LangChain4j

This will lead to this when using OpenAI: > dev.ai4j.openai4j.OpenAiHttpException: { > "error": { > "message": "Invalid parameter: messages with role 'tool' must be a response to a preceeding message...

Can use HtmlTextExtractor under the hood

[BUG] functions and function_call have been deprecated in Azure OpenAI. Now calling through functions or function_call will report an error.

3

  https://learn.microsoft.com/en-au/azure/ai-services/openai/how-to/function-calling?tabs=python#using-function-in-the-chat-completions-api-deprecated

[FEATURE] Make overlap splitter configurable in HierarchicalDocumentSplitter

1

Currently, when calculating overlap, we are first splitting previous segment by **sentences**, then try to fit as many sentences as possible into the overlap. One should be able to configure...

[FEATURE] Ensure Image and ImageContent can be created from all possible sources (url, local file, base64 content) only in a correct state

2

Make sure created `Image` (from all possible sources) can be used as inputs in all places: - `*ChatModel` (wrapped as `ImageContent`) - `*ImageModel` - etc Make sure mime type is...

a way to merge multiple in-memory embedding stores

1

## Summary by CodeRabbit - **New Features** - Introduced enhanced flexibility for managing entries within the `InMemoryEmbeddingStore`. - **Tests** - Added a test to verify the functionality of merging multiple...

OpenAI: OpenAiTokenizer now has a default constructor which should be suitable for all current models

1

## Summary by CodeRabbit - **New Features** - Updated the `OpenAiTokenizer` to support initialization with specific OpenAI model names, enhancing flexibility in model selection.

Anthropic: added AnthropicStreamingChatModel + support for native images

1

## Summary by CodeRabbit - **New Features** - Introduced streaming functionality for creating messages in the Anthropic API. - Added a new Anthropic language model variant `CLAUDE_3_HAIKU_20240307`. - Implemented a...