Use message type during message creation of Mistral AI chat model

[nicolaskrier]

Description

Providers like Azure OpenAI already support flexible conversion of Spring AI messages to provider-specific formats, enabling custom message types.
This PR adds the same capability for Mistral AI, introducing a dedicated conversion method to ensure consistency across the project.

Key Changes

1. New Conversion Method

   • Added createChatCompletionMessages in MistralAiChatModel to map Spring AI Message objects to Mistral AI’s ChatCompletionMessage stream, based on MessageType.
   • Follows the established pattern in Azure OpenAI’s implementation.

2. Unit Tests

   • Added comprehensive tests for the new conversion method, covering:
     • Standard Spring AI messages.
     • Custom message types.
   • Improved existing tests by:
     • Removing redundant @SpringBootTest annotation.
     • Addressing minor code smells.
     • Aligning test class name with project conventions.

3. Backward Compatibility

   • No breaking changes: All existing unit and integration tests pass.

Related Work

This PR aligns with #3999, which implements a similar conversion mechanism for Ollama.
The goal is to standardize message handling across all LLM providers in Spring AI.

[ilayaperumalg]

@ricken07 Could you help with the review please? Thanks!

[ricken07]

@ilayaperumalg All right, no problem

[nicolaskrier]

Thanks @ilayaperumalg and @ricken07 for your reviews!

I have updated the PR description to add more context and clarify the intent behind these changes.

[nicolaskrier]

Hi @ilayaperumalg and @ricken07,

I wanted to apologize for the initial state of this PR. It clearly needed more work. I’ve incorporated as much of your review feedback as possible, and I believe the remaining points are now ready for discussion, especially with the updated PR description.
I’d love to hear your thoughts on both the current implementation and my expectations regarding the described behavior. Does this align with your vision for this feature?

Thanks for your time and guidance!

[nicolaskrier]

Hi @ricken07 and @ilayaperumalg, I’ve addressed all the feedback from the last review. The changes are now ready for another look. Please let me know if everything looks good or if there’s anything else to adjust. Thanks again for your time!

[ilayaperumalg]

@nicolaskrier Thanks for the PR and following up with the review comments. @ricken07 Thanks for reviewing the PR. LGTM, rebased, squashed and merged as f9d87fd