support attn_dp_expert parallelism

[khatwanimohit]

Description

• Add attn_dp_expert mesh axis for expert parallelism: Introduce a new attn_dp_expert axis in the mesh and logical axis rules to support attention DP-aware expert parallelism, particularly when using
  vllm_rpa attention.
• Make expert axis name configurable in MoE layer: Replace hardcoded "expert" axis name in RoutedMoE with a configurable _expert_parallelism_name that switches to "attn_dp_expert" when attention ==
  "vllm_rpa", affecting all collective operations (all_gather, psum_scatter, axis_index, ragged_all_to_all).
• Add rollout_expert_parallelism config for RL training: New config field to specify expert parallelism per replica for rollout, with updated device count validation (tp * dp * ep ==
  num_sampler_devices).
• Pass expert parallelism settings to vLLM rollout: When rollout_expert_parallelism > 1, pass expert_parallel_size and rollout_vllm_enable_expert_parallelism to vLLM kwargs.
• Add vllm_config_path config option: New config field (in base.yml, rl.yml, and types.py) to specify the path to a YAML file for loading vLLM config, defaulting to
  src/maxtext/configs/inference/vllm.yml.
• Update vLLM inference script: Add vllm_swap_space, vllm_async_scheduling, and vllm_config_path flags to vllm_decode.py; remove the hard requirement on hf_config_path.
• Update vLLM logical axis rules: Add attn_dp_expert to sharding rules for activation axes (activation_q_length, activation_kv_batch, decode_batch, embed, exp, etc.) and add moe_mlp rule in the vLLM
  inference config.

Notice 1: Once all tests pass, the "pull ready" label will automatically be assigned.
This label is used for administrative purposes. Please do not add it manually.

Notice 2: For external contributions, our settings currently require an approval from a MaxText maintainer to trigger CI tests.

Tests

Please describe how you tested this change, and include any instructions and/or
commands to reproduce.

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 22.72727% with 17 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/MaxText/rl/train_rl.py	0.00%	9 Missing ⚠️
src/MaxText/layers/moe.py	38.46%	6 Missing and 2 partials ⚠️

📢 Thoughts on this report? Let us know!

[gobbleturk]

Is the tuple case actually supported? Have you tested it?