[component] AI Models(LLMs)

[monikakuster]

As a Workflow Developer I want to be able to use the native connector for OpenAI, Hugging Face, Azure OpenAI, Anthropic LLMs So that I can implement more complex integration use cases where the connector is involved.

Check Spring AI:

• https://docs.spring.io/spring-ai/reference/index.html
• https://spring.io/blog/2024/05/09/spring-ai-structured-output
• https://spring.io/blog/2024/08/09/spring-ai-embraces-openais-structured-outputs-enhancing-json-response
• https://www.baeldung.com/spring-artificial-intelligence-structure-output

Acceptance Criteria

List of implemented components:

• [x] OpenAI
• [x] Hugging Face
• [x] Azure OpenAI
• [x] Anthropic 3
• [x] Amazon Bedrock
• [x] Google Vertex AI
• [x] Groq
• [x] QianFan - remove
• [x] ZhiPu AI - remove
• [x] Watsonx.AI
• [x] Mistral AI
• [x] MiniMax - remove
• [x] Moonshot AI - remove
• [x] Ollama
• [x] NVIDIA
• [x] Stability
• [x] #2273
• [x] #2274
• [x] MistralAI - OCR
• [ ] Amazon Bedrock upgrade https://docs.spring.io/spring-ai/reference/upgrade-notes.html#_removal_of_deprecated_amazon_bedrock_chat_models

Definition of Done:

• [x] Implementation
• [ ] Unit tests
• [ ] Integration tests
• [ ] Documentation

Technical details:

• The components should be built as Spring Components - check ScheduleComponentHandler
• Each LLM should be implemented as a separate connector
• All connectors should have common actions
• If possible, they should share common infrastructure
• They should have additional prompts to define structured output
• Current ChatGPT connector should be replaced with this new approach