feat: Extend chat template preprocessing to support multi-modality content blocks

[guygir]

This PR extends chat template preprocessing to support structured multi-modality content blocks (OpenAI API format).
This is the first stage of multi-modality support - only basic technical feasibility. This PR focuses solely on images - not audio or video, because GAIE already supports images but audio/video support requires additional GAIE changes (These will be addressed in the next stage)

Changes:

• Extended ChatMessage structure to support both string and structured content blocks
• Added ContentBlock and ImageBlock structs matching OpenAI API format
• Maintains backward compatibility with text-only content

What Works:

• Chat template rendering accepts structured content blocks
• Tokenization includes image URLs/base64 in rendered template strings
• Python wrapper correctly handles OpenAI API format
• Backward compatible with text-only requests

Known Limitations (for current stage):

• Tokenization may not match vLLM exactly (images tokenized as text, not vision tokens - this only affects merged preprocessor models like Qwen2-VL)
• Block hashes may not match vLLM exactly (missing mm_hash - we're consistent with ourselves but won't match vLLM's hashes for multimodal blocks, which is not an issue, just FYI)
• These will be addressed in the next stage.

Testing:

• Python wrapper tests for multi-modality support

[hyeongyun0916]

It looks like #219 is working toward the same goal — adding multi-modal content support to the chat template preprocessing.