Add prompt preprocessing and token placeholders for Omni

[hengtaoguo]

Description

Collaborate with original author: @eitanporat at PR #2731.

• Add reformat_prompt_qwen3_omni for prompt template.
• Add get_mm_offsets_qwen3_omni to compute multimodal token lengths.
• Update prepare_text_for_image_fusion, accepting config so it allows more complicated placeholder formatting by using model-specific config fields.

Tests

Follow up in e2e tests in #2729

Added two more test asserts in qwen3_omni_layers_test.py:

• Check the formatted prompts are the same.
• Check the tokenized IDs are the same (with multimodal placeholder paddings inside).

python -m pytest tests/unit/qwen3_omni_layers_test.py::TestQwen3OmniPreprocessing::test_preprocess_mm_data -vv -s

Prompt (with placeholders): https://paste.googleplex.com/6444051776798720
Tokenized prompt (with placeholders): https://diff.googleplex.com/#key=v8U2Vrap2cyR

Checklist

Before submitting this PR, please make sure (put X in square brackets):

• I have performed a self-review of my code. For an optional AI review, add the gemini-review label.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have run end-to-end tests tests and provided workload links above if applicable.
• I have made or will make corresponding changes to the doc if needed, including adding new documentation pages to the relevant Table of Contents (toctree directive) as explained in our documentation.

[codecov]

Codecov Report

❌ Patch coverage is 16.00000% with 21 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/maxtext/multimodal/processor.py	13.63%	17 Missing and 2 partials ⚠️
src/maxtext/decode.py	0.00%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!