[RyanPCo]

• paths/default.yaml: restore root_dir (${oc.env:PROJECT_ROOT,${hydra:runtime.cwd}});
  log_dir had referenced it while it was commented out, so every top-level
  config failed a full resolve.
• data/*.yaml (14 files): ${train_datasets.x} -> ${data.train_datasets.x}. Data
  configs are mounted under data., so the root-relative form raised
  InterpolationKeyError on the first cfg.data.valid_datasets[name] access;
  the repo default data: eva had been unrunnable since 4ca49ca (May).
• data/mecka_pi.yaml, data/scale_pi.yaml: plain-dict / untargeted filters: ->
  DatasetFilter with filter_lambdas (the zarr resolver rejects anything else).
  mecka_pi's task string becomes 'fold_clothes', the value every other mecka
  config filters on; 'folding_clothes' matched nothing.
• train_zarr_keypoints / train_zarr_keypoint_wrist: viz override pointed at
  eva_cartesian_aria_keypoints*, removed in 9788e2b (now keypoints*);
  the two files also had each other's data config.
• viz_language.yaml: add the paths group its data config interpolates.
• tests/unit/test_hydra_configs.py: composes and fully resolves every top-level
  config, every data option under every train config, every model / evaluator /
  viz option, and checks each data config's filters is a DatasetFilter
  (instantiated only for the base class; subclasses may need API keys).
  164 parametrized cases, no data, network, or cluster paths.

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
• fix(configs): fix paths for data hydra yaml configs #622 👈 (View in Graphite)
• 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.