Add T1 dance task config and update motion tool docs

[fanyahao1]

Summary

This PR adds the first T1 dance motion-tracking task configuration to booster_train, updates the play environment registration used by the existing K1 dance task, and refreshes the motion conversion / replay documentation so the workflow matches the current scripts.

What changed

1. Add T1 dance task configuration

Introduces a new T1 dance task package under source/booster_train/booster_train/tasks/manager_based/beyond_mimic/robots/T1/dance/.

Included files:

• __init__.py
  • registers Booster-T1-Dance-v0
  • registers Booster-T1-Dance-v0-Play
• env_cfg.py
  • defines the T1 flat / rough / play environment configs
  • switches the scene robot to BOOSTER_T1_CFG
  • sets the T1 action scale
  • points the motion command to motions/T1/CLIO_Dimitroula_1_stageii.npz
  • configures the T1 anchor body and tracked body list
• tracking_env_cfg.py
  • adds the T1-specific tracking environment base config
  • defines scene, observations, rewards, events, terminations, and curriculum hooks for the T1 motion-tracking setup
• ppo_cfg.py
  • adds the PPO runner config entry point for the new task

2. Add T1 robot support to motion tools

Updates the motion conversion / replay scripts so they can be used with both K1 and T1.

Changes include:

• move the utilities under scripts/mimic/
• scripts/mimic/csv_to_npz.py
  • adds --robot {k1,t1}
  • selects the matching robot config and joint-name mapping
  • supports frame sub-selection with --frame_range
  • keeps configurable output interpolation via --output_fps
• scripts/mimic/replay_npz.py
  • adds --robot {k1,t1}
  • supports replay from either a local .npz file or a Weights & Biases registry artifact
  • automatically appends :latest when the registry alias is omitted
  • resolves the anchor body from the converted motion file for replay

3. Update dance play environment registration

Updates the existing K1 dance config so the play environment uses PlayFlatWoStateEstimationEnvCfg instead of the training config, keeping the play-time setup aligned with the dedicated play behavior.

4. Refresh README documentation

Expands README.md so the documented workflow matches the current scripts:

• updates script paths from scripts/... to scripts/mimic/...
• fixes the conversion argument name to --output_file
• documents --robot, --frame_range, and --output_fps
• adds replay examples for both local motion files and W&B registry artifacts

Commit breakdown

• 3b51bc5 — feat: add PlayFlatWoStateEstimationEnvCfg and update gym registration for play environment
• 54165a9 — docs: update motion conversion and replay tools README
• 0216b24 — feat: add T1 dance task with motion tracking environment config

Diff summary

• 3 commits
• 11 changed files
• 566 additions / 27 deletions

Notes

• local untracked helper scripts were intentionally not included in this PR
• no additional test run is included in this PR description

[fanyahao1]

This change only covers the booster_train side. The motion file configured here in source/booster_train/booster_train/tasks/manager_based/beyond_mimic/robots/T1/dance/env_cfg.py L17: motions/T1/CLIO_Dimitroula_1_stageii.npz is coupled with corresponding updates in the related booster_asset as well, and I plan to open follow-up PRs there separately to keep the full workflow in sync. So if this path looks ahead of the current upstream state in other repos, that is expected for now.