Add phonemes for Czech

rhasspy · Aug 3, 2023 · 41cfb05 · 41cfb05
1 parent 7f7b5bd
commit 41cfb05
Show file tree

Hide file tree

Showing 5 changed files with 31 additions and 9 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -4,7 +4,7 @@ set(CMAKE_VERBOSE_MAKEFILE off)
 
 project(
     piper_phonemize
-    VERSION 1.1.0
+    VERSION 1.2.0
     DESCRIPTION "Phonemization library for Piper text to speech system"
     HOMEPAGE_URL "https://github.com/rhasspy/piper-phonemize"
     LANGUAGES CXX

diff --git a/Dockerfile b/Dockerfile
@@ -40,8 +40,9 @@ RUN curl -L "${ONNXRUNTIME_URL}" | \
 
 # Build minimal version of espeak-ng
 RUN curl -L "https://github.com/rhasspy/espeak-ng/archive/refs/heads/master.tar.gz" | \
-    tar -xzvf - && \
-    cd espeak-ng-master && \
+    tar -xzf -
+
+RUN cd espeak-ng-master && \
     export CFLAGS='-D_FILE_OFFSET_BITS=64' && \
     ./autogen.sh && \
     ./configure \
@@ -51,11 +52,26 @@ RUN curl -L "https://github.com/rhasspy/espeak-ng/archive/refs/heads/master.tar.
         --without-mbrola \
         --without-sonic \
         --with-extdict-cmn \
+        --with-extdict-ru \
         --prefix=/usr && \
     make -j8 src/espeak-ng src/speak-ng && \
     make && \
     make install
 
+# For future, not currently working
+# cmake -Bbuild \
+#     -DCMAKE_INSTALL_PREFIX=/usr \
+#     -DCMAKE_BUILD_TYPE=Release \
+#     -DBUILD_SHARED_LIBS=ON \
+#     -DUSE_MBROLA=OFF \
+#     -DUSE_LIBSONIC=OFF \
+#     -DUSE_LIBPCAUDIO=OFF \
+#     -DUSE_KLATT=OFF \
+#     -DUSE_SPEECHPLAYER=OFF \
+#     -DUSE_ASYNC=OFF \
+#     -DEXTRA_cmn=ON \
+#     -DEXTRA_ru=ON && \
+
 # Build libpiper_phonemize.so
 COPY etc/libtashkeel_model.ort ./etc/
 COPY CMakeLists.txt Makefile ./

diff --git a/Makefile b/Makefile
@@ -2,6 +2,7 @@
 
 LIB_DIR := lib/Linux-$(shell uname -m)
 DOCKER_PLATFORM ?= linux/amd64,linux/arm64,linux/arm/v7
+VENV ?= .venv
 
 release:
 	mkdir -p build
@@ -13,20 +14,20 @@ release:
 
 test:
 	g++ -Wall -o test -Iespeak-ng/build/include -Lespeak-ng/build/lib -I$(LIB_DIR)/onnxruntime/include -L$(LIB_DIR)/onnxruntime/lib -Isrc -std=c++17 src/test.cpp src/phonemize.cpp src/phoneme_ids.cpp src/tashkeel.cpp -lespeak-ng -lonnxruntime
-	LD_LIBRARY_PATH="espeak-ng/build/lib:$(LIB_DIR)/onnxruntime/lib" ./test 'espeak-ng/build/share/espeak-ng-data'
+	LD_LIBRARY_PATH="espeak-ng/build/lib:$(LIB_DIR)/onnxruntime/lib:${LD_LIBRARY_PATH}" ./test 'espeak-ng/build/share/espeak-ng-data'
 
 python:
 	cp -R espeak-ng/build/share/espeak-ng-data piper_phonemize/
 	cp etc/libtashkeel_model.ort piper_phonemize/
-	LD_LIBRARY_PATH='espeak-ng/build/lib' .venv/bin/pip3 install -e .
+	LD_LIBRARY_PATH="espeak-ng/build/lib:${LD_LIBRARY_PATH}" "$(VENV)/bin/pip3" install -e .
 
 python-test:
-	LD_LIBRARY_PATH="espeak-ng/build/lib:$(LIB_DIR)/onnxruntime/lib" .venv/bin/python3 src/python_test.py
+	LD_LIBRARY_PATH="espeak-ng/build/lib:$(LIB_DIR)/onnxruntime/lib:${LD_LIBRARY_PATH}" "$(VENV)/bin/python3" src/python_test.py
 
 python-wheel:
 	cp -R espeak-ng/build/share/espeak-ng-data piper_phonemize/
 	cp etc/libtashkeel_model.ort piper_phonemize/
-	LD_LIBRARY_PATH='espeak-ng/build/lib' .venv/bin/python3 setup.py bdist_wheel
+	LD_LIBRARY_PATH="espeak-ng/build/lib:${LD_LIBRARY_PATH}" "$(VENV)/bin/python3" setup.py bdist_wheel
 
 docker:
 	docker buildx build . --platform "$(DOCKER_PLATFORM)" --output 'type=local,dest=dist'
diff --git a/setup.py b/setup.py
@@ -10,7 +10,7 @@
 _LIB_DIR = _DIR / "lib" / f"Linux-{platform.machine()}"
 _ONNXRUNTIME_DIR = _LIB_DIR / "onnxruntime"
 
-__version__ = "1.1.0"
+__version__ = "1.2.0"
 
 ext_modules = [
     Pybind11Extension(

diff --git a/src/phoneme_ids.cpp b/src/phoneme_ids.cpp
@@ -171,7 +171,12 @@ PhonemeIdMap DEFAULT_PHONEME_ID_MAP = {
     // Luxembourgish
     {U'g', {154}},
     {U'ʦ', {155}},
-    {U'X', {156}}};
+    {U'X', {156}},
+
+    // Czech
+    {U'\u031d', {157}},
+    {U'\u030a', {158}},
+};
 
 // language -> phoneme -> [id, ...]
 std::map<std::string, PhonemeIdMap> DEFAULT_ALPHABET = {