Use piper-phonemize to convert text to token IDs (#453)

cmake/cmake_extension.py

@@ -73,6 +73,10 @@ class BuildExtension(build_ext):
 
         extra_cmake_args = f" -DCMAKE_INSTALL_PREFIX={install_dir} "
         extra_cmake_args += " -DBUILD_SHARED_LIBS=ON "
+        extra_cmake_args += " -DBUILD_PIPER_PHONMIZE_EXE=OFF "
+        extra_cmake_args += " -DBUILD_PIPER_PHONMIZE_TESTS=OFF "
+        extra_cmake_args += " -DBUILD_ESPEAK_NG_EXE=OFF "
+        extra_cmake_args += " -DBUILD_ESPEAK_NG_TESTS=OFF "
 
         extra_cmake_args += " -DSHERPA_ONNX_ENABLE_CHECK=OFF "
         extra_cmake_args += " -DSHERPA_ONNX_ENABLE_PYTHON=ON "
@@ -146,6 +150,9 @@ class BuildExtension(build_ext):
             binaries += ["sherpa-onnx-core.dll"]
             binaries += ["sherpa-onnx-portaudio.dll"]
             binaries += ["onnxruntime.dll"]
+            binaries += ["piper_phonemize.dll"]
+            binaries += ["espeak-ng.dll"]
+            binaries += ["ucd.dll"]
             binaries += ["kaldi-decoder-core.dll"]
             binaries += ["sherpa-onnx-fst.lib"]
             binaries += ["sherpa-onnx-kaldifst-core.lib"]
@@ -161,5 +168,8 @@ class BuildExtension(build_ext):
             shutil.copy(f"{src_file}", f"{out_bin_dir}/")
 
         shutil.rmtree(f"{install_dir}/bin")
+        shutil.rmtree(f"{install_dir}/share")
+        shutil.rmtree(f"{install_dir}/lib/pkgconfig")
+
         if is_windows():
             shutil.rmtree(f"{install_dir}/lib")

cmake/espeak-ng-for-piper.cmake

@@ -86,7 +86,7 @@ function(download_espeak_ng_for_piper)
       -Wno-unused-result
       -Wno-format-overflow
       -Wno-format-truncation
-      -Wno-maybe-uninitialized
+      -Wno-uninitialized
       -Wno-format
     )
 

cmake/sherpa-onnx.pc.in

@@ -13,4 +13,4 @@ Cflags: -I"${includedir}"
 # Note: -lcargs is required only for the following file
 # https://github.com/k2-fsa/sherpa-onnx/blob/master/c-api-examples/decode-file-c-api.c
 # We add it here so that users don't need to specify -lcargs when compiling decode-file-c-api.c
-Libs: -L"${libdir}" -lsherpa-onnx-c-api -lsherpa-onnx-core -lonnxruntime -lkaldi-decoder-core -lsherpa-onnx-kaldifst-core -lsherpa-onnx-fst -lkaldi-native-fbank-core -lcargs -Wl,-rpath,${libdir} @SHERPA_ONNX_PKG_CONFIG_EXTRA_LIBS@
+Libs: -L"${libdir}" -lsherpa-onnx-c-api -lsherpa-onnx-core -lonnxruntime -lkaldi-decoder-core -lsherpa-onnx-kaldifst-core -lsherpa-onnx-fst -lkaldi-native-fbank-core -lpiper_phonemize -lespeak-ng -lucd -lcargs -Wl,-rpath,${libdir} @SHERPA_ONNX_PKG_CONFIG_EXTRA_LIBS@