[RyanPCo]

• Delete setup.py (superseded by [project]) and environment.yaml (a wrapper
  around requirements.txt). requirements.txt stays untouched: the Eva robot
  Dockerfile installs it (it has drifted from the lock: torch 2.6.0 /
  transformers 4.57.3 vs 2.7.1 / 4.53.2, but that image is out of scope).
  requirements-ray.txt stays: the Ray cluster yamls install it.
• packages.find: include = ["egomimic*"]. The wheel had shipped external/
  (lerobot, rpl_vision_utils, scale) and tests/ as top-level packages.
• package-data: hydra_configs/**/.yaml and resources/ (0 of 76 yamls and no
  URDF/XML were in the wheel, so only pip install -e . ever worked).
• Declare scipy (imported by 7 modules, arrived transitively); uv.lock updated.
• openpi is documented, not declared: the fork's lerobot pin needs av>=14.2,
  which conflicts with av==12.0.0, so an extra cannot resolve. README now gives
  the no-deps install + transformers_replace patch the lab venv actually uses.
• README install section points at pyproject; .gitignore build/.
• tests/unit/test_packaging.py builds the wheel via the build backend (from a
  clean build/, removed again afterwards) and checks it ships every hydra yaml
  and the resources, and nothing outside egomimic/.
  Also verified by installing the wheel into a fresh venv.

Not touched (reverted from an earlier revision of this PR): Dockerfile and
requirements.txt. The Eva robot image installs requirements.txt in step 11,
so both stay exactly as on main. The only remaining reference to a deleted
file was the README's conda env create -f environment.yaml, now a plain
conda create -n emimic python=3.11. pip install -e . and the wheel build
go through pyproject's setuptools backend and never needed setup.py.

Co-Authored-By: Claude Fable 5.1 [email protected]
Claude-Session: https://claude.ai/code/session_01UHDiN8yfERcsiFdz4mXuZi

🤖 Generated with Claude Code

https://claude.ai/code/session_01UHDiN8yfERcsiFdz4mXuZi

[RyanPCo]

• import cloud packages lazily to support local training #631 : 2 dependent PRs (#632 , #633 )
• cache dataset resolver for faster training initialization #630
• validate episode hashes in data config before training #628
• modify pi camera map to match data #629
• add tests to ensure training works between every model and data vendor #626
• run CI tests on every PR #625
• add env variables for user-specific paths #624
• build: set pyproject and uv as the dependency managers #623 👈 (View in Graphite)
• fix(configs): fix paths for data hydra yaml configs #622
• refactor(algo): delete ACT #627 : 1 other dependent PR (#621 )
• tests: create dedicated test folder with unit and integration tests #620
• add pyright type checker #619
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.