Refactor WebAssembly build script. (#598)

Make it easier to build WebAssembly for ASR.
2024-02-21 16:51:15 +08:00
parent 25079b5c05
commit 7c4b59932a
12 changed files with 189 additions and 75 deletions
--- a/.github/workflows/wasm-simd-hf-space-de-tts.yaml
+++ b/.github/workflows/wasm-simd-hf-space-de-tts.yaml
@@ -0,0 +1,106 @@
+name: wasm-simd-hf-space-de-tts
+
+on:
+  release:
+    types:
+      - published
+
+  workflow_dispatch:
+
+concurrency:
+  group: wasm-simd-hf-space-de-tts-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  wasm-simd-hf-space-de-tts:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest]
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - name: Install emsdk
+        uses: mymindstorm/setup-emsdk@v14
+
+      - name: View emsdk version
+        shell: bash
+        run: |
+          emcc -v
+          echo "--------------------"
+          emcc --check
+
+      - name: Download model files
+        shell: bash
+        run: |
+          cd wasm/tts/assets
+          ls -lh
+          echo "----------"
+          wget -q https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-piper-de_DE-thorsten_emotional-medium.tar.bz2
+          tar xf vits-piper-de_DE-thorsten_emotional-medium.tar.bz2
+          rm vits-piper-de_DE-thorsten_emotional-medium.tar.bz2
+
+          mv -v vits-piper-de_DE-thorsten_emotional-medium/de_DE-thorsten_emotional-medium.onnx ./model.onnx
+          mv -v vits-piper-de_DE-thorsten_emotional-medium/tokens.txt ./
+          mv vits-piper-de_DE-thorsten_emotional-medium/espeak-ng-data ./
+
+          rm -rf vits-piper-de_DE-thorsten_emotional-medium
+
+          ls -lh
+
+      - name: Build sherpa-onnx for WebAssembly
+        shell: bash
+        run: |
+          ./build-wasm-simd-tts.sh
+
+      - name: collect files
+        shell: bash
+        run: |
+          SHERPA_ONNX_VERSION=v$(grep "SHERPA_ONNX_VERSION" ./CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)
+
+          mv build-wasm-simd-tts/install/bin/wasm sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-de-tts
+          ls -lh sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-de-tts
+          tar cjfv sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-de-tts.tar.bz2 ./sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-de-tts
+
+      - name: Upload wasm files
+        uses: actions/upload-artifact@v4
+        with:
+          name: sherpa-onnx-wasm-simd-de-tts
+          path: ./sherpa-onnx-wasm-simd-*.tar.bz2
+
+      - name: Publish to huggingface
+        env:
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
+        uses: nick-fields/retry@v2
+        with:
+          max_attempts: 20
+          timeout_seconds: 200
+          shell: bash
+          command: |
+            SHERPA_ONNX_VERSION=v$(grep "SHERPA_ONNX_VERSION" ./CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)
+
+            git config --global user.email "csukuangfj@gmail.com"
+            git config --global user.name "Fangjun Kuang"
+
+            rm -rf huggingface
+            export GIT_LFS_SKIP_SMUDGE=1
+
+            git clone https://huggingface.co/spaces/k2-fsa/web-assembly-tts-sherpa-onnx-de huggingface
+            cd huggingface
+            git fetch
+            git pull
+            git merge -m "merge remote" --ff origin main
+
+            cp -v ../sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-de-tts/* .
+
+            git status
+            git lfs track "*.data"
+            git lfs track "*.wasm"
+            ls -lh
+
+            git add .
+            git commit -m "update model"
+            git push https://csukuangfj:$HF_TOKEN@huggingface.co/spaces/k2-fsa/web-assembly-tts-sherpa-onnx-de main
--- a/.github/workflows/wasm-simd-hf-space-en-tts.yaml
+++ b/.github/workflows/wasm-simd-hf-space-en-tts.yaml
@@ -1,9 +1,6 @@
-name: wasm-simd-hf-space-en
+name: wasm-simd-hf-space-en-tts

 on:
-  push:
-    branches:
-      - wasm-2
  release:
    types:
      - published
@@ -11,11 +8,11 @@ on:
  workflow_dispatch:

 concurrency:
-  group: wasm-simd-hf-space-en-${{ github.ref }}
+  group: wasm-simd-hf-space-en-tts-${{ github.ref }}
  cancel-in-progress: true

 jobs:
-  wasm-simd-hf-space-en:
+  wasm-simd-hf-space-en-tts:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
@@ -39,7 +36,7 @@ jobs:
      - name: Download model files
        shell: bash
        run: |
-          cd wasm/assets
+          cd wasm/tts/assets
          ls -lh
          echo "----------"
          wget -q https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-piper-en_US-libritts_r-medium.tar.bz2
@@ -55,21 +52,21 @@ jobs:
      - name: Build sherpa-onnx for WebAssembly
        shell: bash
        run: |
-          ./build-wasm-simd.sh
+          ./build-wasm-simd-tts.sh

      - name: collect files
        shell: bash
        run: |
          SHERPA_ONNX_VERSION=v$(grep "SHERPA_ONNX_VERSION" ./CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)

-          mv build-wasm-simd/install/bin/wasm sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en
-          ls -lh sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en
-          tar cjfv sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en.tar.bz2 ./sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en
+          mv build-wasm-simd-tts/install/bin/wasm sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en-tts
+          ls -lh sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en-tts
+          tar cjfv sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en-tts.tar.bz2 ./sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en-tts

      - name: Upload wasm files
        uses: actions/upload-artifact@v4
        with:
-          name: sherpa-onnx-wasm-simd-en
+          name: sherpa-onnx-wasm-simd-en-tts
          path: ./sherpa-onnx-wasm-simd-*.tar.bz2

      - name: Publish to huggingface
@@ -95,7 +92,7 @@ jobs:
            git pull
            git merge -m "merge remote" --ff origin main

-            cp -v ../sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en/* .
+            cp -v ../sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-en-tts/* .

            git status
            git lfs track "*.data"
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -21,6 +21,7 @@ option(SHERPA_ONNX_ENABLE_C_API "Whether to build C API" ON)
 option(SHERPA_ONNX_ENABLE_WEBSOCKET "Whether to build webscoket server/client" ON)
 option(SHERPA_ONNX_ENABLE_GPU "Enable ONNX Runtime GPU support" OFF)
 option(SHERPA_ONNX_ENABLE_WASM "Whether to enable WASM" OFF)
+option(SHERPA_ONNX_ENABLE_WASM_TTS "Whether to enable WASM for TTS" OFF)
 option(SHERPA_ONNX_ENABLE_BINARY "Whether to build binaries" ON)
 option(SHERPA_ONNX_LINK_LIBSTDCPP_STATICALLY "True to link libstdc++ statically. Used only when BUILD_SHARED_LIBS is OFF on Linux" ON)

@@ -104,6 +105,14 @@ message(STATUS "SHERPA_ONNX_ENABLE_C_API ${SHERPA_ONNX_ENABLE_C_API}")
 message(STATUS "SHERPA_ONNX_ENABLE_WEBSOCKET ${SHERPA_ONNX_ENABLE_WEBSOCKET}")
 message(STATUS "SHERPA_ONNX_ENABLE_GPU ${SHERPA_ONNX_ENABLE_GPU}")
 message(STATUS "SHERPA_ONNX_ENABLE_WASM ${SHERPA_ONNX_ENABLE_WASM}")
+message(STATUS "SHERPA_ONNX_ENABLE_WASM_TTS ${SHERPA_ONNX_ENABLE_WASM_TTS}")
+
+if(SHERPA_ONNX_ENABLE_WASM_TTS)
+  if(NOT SHERPA_ONNX_ENABLE_WASM)
+    message(FATAL_ERROR "Please set SHERPA_ONNX_ENABLE_WASM to ON if you enable WASM for tts")
+  endif()
+endif()
+
 if(SHERPA_ONNX_ENABLE_WASM)
  add_definitions(-DSHERPA_ONNX_ENABLE_WASM=1)
 endif()
@@ -114,7 +123,6 @@ endif()
 set(CMAKE_CXX_EXTENSIONS OFF)
 message(STATUS "C++ Standard version: ${CMAKE_CXX_STANDARD}")

-
 include(CheckIncludeFileCXX)

 if(UNIX AND NOT APPLE AND NOT SHERPA_ONNX_ENABLE_WASM AND NOT CMAKE_SYSTEM_NAME STREQUAL Android)
--- a/build-wasm-simd-tts.sh
+++ b/build-wasm-simd-tts.sh
@@ -31,8 +31,8 @@ if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
  exit 1
 fi

-mkdir -p build-wasm-simd
-pushd build-wasm-simd
+mkdir -p build-wasm-simd-tts
+pushd build-wasm-simd-tts

 export SHERPA_ONNX_IS_USING_BUILD_WASM_SH=ON

@@ -51,6 +51,7 @@ cmake \
  -DSHERPA_ONNX_ENABLE_WEBSOCKET=OFF \
  -DSHERPA_ONNX_ENABLE_GPU=OFF \
  -DSHERPA_ONNX_ENABLE_WASM=ON \
+  -DSHERPA_ONNX_ENABLE_WASM_TTS=ON \
  -DSHERPA_ONNX_ENABLE_BINARY=OFF \
  -DSHERPA_ONNX_LINK_LIBSTDCPP_STATICALLY=OFF \
  ..
--- a/wasm/CMakeLists.txt
+++ b/wasm/CMakeLists.txt
@@ -1,59 +1,3 @@
-if(NOT $ENV{SHERPA_ONNX_IS_USING_BUILD_WASM_SH})
-  message(FATAL_ERROR "Please use ./build-wasm.sh to build for wasm")
+if(SHERPA_ONNX_ENABLE_WASM_TTS)
+  add_subdirectory(tts)
 endif()
-
-if(NOT EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/assets/decoder_jit_trace-pnnx.ncnn.bin")
-  message(WARNING "${CMAKE_CURRENT_SOURCE_DIR}/assets/decoder_jit_trace-pnnx.ncnn.bin does not exist")
-  # message(FATAL_ERROR "Please read ${CMAKE_CURRENT_SOURCE_DIR}/assets/README.md before you continue")
-endif()
-
-set(exported_functions
-  MyPrint
-  SherpaOnnxCreateOfflineTts
-  SherpaOnnxDestroyOfflineTts
-  SherpaOnnxDestroyOfflineTtsGeneratedAudio
-  SherpaOnnxOfflineTtsGenerate
-  SherpaOnnxOfflineTtsGenerateWithCallback
-  SherpaOnnxOfflineTtsNumSpeakers
-  SherpaOnnxOfflineTtsSampleRate
-  SherpaOnnxWriteWave
-)
-set(mangled_exported_functions)
-foreach(x IN LISTS exported_functions)
-  list(APPEND mangled_exported_functions "_${x}")
-endforeach()
-list(JOIN mangled_exported_functions "," all_exported_functions)
-
-
-include_directories(${CMAKE_SOURCE_DIR})
-set(MY_FLAGS " -s FORCE_FILESYSTEM=1 -s INITIAL_MEMORY=512MB -s ALLOW_MEMORY_GROWTH=1")
-string(APPEND MY_FLAGS " -sSTACK_SIZE=10485760 ") # 10MB
-string(APPEND MY_FLAGS " -sEXPORTED_FUNCTIONS=[_CopyHeap,_malloc,_free,${all_exported_functions}] ")
-string(APPEND MY_FLAGS "--preload-file ${CMAKE_CURRENT_SOURCE_DIR}/assets@. ")
-string(APPEND MY_FLAGS " -sEXPORTED_RUNTIME_METHODS=['ccall','stringToUTF8','setValue','getValue'] ")
-
-set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${MY_FLAGS}")
-set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${MY_FLAGS}")
-set(CMAKE_EXECUTBLE_LINKER_FLAGS "${CMAKE_EXECUTBLE_LINKER_FLAGS} ${MY_FLAGS}")
-
-if (NOT CMAKE_EXECUTABLE_SUFFIX STREQUAL ".js")
-  message(FATAL_ERROR "The default suffix for building executables should be .js!")
-endif()
-# set(CMAKE_EXECUTABLE_SUFFIX ".html")
-
-add_executable(sherpa-onnx-wasm-main sherpa-onnx-wasm-main.cc)
-target_link_libraries(sherpa-onnx-wasm-main sherpa-onnx-c-api)
-install(TARGETS sherpa-onnx-wasm-main DESTINATION bin/wasm)
-
-install(
-  FILES
-    "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.js"
-    "index.html"
-    "sherpa-onnx.js"
-    "app.js"
-    "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.wasm"
-    "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.data"
-    # "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.html"
-  DESTINATION
-    bin/wasm
-)
--- a/wasm/tts/CMakeLists.txt
+++ b/wasm/tts/CMakeLists.txt
@@ -0,0 +1,58 @@
+if(NOT $ENV{SHERPA_ONNX_IS_USING_BUILD_WASM_SH})
+  message(FATAL_ERROR "Please use ./build-wasm-simd-tts.sh to build for wasm TTS")
+endif()
+
+if(NOT EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/assets/model.onnx")
+  message(FATAL_ERROR "Please read ${CMAKE_CURRENT_SOURCE_DIR}/assets/README.md before you continue")
+endif()
+
+set(exported_functions
+  MyPrint
+  SherpaOnnxCreateOfflineTts
+  SherpaOnnxDestroyOfflineTts
+  SherpaOnnxDestroyOfflineTtsGeneratedAudio
+  SherpaOnnxOfflineTtsGenerate
+  SherpaOnnxOfflineTtsGenerateWithCallback
+  SherpaOnnxOfflineTtsNumSpeakers
+  SherpaOnnxOfflineTtsSampleRate
+  SherpaOnnxWriteWave
+)
+set(mangled_exported_functions)
+foreach(x IN LISTS exported_functions)
+  list(APPEND mangled_exported_functions "_${x}")
+endforeach()
+list(JOIN mangled_exported_functions "," all_exported_functions)
+
+
+include_directories(${CMAKE_SOURCE_DIR})
+set(MY_FLAGS " -s FORCE_FILESYSTEM=1 -s INITIAL_MEMORY=512MB -s ALLOW_MEMORY_GROWTH=1")
+string(APPEND MY_FLAGS " -sSTACK_SIZE=10485760 ") # 10MB
+string(APPEND MY_FLAGS " -sEXPORTED_FUNCTIONS=[_CopyHeap,_malloc,_free,${all_exported_functions}] ")
+string(APPEND MY_FLAGS "--preload-file ${CMAKE_CURRENT_SOURCE_DIR}/assets@. ")
+string(APPEND MY_FLAGS " -sEXPORTED_RUNTIME_METHODS=['ccall','stringToUTF8','setValue','getValue'] ")
+
+set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${MY_FLAGS}")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${MY_FLAGS}")
+set(CMAKE_EXECUTBLE_LINKER_FLAGS "${CMAKE_EXECUTBLE_LINKER_FLAGS} ${MY_FLAGS}")
+
+if (NOT CMAKE_EXECUTABLE_SUFFIX STREQUAL ".js")
+  message(FATAL_ERROR "The default suffix for building executables should be .js!")
+endif()
+# set(CMAKE_EXECUTABLE_SUFFIX ".html")
+
+add_executable(sherpa-onnx-wasm-main sherpa-onnx-wasm-main.cc)
+target_link_libraries(sherpa-onnx-wasm-main sherpa-onnx-c-api)
+install(TARGETS sherpa-onnx-wasm-main DESTINATION bin/wasm)
+
+install(
+  FILES
+    "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.js"
+    "index.html"
+    "sherpa-onnx.js"
+    "app.js"
+    "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.wasm"
+    "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.data"
+    # "$<TARGET_FILE_DIR:sherpa-onnx-wasm-main>/sherpa-onnx-wasm-main.html"
+  DESTINATION
+    bin/wasm
+)
--- a/wasm/tts/app.js
+++ b/wasm/tts/app.js
--- a/wasm/tts/assets/.gitignore
+++ b/wasm/tts/assets/.gitignore
--- a/wasm/tts/assets/README.md
+++ b/wasm/tts/assets/README.md
@@ -18,7 +18,7 @@ rm -rf vits-piper-en_US-libritts_r-medium
 ```

 You should have the following files in `assets` before you can run
-`build-wasm-simd.sh`
+`build-wasm-simd-tts.sh`

 ```
 assets fangjun$ tree -L 1
--- a/wasm/tts/index.html
+++ b/wasm/tts/index.html
--- a/wasm/tts/sherpa-onnx-wasm-main.cc
+++ b/wasm/tts/sherpa-onnx-wasm-main.cc
--- a/wasm/tts/sherpa-onnx.js
+++ b/wasm/tts/sherpa-onnx.js