Merge pull request #13 from mideind/nocache

Allow audio cache size of 0 (disables caching)

.github/workflows/main.yml

@@ -15,12 +15,12 @@ jobs:
 
     strategy:
       matrix:
-        python-version: ["3.9", "3.10", "3.11", "3.12", "pypy-3.9"]
+        python-version: ["3.9", "3.12", "pypy-3.10"]
 
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python-version }}
       - name: Install dependencies
@@ -36,18 +36,18 @@ jobs:
       - name: Test with pytest
         run: |
           pytest --run-slow -vvvrP --log-level=DEBUG --capture=tee-sys
-      - name: Lint with pre-commit hooks
-        run: |
-          pre-commit run --all-files
+      # - name: Lint with pre-commit hooks
+      #   run: |
+      #     pre-commit run --all-files
 
   network:
     runs-on: ubuntu-22.04
     steps:
       - uses: actions/checkout@v4
-      - name: Set up Python 3.10
-        uses: actions/setup-python@v4
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v5
         with:
-          python-version: '3.10'
+          python-version: '3.12'
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip wheel setuptools

pyproject.toml

@@ -30,9 +30,9 @@ dependencies = [
     "typing-extensions>=4.7.1",
     "pydantic==2.4.0",
     "pydantic-settings>=2.0.3",
-    "cachetools>=5.3.1",
+    "cachetools>=5.5.0",
     # For parsing Icelandic text
-    "islenska<1.0.0",
+    "islenska==1.0.3",
     "reynir<4.0.0",
     "tokenizer<4.0.0",
     # Azure TTS

src/icespeak/settings.py

@@ -24,7 +24,6 @@
 # We dont import annotations from __future__ here
 # due to pydantic
 from typing import Any, Optional
-from typing_extensions import Literal
 
 import json
 import os
@@ -93,7 +92,9 @@ class Settings(BaseSettings):
         extra="ignore",
     )
 
-    DEFAULT_VOICE: str = Field(default="Gudrun", description="Default TTS voice if none is requested.")
+    DEFAULT_VOICE: str = Field(
+        default="Gudrun", description="Default TTS voice if none is requested."
+    )
     DEFAULT_VOICE_SPEED: float = Field(
         default=1.0,
         le=MAX_SPEED,
@@ -104,7 +105,9 @@ class Settings(BaseSettings):
         default=TextFormats.SSML,
         description="Default format to interpret input text as.",
     )
-    DEFAULT_AUDIO_FORMAT: AudioFormats = Field(default=AudioFormats.MP3, description="Default audio output format.")
+    DEFAULT_AUDIO_FORMAT: AudioFormats = Field(
+        default=AudioFormats.MP3, description="Default audio output format."
+    )
 
     AUDIO_DIR: Optional[Path] = Field(
         default=None,
@@ -113,10 +116,16 @@ class Settings(BaseSettings):
             "If not set, creates a directory in the platform's temporary directory."
         ),
     )
-    AUDIO_CACHE_SIZE: int = Field(default=300, gt=0, description="Max number of audio files to cache.")
-    AUDIO_CACHE_CLEAN: bool = Field(default=True, description="If True, cleans up generated audio files upon exit.")
+    AUDIO_CACHE_SIZE: int = Field(
+        default=300, gt=-1, description="Max number of audio files to cache."
+    )
+    AUDIO_CACHE_CLEAN: bool = Field(
+        default=True, description="If True, cleans up generated audio files upon exit."
+    )
 
-    KEYS_DIR: Path = Field(default=Path("keys"), description="Where to look for API keys.")
+    KEYS_DIR: Path = Field(
+        default=Path("keys"), description="Where to look for API keys."
+    )
     AWSPOLLY_KEY_FILENAME: str = Field(
         default="AWSPollyServerKey.json",
         description="Name of the AWS Polly API key file.",
@@ -181,26 +190,24 @@ class Keys(BaseModel):
 
     azure: Optional[AzureKey] = Field(default=None, description="Azure API key.")
     aws: Optional[AWSPollyKey] = Field(default=None, description="AWS Polly API key.")
-    google: Optional[dict[str, Any]] = Field(default=None, description="Google API key.")
-    # TODO: Re-implement TTS with Tiro
-    tiro: Literal[None] = Field(default=None)
+    google: Optional[dict[str, Any]] = Field(
+        default=None, description="Google API key."
+    )
     openai: Optional[OpenAIKey] = Field(default=None, description="OpenAI API key.")
 
     def __hash__(self):
-        return hash((self.azure, self.aws, self.google, self.tiro, self.openai))
+        return hash((self.azure, self.aws, self.google, self.openai))
 
     def __eq__(self, other: object):
         return isinstance(other, Keys) and (
             self.azure,
             self.aws,
             self.google,
-            self.tiro,
             self.openai,
         ) == (
             other.azure,
             other.aws,
             other.google,
-            other.tiro,
             other.openai,
         )
 
@@ -209,23 +216,33 @@ def __eq__(self, other: object):
 
 _kd = SETTINGS.KEYS_DIR
 if not (_kd.exists() and _kd.is_dir()):
-    _LOG.warning("Keys directory missing or incorrect, TTS will not work! Set to: %s", _kd)
+    _LOG.warning(
+        "Keys directory missing or incorrect, TTS will not work! Set to: %s", _kd
+    )
 else:
     # Load API keys, logging exceptions in level DEBUG so they aren't logged twice,
     # as exceptions are logged as warnings when voice modules are initialized
     try:
-        API_KEYS.aws = AWSPollyKey.model_validate_json((_kd / SETTINGS.AWSPOLLY_KEY_FILENAME).read_text().strip())
+        API_KEYS.aws = AWSPollyKey.model_validate_json(
+            (_kd / SETTINGS.AWSPOLLY_KEY_FILENAME).read_text().strip()
+        )
     except Exception as err:
         _LOG.debug(
             "Could not load AWS Polly API key, ASR with AWS Polly will not work. Error: %s",
             err,
         )
     try:
-        API_KEYS.azure = AzureKey.model_validate_json((_kd / SETTINGS.AZURE_KEY_FILENAME).read_text().strip())
+        API_KEYS.azure = AzureKey.model_validate_json(
+            (_kd / SETTINGS.AZURE_KEY_FILENAME).read_text().strip()
+        )
     except Exception as err:
-        _LOG.debug("Could not load Azure API key, ASR with Azure will not work. Error: %s", err)
+        _LOG.debug(
+            "Could not load Azure API key, ASR with Azure will not work. Error: %s", err
+        )
     try:
-        API_KEYS.google = json.loads((_kd / SETTINGS.GOOGLE_KEY_FILENAME).read_text().strip())
+        API_KEYS.google = json.loads(
+            (_kd / SETTINGS.GOOGLE_KEY_FILENAME).read_text().strip()
+        )
     except Exception as err:
         _LOG.debug(
             "Could not load Google API key, ASR with Google will not work. Error: %s",
@@ -236,7 +253,9 @@ def __eq__(self, other: object):
         if key := os.getenv("OPENAI_API_KEY"):
             API_KEYS.openai = OpenAIKey(api_key=SecretStr(key))
         else:
-            API_KEYS.openai = OpenAIKey.model_validate_json((_kd / SETTINGS.OPENAI_KEY_FILENAME).read_text().strip())
+            API_KEYS.openai = OpenAIKey.model_validate_json(
+                (_kd / SETTINGS.OPENAI_KEY_FILENAME).read_text().strip()
+            )
     except Exception as err:
         _LOG.debug(
             "Could not load OpenAI API key, ASR with OpenAI will not work. Error: %s",

src/icespeak/tts.py

@@ -37,9 +37,14 @@
 
 from .settings import SETTINGS, TRACE, Keys
 from .transcribe import TranscriptionOptions
-
-# TODO: Re implement Tiro
-from .voices import BaseVoice, TTSOptions, VoiceInfoT, aws_polly, azure, openai  # , google
+from .voices import (
+    BaseVoice,
+    TTSOptions,
+    VoiceInfoT,
+    aws_polly,
+    azure,
+    openai,
+)
 
 if TYPE_CHECKING:
     from pathlib import Path
@@ -60,7 +65,6 @@ def _setup_voices() -> tuple[VoicesT, ServicesT]:
         azure.AzureVoice(),
         openai.OpenAIVoice(),
         # google.GoogleVoice(),
-        # tiro.TiroVoice(),
     )
     voices: VoicesT = {}
     for service in services:
@@ -72,7 +76,8 @@ def _setup_voices() -> tuple[VoicesT, ServicesT]:
             # Info about each voice
             if voice in voices:
                 _LOG.warning(
-                    "Voice named %r already exists! " + "Skipping the one defined in module %s.",
+                    "Voice named %r already exists! "
+                    + "Skipping the one defined in module %s.",
                     voice,
                     service.name,
                 )
@@ -122,7 +127,9 @@ def _cleanup():
             audiofile.unlink(missing_ok=True)
 
     # Small daemon thread which deletes files sent to the expired queue
-    _cleanup_thread = threading.Thread(target=_cleanup, name="audio_cleanup", daemon=True)
+    _cleanup_thread = threading.Thread(
+        target=_cleanup, name="audio_cleanup", daemon=True
+    )
     _cleanup_thread.start()
 
     def _evict_all():
@@ -168,11 +175,20 @@ def tts_to_file(
     """
     if _LOG.isEnabledFor(DEBUG):
         _LOG.debug(
-            "tts_to_file, text: %r, TTS options: %s, " + "transcribe: %r, transcription options: %s",
+            "tts_to_file, text: %r, TTS options: %s, "
+            + "transcribe: %r, transcription options: %s",
             text,
-            tts_options.model_dump(exclude_defaults=True) or "<default>" if tts_options else "None",
+            (
+                tts_options.model_dump(exclude_defaults=True) or "<default>"
+                if tts_options
+                else "None"
+            ),
             transcribe,
-            transcription_options.model_dump(exclude_defaults=True) or "<default>" if transcription_options else "None",
+            (
+                transcription_options.model_dump(exclude_defaults=True) or "<default>"
+                if transcription_options
+                else "None"
+            ),
         )
     tts_options = tts_options or TTSOptions()
     try:
@@ -181,7 +197,9 @@ def tts_to_file(
         raise ValueError(f"Voice {tts_options.voice!r} not available.") from e
 
     if tts_options.audio_format not in service.audio_formats:
-        raise ValueError(f"Service {service.name} doesn't support audio format {tts_options.audio_format}.")
+        raise ValueError(
+            f"Service {service.name} doesn't support audio format {tts_options.audio_format}."
+        )
 
     if transcribe:
         transcription_options = transcription_options or TranscriptionOptions()

tests/test_tts.py

@@ -125,20 +125,6 @@ def test_Google_speech_synthesis():
     path.unlink()
 
 
-@pytest.mark.skipif(condition=True, reason="Missing Tiro API Key.")
-@pytest.mark.network
-def test_Tiro_speech_synthesis():
-    # Test Tiro
-    tts_out = tts_to_file(
-        _TEXT,
-        TTSOptions(text_format=TextFormats.TEXT, audio_format="mp3", voice="Alfur"),
-    )
-    path = tts_out.file
-    assert path.is_file(), "Expected audio file to exist"
-    assert path.stat().st_size > _MIN_AUDIO_SIZE, "Expected longer audio data"
-    path.unlink()
-
-
 @pytest.mark.skipif(API_KEYS.openai is None, reason="Missing OpenAI API Key.")
 @pytest.mark.network
 def test_OpenAI_speech_synthesis():