Fix C# to support Chinese tts models using jieba (#815)

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

@@ -47,6 +47,6 @@ jobs:
         env:
           API_KEY: ${{ secrets.NUGET_API_KEY }}
         run: |
-          # API_KEY is valid until 2024.05.02
+          # API_KEY is valid until 2025.04.26
           cd /tmp/packages
           dotnet nuget push ./org.k2fsa.sherpa.onnx.*.nupkg --skip-duplicate --api-key $API_KEY --source https://api.nuget.org/v3/index.json

@@ -28,6 +28,9 @@ class OfflineTtsPlayDemo
     [Option("tts-rule-fsts", Required = false, Default = "", HelpText = "path to rule.fst")]
     public string RuleFsts { get; set; }
 
+    [Option("vits-dict-dir", Required = false, Default = "", HelpText = "Path to the directory containing dict for jieba.")]
+    public string DictDir { get; set; }
+
     [Option("vits-data-dir", Required = false, Default = "", HelpText = "Path to the directory containing dict for espeak-ng.")]
     public string DataDir { get; set; }
 
@@ -129,6 +132,7 @@ to download more models.
     config.Model.Vits.Lexicon = options.Lexicon;
     config.Model.Vits.Tokens = options.Tokens;
     config.Model.Vits.DataDir = options.DataDir;
+    config.Model.Vits.DictDir = options.DictDir;
     config.Model.Vits.NoiseScale = options.NoiseScale;
     config.Model.Vits.NoiseScaleW = options.NoiseScaleW;
     config.Model.Vits.LengthScale = options.LengthScale;

+#!/usr/bin/env bash
+set -ex
+if [ ! -f ./vits-zh-hf-fanchen-C/vits-zh-hf-fanchen-C.onnx ]; then
+  curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-zh-hf-fanchen-C.tar.bz2
+  tar xf vits-zh-hf-fanchen-C.tar.bz2
+  rm vits-zh-hf-fanchen-C.tar.bz2
+fi
+
+dotnet run \
+  --vits-model=./vits-zh-hf-fanchen-C/vits-zh-hf-fanchen-C.onnx \
+  --vits-tokens=./vits-zh-hf-fanchen-C/tokens.txt \
+  --vits-lexicon=./vits-zh-hf-fanchen-C/lexicon.txt \
+  --tts-rule-fsts=./vits-zh-hf-fanchen-C/phone.fst,./vits-zh-hf-fanchen-C/date.fst,./vits-zh-hf-fanchen-C/number.fst \
+  --vits-dict-dir=./vits-zh-hf-fanchen-C/dict \
+  --sid=100 \
+  --debug=1 \
+  --output-filename=./fanchen-100.wav \
+  --text="这是一个语音合成测试, 写于公元2024年4月26号, 11点05分，星期5。小米的使命是，始终坚持做'感动人心、价格厚道'的好产品，让全球每个人都能享受科技带来的美好生活。"

@@ -2,7 +2,6 @@
 
 set -ex
 if [ ! -f ./vits-piper-en_US-amy-low/en_US-amy-low.onnx ]; then
-  # wget -qq https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-piper-en_US-amy-low.tar.bz2
   curl -OL https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-piper-en_US-amy-low.tar.bz2
   tar xf vits-piper-en_US-amy-low.tar.bz2
   rm vits-piper-en_US-amy-low.tar.bz2

@@ -23,6 +23,9 @@ class OfflineTtsDemo
     [Option("tts-rule-fars", Required = false, Default = "", HelpText = "path to rule.far")]
     public string RuleFars { get; set; }
 
+    [Option("vits-dict-dir", Required = false, Default = "", HelpText = "Path to the directory containing dict for jieba.")]
+    public string DictDir { get; set; }
+
     [Option("vits-data-dir", Required = false, Default = "", HelpText = "Path to the directory containing dict for espeak-ng.")]
     public string DataDir { get; set; }
 
@@ -124,6 +127,7 @@ to download more models.
     config.Model.Vits.Lexicon = options.Lexicon;
     config.Model.Vits.Tokens = options.Tokens;
     config.Model.Vits.DataDir = options.DataDir;
+    config.Model.Vits.DictDir = options.DictDir;
     config.Model.Vits.NoiseScale = options.NoiseScale;
     config.Model.Vits.NoiseScaleW = options.NoiseScaleW;
     config.Model.Vits.LengthScale = options.LengthScale;

+#!/usr/bin/env bash
+set -ex
+if [ ! -f ./vits-zh-hf-fanchen-C/vits-zh-hf-fanchen-C.onnx ]; then
+  curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-zh-hf-fanchen-C.tar.bz2
+  tar xf vits-zh-hf-fanchen-C.tar.bz2
+  rm vits-zh-hf-fanchen-C.tar.bz2
+fi
+
+dotnet run \
+  --vits-model=./vits-zh-hf-fanchen-C/vits-zh-hf-fanchen-C.onnx \
+  --vits-tokens=./vits-zh-hf-fanchen-C/tokens.txt \
+  --vits-lexicon=./vits-zh-hf-fanchen-C/lexicon.txt \
+  --tts-rule-fsts=./vits-zh-hf-fanchen-C/phone.fst,./vits-zh-hf-fanchen-C/date.fst,./vits-zh-hf-fanchen-C/number.fst \
+  --vits-dict-dir=./vits-zh-hf-fanchen-C/dict \
+  --sid=100 \
+  --debug=1 \
+  --output-filename=./fanchen-100.wav \
+  --text="这是一个语音合成测试, 写于公元2024年4月26号, 11点05分，星期5。小米的使命是，始终坚持做'感动人心、价格厚道'的好产品，让全球每个人都能享受科技带来的美好生活。"

@@ -2,6 +2,7 @@
 # Copyright (c)  2023  Xiaomi Corporation
 
 import argparse
+import os
 import re
 from pathlib import Path
 
@@ -9,6 +10,8 @@ import jinja2
 
 SHERPA_ONNX_DIR = Path(__file__).resolve().parent.parent.parent
 
+src_dir = os.environ.get("src_dir", "/tmp")
+
 
 def get_version():
     cmake_file = SHERPA_ONNX_DIR / "CMakeLists.txt"
@@ -45,7 +48,7 @@ def process_linux(s):
         "libsherpa-onnx-kaldifst-core.so",
         "libucd.so",
     ]
-    prefix = "/tmp/linux/"
+    prefix = f"{src_dir}/linux/"
     libs = [prefix + lib for lib in libs]
     libs = "\n      ;".join(libs)
 
@@ -74,7 +77,7 @@ def process_macos(s):
         "libsherpa-onnx-kaldifst-core.dylib",
         "libucd.dylib",
     ]
-    prefix = f"/tmp/macos/"
+    prefix = f"{src_dir}/macos/"
     libs = [prefix + lib for lib in libs]
     libs = "\n      ;".join(libs)
 
@@ -106,7 +109,7 @@ def process_windows(s, rid):
 
     version = get_version()
 
-    prefix = f"/tmp/windows-{rid}/"
+    prefix = f"{src_dir}/windows-{rid}/"
     libs = [prefix + lib for lib in libs]
     libs = "\n      ;".join(libs)
 

@@ -10,23 +10,36 @@ echo "SHERPA_ONNX_DIR: $SHERPA_ONNX_DIR"
 
 SHERPA_ONNX_VERSION=$(grep "SHERPA_ONNX_VERSION" $SHERPA_ONNX_DIR/CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)
 
-# HF_MIRROR=hf-mirror.com
-HF_MIRROR=hf.co
+# You can pre-download the required wheels to $src_dir
+
+if [ $(hostname) == fangjuns-MacBook-Pro.local ]; then
+  HF_MIRROR=hf-mirror.com
+  src_dir=/Users/fangjun/open-source/sherpa-onnx/scripts/dotnet/tmp
+else
+  src_dir=/tmp
+  HF_MIRROR=hf.co
+fi
+export src_dir
 
-mkdir -p /tmp/
-pushd /tmp
+mkdir -p $src_dir
+pushd $src_dir
 
 mkdir -p linux macos windows-x64 windows-x86
 
-# You can pre-download the required wheels to /tmp
-src_dir=/tmp
 
-linux_wheel=$src_dir/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
-macos_wheel=$src_dir/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-macosx_11_0_x86_64.whl
-windows_x64_wheel=$src_dir/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win_amd64.whl
-windows_x86_wheel=$src_dir/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win32.whl
+linux_wheel_filename=sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
+linux_wheel=$src_dir/$linux_wheel_filename
+
+macos_wheel_filename=sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-macosx_11_0_x86_64.whl
+macos_wheel=$src_dir/$macos_wheel_filename
+
+windows_x64_wheel_filename=sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win_amd64.whl
+windows_x64_wheel=$src_dir/$windows_x64_wheel_filename
+
+windows_x86_wheel_filename=sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win32.whl
+windows_x86_wheel=$src_dir/$windows_x86_wheel_filename
 
-if [ ! -f /tmp/linux/libsherpa-onnx-core.so ]; then
+if [ ! -f $src_dir/linux/libsherpa-onnx-core.so ]; then
   echo "---linux x86_64---"
   cd linux
   mkdir -p wheel
@@ -34,9 +47,9 @@ if [ ! -f /tmp/linux/libsherpa-onnx-core.so ]; then
   if [ -f $linux_wheel ]; then
     cp -v $linux_wheel .
   else
-    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
+    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/$linux_wheel_filename
   fi
-  unzip sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
+  unzip $linux_wheel_filename
   cp -v sherpa_onnx/lib/*.so* ../
   cd ..
   rm -v libpiper_phonemize.so libpiper_phonemize.so.1.2.0
@@ -49,7 +62,7 @@ if [ ! -f /tmp/linux/libsherpa-onnx-core.so ]; then
   cd ..
 fi
 
-if [ ! -f /tmp/macos/libsherpa-onnx-core.dylib ]; then
+if [ ! -f $src_dir/macos/libsherpa-onnx-core.dylib ]; then
   echo "---macOS x86_64---"
   cd macos
   mkdir -p wheel
@@ -57,9 +70,9 @@ if [ ! -f /tmp/macos/libsherpa-onnx-core.dylib ]; then
   if [ -f $macos_wheel  ]; then
     cp -v $macos_wheel .
   else
-    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-macosx_11_0_x86_64.whl
+    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/$macos_wheel_filename
   fi
-  unzip sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-macosx_11_0_x86_64.whl
+  unzip $macos_wheel_filename
   cp -v sherpa_onnx/lib/*.dylib ../
 
   cd ..
@@ -75,7 +88,7 @@ if [ ! -f /tmp/macos/libsherpa-onnx-core.dylib ]; then
 fi
 
 
-if [ ! -f /tmp/windows-x64/sherpa-onnx-core.dll ]; then
+if [ ! -f $src_dir/windows-x64/sherpa-onnx-core.dll ]; then
   echo "---windows x64---"
   cd windows-x64
   mkdir -p wheel
@@ -83,9 +96,9 @@ if [ ! -f /tmp/windows-x64/sherpa-onnx-core.dll ]; then
   if [ -f $windows_x64_wheel ]; then
     cp -v $windows_x64_wheel .
   else
-    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win_amd64.whl
+    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/$windows_x64_wheel_filename
   fi
-  unzip sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win_amd64.whl
+  unzip $windows_x64_wheel_filename
   cp -v sherpa_onnx-${SHERPA_ONNX_VERSION}.data/data/bin/*.dll ../
   cp -v sherpa_onnx-${SHERPA_ONNX_VERSION}.data/data/bin/*.lib ../
   cd ..
@@ -95,7 +108,7 @@ if [ ! -f /tmp/windows-x64/sherpa-onnx-core.dll ]; then
   cd ..
 fi
 
-if [ ! -f /tmp/windows-x86/sherpa-onnx-core.dll ]; then
+if [ ! -f $src_dir/windows-x86/sherpa-onnx-core.dll ]; then
   echo "---windows x86---"
   cd windows-x86
   mkdir -p wheel
@@ -103,9 +116,9 @@ if [ ! -f /tmp/windows-x86/sherpa-onnx-core.dll ]; then
   if [ -f $windows_x86_wheel ]; then
     cp -v $windows_x86_wheel .
   else
-    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win32.whl
+    curl -OL https://$HF_MIRROR/csukuangfj/sherpa-onnx-wheels/resolve/main/$windows_x86_wheel_filename
   fi
-  unzip sherpa_onnx-${SHERPA_ONNX_VERSION}-cp38-cp38-win32.whl
+  unzip $windows_x86_wheel_filename
   cp -v sherpa_onnx-${SHERPA_ONNX_VERSION}.data/data/bin/*.dll ../
   cp -v sherpa_onnx-${SHERPA_ONNX_VERSION}.data/data/bin/*.lib ../
   cd ..

@@ -38,29 +38,29 @@ static void Handler(int32_t sig) {
   fprintf(stderr, "\nCaught Ctrl + C. Exiting...\n");
 }
 
-static std::string tolowerUnicode(const std::string& input_str) {
-    
-    // Use system locale
-    std::setlocale(LC_ALL, ""); 
-    
-    // From char string to wchar string
-    std::wstring input_wstr(input_str.size()+1, '\0');
-    std::mbstowcs(&input_wstr[0], input_str.c_str(), input_str.size()); 
-    std::wstring lowercase_wstr;
-
-    for (wchar_t wc : input_wstr) {
-        if (std::iswupper(wc)) {
-            lowercase_wstr += std::towlower(wc);
-        } else {
-            lowercase_wstr += wc;
-        }
+static std::string tolowerUnicode(const std::string &input_str) {
+  // Use system locale
+  std::setlocale(LC_ALL, "");
+
+  // From char string to wchar string
+  std::wstring input_wstr(input_str.size() + 1, '\0');
+  std::mbstowcs(&input_wstr[0], input_str.c_str(), input_str.size());
+  std::wstring lowercase_wstr;
+
+  for (wchar_t wc : input_wstr) {
+    if (std::iswupper(wc)) {
+      lowercase_wstr += std::towlower(wc);
+    } else {
+      lowercase_wstr += wc;
     }
-    
-    // Back to char string
-    std::string lowercase_str(input_str.size()+1, '\0');
-    std:wcstombs(&lowercase_str[0], lowercase_wstr.c_str(), lowercase_wstr.size());
+  }
+
+  // Back to char string
+  std::string lowercase_str(input_str.size() + 1, '\0');
+  std::wcstombs(&lowercase_str[0], lowercase_wstr.c_str(),
+                lowercase_wstr.size());
 
-    return lowercase_str;
+  return lowercase_str;
 }
 
 int32_t main(int32_t argc, char *argv[]) {