[Misc] Consolidate LingBot text-to-video runner

[hsliuustc0106]

Purpose

Consolidate LingBot-Video offline inference into the shared text_to_video.py runner and remove the dedicated text_to_video_lingbot.py script.

• add LingBot model/class detection and preserve the dedicated example's defaults
• build the canonical video request envelope in the shared runner
• route LingBot-only options through --extra-body
• preserve already-normalized [0, 1] frame tensors during export
• scope distilled preset detection to the LTX family and support class-based preset detection
• update the LingBot recipe, user docs, and dense parity benchmark reference

Test Plan

• Run focused CPU coverage for shared T2V prompts/presets/export normalization, LingBot request normalization, and model-extra routing.
• Run the shared runner on one NVIDIA L20X with the cached LingBot dense checkpoint.
• Compare the default shared-runner artifact against the retired dedicated runner using the same model, seed, and sampling settings.
• Exercise migrated LingBot batch_cfg and output_type options through --extra-body.
• Run a non-LingBot shared-runner smoke with the tiny Helios pipeline.

vLLM Version: 0.26.0

vLLM-Omni Commit: 0a593bd

Test Result

• 127 passed, 14 warnings in 1.81s
• pre-commit hooks passed on all changed files
• LingBot default: valid H.264, 320x192, 24 FPS, 9 frames; byte-identical MP4 and decoded frame hashes versus the dedicated runner
• LingBot model extras: --extra-body '{"batch_cfg": true, "output_type": "np"}' produced a valid H.264, 320x192, 24 FPS, 9-frame artifact
• Helios non-LingBot guardrail: valid H.264, 64x64, 16 FPS artifact using the tiny modular pipeline with SDPA

[chatgpt-codex-connector]

Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits.
Credits must be used to enable repository wide code reviews.

[vllm-omni-review-bot]

This PR appears to belong to: docs/design/module/benchmarking.md.

Module owners: @alex-jw-brooks @Bounty-hunter

Please take a look when you have a chance. If you would like an automated review, mention @vllm-omni-review-bot in a comment.