wyoming-microsoft-tts 1.3.4__tar.gz → 1.4.0__tar.gz

{wyoming_microsoft_tts-1.3.4 → wyoming_microsoft_tts-1.4.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wyoming-microsoft-tts
-Version: 1.3.4
+Version: 1.4.0
 Summary: Add your description here
 Home-page: https://github.com/hugobloem/wyoming-microsoft-tts
 Author: Hugo Bloem

{wyoming_microsoft_tts-1.3.4 → wyoming_microsoft_tts-1.4.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "wyoming-microsoft-tts"
-version = "1.3.4"
+version = "1.4.0"
 description = "Add your description here"
 readme = "README.md"
 requires-python = ">=3.13"

wyoming_microsoft_tts-1.4.0/requirements.txt

@@ -0,0 +1,7 @@
+wyoming==1.8.0
+azure-cognitiveservices-speech==1.47.0
+black>=24,<27
+lxml>=5,<7
+pycountry>=23.12.11
+ruff
+regex

{wyoming_microsoft_tts-1.3.4 → wyoming_microsoft_tts-1.4.0}/setup.py

@@ -20,7 +20,7 @@ data_files = [module_dir / "voices.json"]
 
 setup(
     name="wyoming_microsoft_tts",
-    version="1.3.4",
+    version="1.4.0",
     description="Wyoming Server for Microsoft TTS",
     url="https://github.com/hugobloem/wyoming-microsoft-tts",
     author="Hugo Bloem",

{wyoming_microsoft_tts-1.3.4 → wyoming_microsoft_tts-1.4.0}/tests/conftest.py

@@ -21,6 +21,11 @@ def microsoft_tts(configuration):
         subscription_key=os.environ.get("SPEECH_KEY"),
         service_region=os.environ.get("SPEECH_REGION"),
         download_dir="/tmp/",
+        rate=None,
+        pitch=None,
+        volume=None,
+        style=None,
+        style_degree=None,
         **configuration,
     )
     return MicrosoftTTS(args)

wyoming_microsoft_tts-1.4.0/tests/test_microsoft_tts.py

@@ -0,0 +1,352 @@
+"""Tests for the MicrosoftTTS class."""
+
+from types import SimpleNamespace
+import os
+import pytest
+from wyoming_microsoft_tts.microsoft_tts import MicrosoftTTS
+
+
+def test_initialize(microsoft_tts, configuration):
+    """Test initialization."""
+    assert microsoft_tts.args.voice == configuration["voice"]
+    assert microsoft_tts.speech_config is not None
+    assert microsoft_tts.output_dir is not None
+
+
+@pytest.mark.skipif(
+    not os.environ.get("SPEECH_KEY") or not os.environ.get("SPEECH_REGION"),
+    reason="SPEECH_KEY and SPEECH_REGION environment variables required",
+)
+def test_synthesize(microsoft_tts):
+    """Test synthesize."""
+    text = "Hello, world!"
+    voice = "en-US-JennyNeural"
+
+    result = microsoft_tts.synthesize(text, voice)
+    assert result.endswith(".wav")
+
+
+# SSML Building Tests
+
+
+def test_build_ssml_with_rate():
+    """Test SSML generation with rate parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate="+30%",
+        pitch=None,
+        volume=None,
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("Hello, world!", "en-US-JennyNeural")
+
+    assert '<?xml version="1.0" encoding="UTF-8"?>' in ssml
+    assert '<speak version="1.0"' in ssml
+    assert '<prosody rate="+30%">' in ssml
+    assert "</prosody>" in ssml
+    assert "Hello, world!" in ssml
+    assert "xmlns:mstts" not in ssml  # No style, so no mstts namespace
+
+
+def test_build_ssml_with_pitch():
+    """Test SSML generation with pitch parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch="+10%",
+        volume=None,
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("Testing pitch", "en-US-JennyNeural")
+
+    assert '<prosody pitch="+10%">' in ssml
+    assert "</prosody>" in ssml
+    assert "Testing pitch" in ssml
+
+
+def test_build_ssml_with_volume():
+    """Test SSML generation with volume parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch=None,
+        volume="loud",
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("Volume test", "en-US-JennyNeural")
+
+    assert '<prosody volume="loud">' in ssml
+    assert "</prosody>" in ssml
+    assert "Volume test" in ssml
+
+
+def test_build_ssml_with_all_prosody():
+    """Test SSML generation with all prosody parameters."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate="fast",
+        pitch="high",
+        volume="+20%",
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("All prosody", "en-US-JennyNeural")
+
+    assert '<prosody rate="fast" pitch="high" volume="+20%">' in ssml
+    assert "</prosody>" in ssml
+    assert "All prosody" in ssml
+
+
+def test_build_ssml_with_style():
+    """Test SSML generation with style parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch=None,
+        volume=None,
+        style="cheerful",
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("Style test", "en-US-JennyNeural")
+
+    assert 'xmlns:mstts="https://www.w3.org/2001/mstts"' in ssml
+    assert '<mstts:express-as style="cheerful">' in ssml
+    assert "</mstts:express-as>" in ssml
+    assert "Style test" in ssml
+
+
+def test_build_ssml_with_style_and_degree():
+    """Test SSML generation with style and style_degree parameters."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch=None,
+        volume=None,
+        style="sad",
+        style_degree=1.5,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("Sad voice", "en-US-JennyNeural")
+
+    assert 'xmlns:mstts="https://www.w3.org/2001/mstts"' in ssml
+    assert '<mstts:express-as style="sad" styledegree="1.5">' in ssml
+    assert "</mstts:express-as>" in ssml
+    assert "Sad voice" in ssml
+
+
+def test_build_ssml_with_prosody_and_style():
+    """Test SSML generation with both prosody and style parameters."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate="slow",
+        pitch="low",
+        volume="soft",
+        style="calm",
+        style_degree=0.5,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("Combined test", "en-US-JennyNeural")
+
+    assert 'xmlns:mstts="https://www.w3.org/2001/mstts"' in ssml
+    assert '<mstts:express-as style="calm" styledegree="0.5">' in ssml
+    assert '<prosody rate="slow" pitch="low" volume="soft">' in ssml
+    assert "</prosody>" in ssml
+    assert "</mstts:express-as>" in ssml
+    assert "Combined test" in ssml
+
+
+def test_build_ssml_voice_key_and_lang():
+    """Test that SSML uses correct voice key and language."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-GB-SoniaNeural",
+        rate="+10%",
+        pitch=None,
+        volume=None,
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    ssml = tts._build_ssml("UK voice", "en-GB-SoniaNeural")
+
+    # Should contain the voice key from the voices.json
+    assert 'xml:lang="en-GB"' in ssml
+    assert '<voice name="en-GB-SoniaNeural">' in ssml
+
+
+# Integration Tests with Synthesize
+
+
+@pytest.mark.skipif(
+    not os.environ.get("SPEECH_KEY") or not os.environ.get("SPEECH_REGION"),
+    reason="SPEECH_KEY and SPEECH_REGION environment variables required",
+)
+def test_synthesize_with_rate():
+    """Test synthesize with rate parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate="+30%",
+        pitch=None,
+        volume=None,
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    result = tts.synthesize("Testing rate parameter", "en-US-JennyNeural")
+
+    assert result is not None
+    assert result.endswith(".wav")
+
+
+@pytest.mark.skipif(
+    not os.environ.get("SPEECH_KEY") or not os.environ.get("SPEECH_REGION"),
+    reason="SPEECH_KEY and SPEECH_REGION environment variables required",
+)
+def test_synthesize_with_pitch():
+    """Test synthesize with pitch parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch="+5%",
+        volume=None,
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    result = tts.synthesize("Testing pitch parameter", "en-US-JennyNeural")
+
+    assert result is not None
+    assert result.endswith(".wav")
+
+
+@pytest.mark.skipif(
+    not os.environ.get("SPEECH_KEY") or not os.environ.get("SPEECH_REGION"),
+    reason="SPEECH_KEY and SPEECH_REGION environment variables required",
+)
+def test_synthesize_with_volume():
+    """Test synthesize with volume parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch=None,
+        volume="loud",
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    result = tts.synthesize("Testing volume parameter", "en-US-JennyNeural")
+
+    assert result is not None
+    assert result.endswith(".wav")
+
+
+@pytest.mark.skipif(
+    not os.environ.get("SPEECH_KEY") or not os.environ.get("SPEECH_REGION"),
+    reason="SPEECH_KEY and SPEECH_REGION environment variables required",
+)
+def test_synthesize_with_style():
+    """Test synthesize with style parameter."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch=None,
+        volume=None,
+        style="cheerful",
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    result = tts.synthesize("Testing style parameter", "en-US-JennyNeural")
+
+    assert result is not None
+    assert result.endswith(".wav")
+
+
+@pytest.mark.skipif(
+    not os.environ.get("SPEECH_KEY") or not os.environ.get("SPEECH_REGION"),
+    reason="SPEECH_KEY and SPEECH_REGION environment variables required",
+)
+def test_synthesize_with_combined_parameters():
+    """Test synthesize with multiple parameters combined."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate="fast",
+        pitch="+10%",
+        volume="loud",
+        style="excited",
+        style_degree=1.2,
+    )
+    tts = MicrosoftTTS(args)
+    result = tts.synthesize("Testing all parameters together", "en-US-JennyNeural")
+
+    assert result is not None
+    assert result.endswith(".wav")
+
+
+@pytest.mark.skipif(
+    not os.environ.get("SPEECH_KEY") or not os.environ.get("SPEECH_REGION"),
+    reason="SPEECH_KEY and SPEECH_REGION environment variables required",
+)
+def test_synthesize_without_parameters_still_works():
+    """Test that synthesize still works without any new parameters."""
+    args = SimpleNamespace(
+        subscription_key=os.environ.get("SPEECH_KEY"),
+        service_region=os.environ.get("SPEECH_REGION"),
+        download_dir="/tmp/",
+        voice="en-US-JennyNeural",
+        rate=None,
+        pitch=None,
+        volume=None,
+        style=None,
+        style_degree=None,
+    )
+    tts = MicrosoftTTS(args)
+    result = tts.synthesize("Testing without parameters", "en-US-JennyNeural")
+
+    assert result is not None
+    assert result.endswith(".wav")

{wyoming_microsoft_tts-1.3.4 → wyoming_microsoft_tts-1.4.0}/wyoming_microsoft_tts/__main__.py

@@ -68,6 +68,32 @@ def parse_arguments():
     )
     parser.add_argument("--samples-per-chunk", type=int, default=1024)
     #
+    parser.add_argument(
+        "--rate",
+        type=str,
+        help="Speech rate (e.g., '+30%', '0.5', 'fast', 'slow')",
+    )
+    parser.add_argument(
+        "--pitch",
+        type=str,
+        help="Speech pitch (e.g., '+10%', 'high', 'low', '+80Hz')",
+    )
+    parser.add_argument(
+        "--volume",
+        type=str,
+        help="Speech volume (e.g., '+20%', 'loud', 'soft', '75')",
+    )
+    parser.add_argument(
+        "--style",
+        type=str,
+        help="Speaking style (e.g., 'cheerful', 'sad', 'angry', 'calm')",
+    )
+    parser.add_argument(
+        "--style-degree",
+        type=float,
+        help="Style intensity from 0.01 to 2 (default: 1)",
+    )
+    #
     parser.add_argument(
         "--update-voices",
         action="store_true",

wyoming_microsoft_tts-1.4.0/wyoming_microsoft_tts/microsoft_tts.py

@@ -0,0 +1,115 @@
+"""Microsoft TTS."""
+
+import logging
+import tempfile
+import time
+from pathlib import Path
+
+import azure.cognitiveservices.speech as speechsdk
+
+from .download import get_voices
+
+_LOGGER = logging.getLogger(__name__)
+
+
+class MicrosoftTTS:
+    """Class to handle Microsoft TTS."""
+
+    def __init__(self, args) -> None:
+        """Initialize."""
+        _LOGGER.debug("Initialize Microsoft TTS")
+        self.args = args
+        self.speech_config = speechsdk.SpeechConfig(
+            subscription=args.subscription_key, region=args.service_region
+        )
+
+        output_dir = str(tempfile.TemporaryDirectory())
+        output_dir = Path(output_dir)
+        output_dir.mkdir(parents=True, exist_ok=True)
+        self.output_dir = output_dir
+
+        self.voices = get_voices(args.download_dir)
+
+    def _build_ssml(self, text, voice):
+        """Build SSML with prosody and style parameters."""
+        voice_key = self.voices[voice]["key"]
+        voice_lang = self.voices[voice]["language"]["code"]
+
+        ssml_parts = [
+            '<?xml version="1.0" encoding="UTF-8"?>',
+            '<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis"',
+        ]
+
+        if self.args.style or self.args.style_degree:
+            ssml_parts.append(' xmlns:mstts="https://www.w3.org/2001/mstts"')
+
+        ssml_parts.append(f' xml:lang="{voice_lang}">')
+        ssml_parts.append(f'<voice name="{voice_key}">')
+
+        has_style = self.args.style is not None
+        has_prosody = any([self.args.rate, self.args.pitch, self.args.volume])
+
+        if has_style:
+            style_attrs = [f'style="{self.args.style}"']
+            if self.args.style_degree is not None:
+                style_attrs.append(f'styledegree="{self.args.style_degree}"')
+            ssml_parts.append(f'<mstts:express-as {" ".join(style_attrs)}>')
+
+        if has_prosody:
+            prosody_attrs = []
+            if self.args.rate:
+                prosody_attrs.append(f'rate="{self.args.rate}"')
+            if self.args.pitch:
+                prosody_attrs.append(f'pitch="{self.args.pitch}"')
+            if self.args.volume:
+                prosody_attrs.append(f'volume="{self.args.volume}"')
+            ssml_parts.append(f'<prosody {" ".join(prosody_attrs)}>')
+
+        ssml_parts.append(text)
+
+        if has_prosody:
+            ssml_parts.append('</prosody>')
+
+        if has_style:
+            ssml_parts.append('</mstts:express-as>')
+
+        ssml_parts.append('</voice>')
+        ssml_parts.append('</speak>')
+
+        return ''.join(ssml_parts)
+
+    def synthesize(self, text, voice=None):
+        """Synthesize text to speech."""
+        _LOGGER.debug(f"Requested TTS for [{text}]")
+        if voice is None:
+            voice = self.args.voice
+
+        # Convert the requested voice to the key microsoft use.
+        self.speech_config.speech_synthesis_voice_name = self.voices[voice]["key"]
+
+        file_name = self.output_dir / f"{time.monotonic_ns()}.wav"
+        audio_config = speechsdk.audio.AudioOutputConfig(filename=str(file_name))
+
+        speech_synthesizer = speechsdk.SpeechSynthesizer(
+            speech_config=self.speech_config, audio_config=audio_config
+        )
+
+        if any([self.args.rate, self.args.pitch, self.args.volume, self.args.style, self.args.style_degree]):
+            ssml = self._build_ssml(text, voice)
+            _LOGGER.debug(f"Using SSML: {ssml}")
+            speech_synthesis_result = speech_synthesizer.speak_ssml_async(ssml).get()
+        else:
+            speech_synthesis_result = speech_synthesizer.speak_text_async(text).get()
+
+        if (
+            speech_synthesis_result.reason
+            == speechsdk.ResultReason.SynthesizingAudioCompleted
+        ):
+            _LOGGER.debug(f"Speech synthesized for text [{text}]")
+            return str(file_name)
+
+        elif speech_synthesis_result.reason == speechsdk.ResultReason.Canceled:
+            cancellation_details = speech_synthesis_result.cancellation_details
+            _LOGGER.warning(f"Speech synthesis canceled: {cancellation_details.reason}")
+            if cancellation_details.reason == speechsdk.CancellationReason.Error:
+                _LOGGER.warning(f"Error details: {cancellation_details.error_details}")

{wyoming_microsoft_tts-1.3.4 → wyoming_microsoft_tts-1.4.0}/wyoming_microsoft_tts/version.py

@@ -1,3 +1,3 @@
 """Version information."""
 
-__version__ = "1.3.4"
+__version__ = "1.4.0"

{wyoming_microsoft_tts-1.3.4 → wyoming_microsoft_tts-1.4.0}/wyoming_microsoft_tts.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wyoming-microsoft-tts
-Version: 1.3.4
+Version: 1.4.0
 Summary: Add your description here
 Home-page: https://github.com/hugobloem/wyoming-microsoft-tts
 Author: Hugo Bloem

wyoming_microsoft_tts-1.3.4/requirements.txt

@@ -1,7 +0,0 @@
-wyoming==1.7.2
-azure-cognitiveservices-speech==1.46.0
-black>=24,<26
-lxml>=5,<7
-pycountry>=23.12.11
-ruff
-regex

wyoming_microsoft_tts-1.3.4/tests/test_microsoft_tts.py

@@ -1,17 +0,0 @@
-"""Tests for the MicrosoftTTS class."""
-
-
-def test_initialize(microsoft_tts, configuration):
-    """Test initialization."""
-    assert microsoft_tts.args.voice == configuration["voice"]
-    assert microsoft_tts.speech_config is not None
-    assert microsoft_tts.output_dir is not None
-
-
-def test_synthesize(microsoft_tts):
-    """Test synthesize."""
-    text = "Hello, world!"
-    voice = "en-US-JennyNeural"
-
-    result = microsoft_tts.synthesize(text, voice)
-    assert result.endswith(".wav")

wyoming_microsoft_tts-1.3.4/wyoming_microsoft_tts/microsoft_tts.py

@@ -1,62 +0,0 @@
-"""Microsoft TTS."""
-
-import logging
-import tempfile
-import time
-from pathlib import Path
-
-import azure.cognitiveservices.speech as speechsdk
-
-from .download import get_voices
-
-_LOGGER = logging.getLogger(__name__)
-
-
-class MicrosoftTTS:
-    """Class to handle Microsoft TTS."""
-
-    def __init__(self, args) -> None:
-        """Initialize."""
-        _LOGGER.debug("Initialize Microsoft TTS")
-        self.args = args
-        self.speech_config = speechsdk.SpeechConfig(
-            subscription=args.subscription_key, region=args.service_region
-        )
-
-        output_dir = str(tempfile.TemporaryDirectory())
-        output_dir = Path(output_dir)
-        output_dir.mkdir(parents=True, exist_ok=True)
-        self.output_dir = output_dir
-
-        self.voices = get_voices(args.download_dir)
-
-    def synthesize(self, text, voice=None):
-        """Synthesize text to speech."""
-        _LOGGER.debug(f"Requested TTS for [{text}]")
-        if voice is None:
-            voice = self.args.voice
-
-        # Convert the requested voice to the key microsoft use.
-        self.speech_config.speech_synthesis_voice_name = self.voices[voice]["key"]
-
-        file_name = self.output_dir / f"{time.monotonic_ns()}.wav"
-        audio_config = speechsdk.audio.AudioOutputConfig(filename=str(file_name))
-
-        speech_synthesizer = speechsdk.SpeechSynthesizer(
-            speech_config=self.speech_config, audio_config=audio_config
-        )
-
-        speech_synthesis_result = speech_synthesizer.speak_text_async(text).get()
-
-        if (
-            speech_synthesis_result.reason
-            == speechsdk.ResultReason.SynthesizingAudioCompleted
-        ):
-            _LOGGER.debug(f"Speech synthesized for text [{text}]")
-            return str(file_name)
-
-        elif speech_synthesis_result.reason == speechsdk.ResultReason.Canceled:
-            cancellation_details = speech_synthesis_result.cancellation_details
-            _LOGGER.warning(f"Speech synthesis canceled: {cancellation_details.reason}")
-            if cancellation_details.reason == speechsdk.CancellationReason.Error:
-                _LOGGER.warning(f"Error details: {cancellation_details.error_details}")