Release 0.4.0
Until version 1.0 (at least) library API is subject to potentially incompatible changes, while the code gets more stable.
However, functionalities from older versions should always be available, with minor code adjustments.
- Refactoring:
Endpointhas been renamedAiEndpoint, newEndpointinterface is now just a marker. - Refactoring:
LlmUtilis nowChunkUtil. Methods added to support chunking with moving window and overlapping text. EmbeddingServicenow uses the above methods to embed text in a more flexible way.- Web Search:
SearchEndpointadded. CreatedGoogleSearchServiceto provide online search capabilities. - Added Text to Speech OpenAI API.
- Added support for multiple function calls (tools) in OpenAI chat completion API.
- OpenAI image service now supports DALL-E 3 and uses it by default.
OpenAiChatServiceuses GPT-4 8k by default.- Fixed broken OpenAI audio API (
whisper-1model). - Added new categories for OpenAI moderation API.
- Added "slot filling" capabilities to
CompletionServiceandChatService. - Robustness: Added automatic "throttling" when API call rate limit is exceeded (automatic handling of HTTP errors 429, 500 & 503).
- Robustness: Added automatic adjustment of completion size for OpenAI completion services when context size is exceeded.
- Added extensive example (
EssayWriter). - Bug fixes.