Skip to content

TESTING.md

Latest commit

 

History

History
148 lines (101 loc) · 4.28 KB

TESTING.md

File metadata and controls

148 lines (101 loc) · 4.28 KB

Testing and Examples

This document explains how to test the TTS wrapper and run examples for different TTS engines.

Prerequisites

Before running tests or examples, make sure you have set up the necessary environment variables for the TTS engines you want to test:

Azure TTS

MICROSOFT_TOKEN=your_azure_subscription_key
MICROSOFT_REGION=your_azure_region

ElevenLabs TTS

ELEVENLABS_API_KEY=your_elevenlabs_api_key

Google TTS

GOOGLE_APPLICATION_CREDENTIALS=path_to_your_google_credentials_json
# or
GOOGLE_SA_PATH=path_to_your_google_service_account_json

CereVoice TTS

CEREVOICE_EMAIL=your_cerevoice_email
CEREVOICE_PASSWORD=your_cerevoice_password
# Optional
CEREVOICE_ACCESS_TOKEN=existing_access_token
CEREVOICE_REFRESH_TOKEN=existing_refresh_token

You can set these environment variables in a .env file in the root of the project.

Running Tests

The project includes a unified test framework that can test all TTS engines or specific engines.

Credential Validation

Before running tests, the framework automatically validates your credentials for each TTS engine using the checkCredentials() method. This ensures that only engines with valid credentials are tested.

Testing All Engines

To test all available TTS engines:

npm run test:tts

This will run tests for all engines for which you have provided valid credentials.

Testing Specific Engines

To test a specific TTS engine:

# Test Azure TTS
npm run test:azure

# Test ElevenLabs TTS
npm run test:elevenlabs

# Test Google TTS
npm run test:google

# Test CereVoice unit coverage
npm run test:cerevoice

If your credentials for the specified engine are invalid, the tests will be skipped with a clear message.

npm run test:cerevoice runs mocked unit coverage for auth, voice mapping, synthesis request shape, metadata fetching, and word-boundary conversion. It does not require live CereVoice credentials.

Running Examples

The project includes a unified example framework that can demonstrate all TTS engines or specific engines.

Credential Validation

Before running examples, the framework automatically validates your credentials for each TTS engine using the checkCredentials() method. This ensures that only engines with valid credentials are demonstrated.

Running Examples for All Engines

To run examples for all available TTS engines:

npm run example

This will run examples for all engines for which you have provided valid credentials.

Running Examples for Specific Engines

To run an example for a specific TTS engine:

# Run Azure TTS example
npm run example:azure

# Run ElevenLabs TTS example
npm run example:elevenlabs

# Run Google TTS example
npm run example:google

If your credentials for the specified engine are invalid, the example will be skipped with a clear message.

Adding a New TTS Engine

To add a new TTS engine to the unified test and example framework:

  1. Create a new engine implementation in src/engines/
  2. Implement the _getVoices() method to retrieve voices from the provider
  3. Implement the _mapVoicesToUnified() method to map the raw voice data to a unified format
  4. Implement the checkCredentials() method to validate credentials (or use the default implementation)
  5. Update the factory function in src/__tests__/tts-engine.test.ts to include your new engine
  6. Update the factory function in examples/tts-example.js to include your new engine
  7. Add the new engine to the validEngines array in run-tts-tests.js
  8. Add new npm scripts in package.json for testing and running examples with your engine

Test Coverage

The tests cover the following functionality for each TTS engine:

  • Listing available voices
  • Getting voices by language
  • Setting and getting properties
  • Synthesizing text using non-streaming approach
  • Synthesizing SSML to speech (for engines that support SSML)
  • Synthesizing text using streaming approach
  • Handling word boundary events

Example Functionality

The examples demonstrate the following functionality for each TTS engine:

  • Listing available voices
  • Getting voices by language
  • Setting a voice
  • Converting text to speech
  • Converting SSML to speech (for engines that support SSML)
  • Converting text to speech using streaming
  • Handling word boundary events