v1.0.3

Added

• Customizable Analysis Response: Introduced a major new feature allowing clients to define their own data structure for the analysis result.
  • Added a new generic method, newCustomAudioReviewer<T>(), which takes a client-defined data class (Class<T>) and returns a VoiceReviewer<T>.
  • Introduced the @FieldDescription annotation, which allows clients to describe the fields in their custom data class. These descriptions are used to dynamically generate a precise prompt for the AI, ensuring an accurate and structured JSON response.

Changed

• The core AnalysisEngine and VoiceReviewer interfaces are now generic (AnalysisEngine<T>, VoiceReviewer<T>) to support the new customizable analysis feature.
• The VoiceReviewState.Success class is now also generic (Success<T>) and will contain the client-defined data model upon a successful analysis.

Full Changelog: 1.0.2...1.0.3

v1.0.1.04

Fixed

• Resolved a critical reflection crash (KotlinReflectionInternalError) in release builds by integrating the KSP annotation processor and removing default arguments from the GptAnalysisResponse data class. This ensures compatibility with code shrinkers like ProGuard.

Changed

• Reduced excessive logging in the OpenAIAnalysisEngine to provide a cleaner and more focused output.

Full Changelog: 1.0.1.03...1.0.1.04

v1.0.1

Added

• Comprehensive logging within the OpenAIAnalysisEngine to provide deeper insights into the transcription and analysis process.

Fixed

• Resolved a critical bug causing JSON parsing failures from the OpenAI service by refining the internal GptAnalysisResponse data model to handle field mismatches gracefully.

Changed

• Improved error propagation within the analysis engine by wrapping failures in a specific AnalysisFailedException for clearer debugging context.

Full Changelog: 1.0.0...1.0.1

VeView Initial Release

• Initial release of the VeView SDK.
• Core functionality for recording voice reviews and managing the lifecycle through a reactive StateFlow.
• Integration with OpenAI for AI-powered audio transcription.

Full Changelog: https://github.com/SunnyBe/VeView/commits/0.0.2