[FEATURE] Allow interleaving text and attachments in message

[Antsiscool]

Scope check

• This is core LLM communication (not application logic)
• This benefits most users (not just my use case)
• This can't be solved in application code with current RubyLLM
• I read the Contributing Guide

Due diligence

• I searched existing issues
• I checked the documentation

What problem does this solve?

Currently when a message contains both text and attachments, the text is at the beginning of the message followed by all of the attachments.

Allowing interleaving would mean we could place text closer to the correct attachment.

Proposed solution

RubyLLM::Content has a text and attachments attribute.

When formatting for a message, it first adds the text attribute and then loops through all of the attachments.

This could be changed to have a single attribute which contains both the text and attachments as an ordered array.

Why this belongs in RubyLLM

There is no way to bypass the current implementation of RubyLLM::Content and interleave text and attachments.

[crmne]

Is this supported by the LLM APIs?

[tpaulshippy]

Yes

Just as with document-query placement, Claude works best when images come before text. Images placed after text or interpolated with text will still perform well, but if your use case allows it, we recommend an image-then-text structure.

https://docs.claude.com/en/docs/build-with-claude/vision