yes | v10|
+|https://www.npmjs.com/package/sherpa-onnx| WebAssembly | No | v18|

.github/scripts/node-addon/index.js

@@ -0,0 +1 @@
+module.exports = require('./sherpa-onnx.node');

.github/scripts/node-addon/notes.md

@@ -0,0 +1,9 @@
+# Introduction
+
+See also
+
+  - https://github.com/WonderInventions/node-webrtc/blob/develop/package.json
+  - https://stackoverflow.com/questions/15176082/npm-package-json-os-specific-dependency
+  - https://github.com/WonderInventions/node-webrtc/blob/develop/lib/binding.js
+  - cross-compiling https://github.com/nodejs/node-gyp/issues/829#issuecomment-665527032
+  - https://nodejs.github.io/node-addon-examples/build-tools/cmake-js

.github/scripts/node-addon/package-optional.json

@@ -0,0 +1,53 @@
+{
+  "name": "sherpa-onnx-PLATFORM2-ARCH",
+  "version": "SHERPA_ONNX_VERSION",
+  "description": "Speech-to-text and text-to-speech using Next-gen Kaldi without internet connection",
+  "main": "index.js",
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/k2-fsa/sherpa-onnx.git"
+  },
+  "keywords": [
+    "speech to text",
+    "text to speech",
+    "transcription",
+    "real-time speech recognition",
+    "without internet connection",
+    "embedded systems",
+    "open source",
+    "zipformer",
+    "asr",
+    "tts",
+    "stt",
+    "c++",
+    "onnxruntime",
+    "onnx",
+    "ai",
+    "next-gen kaldi",
+    "offline",
+    "privacy",
+    "open source",
+    "vad",
+    "speaker id",
+    "language id",
+    "node-addon-api",
+    "streaming speech recognition",
+    "speech",
+    "recognition"
+  ],
+  "author": "The next-gen Kaldi team",
+  "license": "Apache-2.0",
+  "bugs": {
+    "url": "https://github.com/k2-fsa/sherpa-onnx/issues"
+  },
+  "homepage": "https://github.com/k2-fsa/sherpa-onnx#readme",
+   "os": [
+    "PLATFORM"
+  ],
+  "cpu": [
+    "ARCH"
+  ]
+}

.github/scripts/node-addon/package.json

@@ -0,0 +1,54 @@
+{
+  "name": "sherpa-onnx-node",
+  "version": "SHERPA_ONNX_VERSION",
+  "description": "Speech-to-text and text-to-speech using Next-gen Kaldi without internet connection",
+  "main": "sherpa-onnx.js",
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/k2-fsa/sherpa-onnx.git"
+  },
+  "keywords": [
+    "speech to text",
+    "text to speech",
+    "transcription",
+    "real-time speech recognition",
+    "without internet connection",
+    "embedded systems",
+    "open source",
+    "zipformer",
+    "asr",
+    "tts",
+    "stt",
+    "c++",
+    "onnxruntime",
+    "onnx",
+    "ai",
+    "next-gen kaldi",
+    "offline",
+    "privacy",
+    "open source",
+    "vad",
+    "speaker id",
+    "language id",
+    "node-addon-api",
+    "streaming speech recognition",
+    "speech",
+    "recognition"
+  ],
+  "author": "The next-gen Kaldi team",
+  "license": "Apache-2.0",
+  "bugs": {
+    "url": "https://github.com/k2-fsa/sherpa-onnx/issues"
+  },
+  "homepage": "https://github.com/k2-fsa/sherpa-onnx#readme",
+  "optionalDependencies": {
+    "sherpa-onnx-darwin-arm64": "^SHERPA_ONNX_VERSION",
+    "sherpa-onnx-darwin-x64": "^SHERPA_ONNX_VERSION",
+    "sherpa-onnx-linux-x64": "^SHERPA_ONNX_VERSION",
+    "sherpa-onnx-linux-arm64": "^SHERPA_ONNX_VERSION",
+    "sherpa-onnx-win-x64": "^SHERPA_ONNX_VERSION"
+  }
+}

.github/scripts/node-addon/run.sh

@@ -0,0 +1,45 @@
+#!/usr/bin/env bash
+
+sherpa_onnx_dir=$PWD
+echo "sherpa_onnx_dir: $sherpa_onnx_dir"
+
+src_dir=$sherpa_onnx_dir/.github/scripts/node-addon
+
+platform=$(node -p "require('os').platform()")
+
+arch=$(node -p "require('os').arch()")
+
+platform2=$platform
+
+if [[ $platform == win32 ]]; then
+  platform2=win
+fi
+
+SHERPA_ONNX_VERSION=$(grep "SHERPA_ONNX_VERSION" ./CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)
+echo "SHERPA_ONNX_VERSION $SHERPA_ONNX_VERSION"
+
+if [ -z $owner ]; then
+  owner=k2-fsa
+fi
+
+sed -i.bak s/SHERPA_ONNX_VERSION/$SHERPA_ONNX_VERSION/g $src_dir/package-optional.json
+sed -i.bak s/k2-fsa/$owner/g $src_dir/package-optional.json
+sed -i.bak s/PLATFORM2/$platform2/g $src_dir/package-optional.json
+sed -i.bak s/PLATFORM/$platform/g $src_dir/package-optional.json
+sed -i.bak s/ARCH/$arch/g $src_dir/package-optional.json
+
+git diff $src_dir/package-optional.json
+
+dst=$sherpa_onnx_dir/sherpa-onnx-node
+mkdir -p $dst
+
+cp $src_dir/package-optional.json $dst/package.json
+cp $src_dir/index.js $dst/
+cp $src_dir/README-optional.md $dst/README.md
+
+cp -fv build/install/lib/lib* $dst/ || true
+cp -fv build/install/lib/*dll $dst/ || true
+
+cp scripts/node-addon-api/build/Release/sherpa-onnx.node $dst/
+
+ls $dst

.github/scripts/test-dot-net.sh

@@ -24,6 +24,7 @@ cd ../offline-decode-files
 cd ../offline-tts
 ./run-aishell3.sh
 ./run-piper.sh
+./run-hf-fanchen.sh
 ls -lh
 
 cd ../..

.github/scripts/test-nodejs-addon-npm.sh

@@ -0,0 +1,24 @@
+#!/usr/bin/env bash
+
+set -ex
+
+d=nodejs-addon-examples
+echo "dir: $d"
+cd $d
+npm install --verbose
+git status
+ls -lh
+ls -lh node_modules
+
+export DYLD_LIBRARY_PATH=$PWD/node_modules/sherpa-onnx-darwin-x64:$DYLD_LIBRARY_PATH
+export DYLD_LIBRARY_PATH=$PWD/node_modules/sherpa-onnx-darwin-arm64:$DYLD_LIBRARY_PATH
+export LD_LIBRARY_PATH=$PWD/node_modules/sherpa-onnx-linux-x64:$LD_LIBRARY_PATH
+export LD_LIBRARY_PATH=$PWD/node_modules/sherpa-onnx-linux-arm64:$LD_LIBRARY_PATH
+
+curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20.tar.bz2
+tar xvf sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20.tar.bz2
+rm sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20.tar.bz2
+
+node test_asr_streaming_transducer.js
+
+rm -rf sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20

.github/scripts/test-offline-ctc.sh

@@ -13,14 +13,111 @@ echo "PATH: $PATH"
 
 which $EXE
 
+log "-----------------------------------------------------------------"
+log "Run Nemo fast conformer hybrid transducer ctc models (CTC branch)"
+log "-----------------------------------------------------------------"
+
+url=https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-nemo-fast-conformer-ctc-be-de-en-es-fr-hr-it-pl-ru-uk-20k.tar.bz2
+name=$(basename $url)
+curl -SL -O $url
+tar xvf $name
+rm $name
+repo=$(basename -s .tar.bz2 $name)
+ls -lh $repo
+
+log "test $repo"
+test_wavs=(
+de-german.wav
+es-spanish.wav
+hr-croatian.wav
+po-polish.wav
+uk-ukrainian.wav
+en-english.wav
+fr-french.wav
+it-italian.wav
+ru-russian.wav
+)
+for w in ${test_wavs[@]}; do
+  time $EXE \
+    --tokens=$repo/tokens.txt \
+    --nemo-ctc-model=$repo/model.onnx \
+    --debug=1 \
+    $repo/test_wavs/$w
+done
+
+rm -rf $repo
+
+url=https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-nemo-fast-conformer-ctc-en-24500.tar.bz2
+name=$(basename $url)
+curl -SL -O $url
+tar xvf $name
+rm $name
+repo=$(basename -s .tar.bz2 $name)
+ls -lh $repo
+
+log "Test $repo"
+
+time $EXE \
+  --tokens=$repo/tokens.txt \
+  --nemo-ctc-model=$repo/model.onnx \
+  --debug=1 \
+  $repo/test_wavs/en-english.wav
+
+rm -rf $repo
+
+url=https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-nemo-fast-conformer-ctc-es-1424.tar.bz2
+name=$(basename $url)
+curl -SL -O $url
+tar xvf $name
+rm $name
+repo=$(basename -s .tar.bz2 $name)
+ls -lh $repo
+
+log "test $repo"
+
+time $EXE \
+  --tokens=$repo/tokens.txt \
+  --nemo-ctc-model=$repo/model.onnx \
+  --debug=1 \
+  $repo/test_wavs/es-spanish.wav
+
+rm -rf $repo
+
+url=https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-nemo-fast-conformer-ctc-en-de-es-fr-14288.tar.bz2
+name=$(basename $url)
+curl -SL -O $url
+tar xvf $name
+rm $name
+repo=$(basename -s .tar.bz2 $name)
+ls -lh $repo
+
+log "Test $repo"
+
+test_wavs=(
+en-english.wav
+de-german.wav
+fr-french.wav
+es-spanish.wav
+)
+
+for w in ${test_wavs[@]}; do
+  time $EXE \
+    --tokens=$repo/tokens.txt \
+    --nemo-ctc-model=$repo/model.onnx \
+    --debug=1 \
+    $repo/test_wavs/$w
+done
+
+rm -rf $repo
+
 log "------------------------------------------------------------"
 log "Run Wenet models"
 log "------------------------------------------------------------"
 wenet_models=(
 sherpa-onnx-zh-wenet-aishell
-sherpa-onnx-zh-wenet-aishell2
+# sherpa-onnx-zh-wenet-aishell2
 # sherpa-onnx-zh-wenet-wenetspeech
-sherpa-onnx-zh-wenet-multi-cn
+# sherpa-onnx-zh-wenet-multi-cn
 sherpa-onnx-en-wenet-librispeech
 # sherpa-onnx-en-wenet-gigaspeech
 )

.github/scripts/test-spoken-language-identification.sh

@@ -62,6 +62,11 @@ for wav in ${waves[@]}; do
   ls -lh *.wav
 done
 
+curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/spoken-language-identification-test-wavs.tar.bz2
+tar xvf spoken-language-identification-test-wavs.tar.bz2
+rm spoken-language-identification-test-wavs.tar.bz2
+data=spoken-language-identification-test-wavs
+
 for name in ${names[@]}; do
   log "------------------------------------------------------------"
   log "Run $name"
@@ -85,14 +90,14 @@ for name in ${names[@]}; do
     time $EXE \
       --whisper-encoder=$repo/${name}-encoder.onnx \
       --whisper-decoder=$repo/${name}-decoder.onnx \
-      $wav
+      $data/$wav
 
     log "test int8 onnx"
 
     time $EXE \
       --whisper-encoder=$repo/${name}-encoder.int8.onnx \
       --whisper-decoder=$repo/${name}-decoder.int8.onnx \
-      $wav
+      $data/$wav
   done
   rm -rf $repo
 done

.github/workflows/android.yaml

@@ -57,6 +57,7 @@ jobs:
           ./build-android-arm64-v8a.sh
           mkdir -p jniLibs/arm64-v8a/
           cp -v ./build-android-arm64-v8a/install/lib/*.so ./jniLibs/arm64-v8a/
+          rm -rf  ./build-android-arm64-v8a/
 
       - name: build android armv7-eabi
         shell: bash
@@ -65,6 +66,7 @@ jobs:
           ./build-android-armv7-eabi.sh
           mkdir -p ./jniLibs/armeabi-v7a/
           cp -v ./build-android-armv7-eabi/install/lib/*.so ./jniLibs/armeabi-v7a/
+          rm -rf ./build-android-armv7-eabi
 
       - name: build android x86_64
         shell: bash
@@ -73,6 +75,7 @@ jobs:
           ./build-android-x86-64.sh
           mkdir -p ./jniLibs/x86_64
           cp -v ./build-android-x86-64/install/lib/*.so ./jniLibs/x86_64
+          rm -rf ./build-android-x86-64
 
       - name: build android x86
         shell: bash
@@ -81,6 +84,7 @@ jobs:
           ./build-android-x86.sh
           mkdir -p ./jniLibs/x86
           cp -v ./build-android-x86/install/lib/*.so ./jniLibs/x86
+          rm -rf ./build-android-x86
 
       - name: Copy files
         shell: bash
@@ -112,6 +116,8 @@ jobs:
           command: |
             git config --global user.email "[email protected]"
             git config --global user.name "Fangjun Kuang"
+            du -h -d1 .
+            ls -lh
 
             rm -rf huggingface
             GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/csukuangfj/sherpa-onnx-libs huggingface