Add ShakeBottle selected-16000 release metadata (upload in progress)

real_data/shakebottle/DATASET_INFO.json

@@ -0,0 +1,28 @@
+{
+  "action_dim": 14,
+  "cameras": [
+    "head_camera",
+    "left_camera",
+    "right_camera"
+  ],
+  "excluded_surplus_candidates": 5256,
+  "formal_episodes": 16000,
+  "hdf5_image_shapes": [
+    240,
+    320,
+    3
+  ],
+  "left": 8000,
+  "num_archives": 112,
+  "physics_hz": 250,
+  "processed_cache_included": false,
+  "publication": "Byte-identical selected HDF5 episodes, paired original JSON configs and episode metadata.",
+  "right": 8000,
+  "robot": "Piper user real URDF",
+  "robot_urdf_sha256": "7b40446de9af155391a4c4efe5a254fb2812c56ef8b66551e3a9f4a9ea58fded",
+  "save_freq": 15,
+  "simulated": true,
+  "source": "/data/huangzengxiang/intact-flash-sim2real/dataset/robotwin-high-entropy/single-arm/shake-bottle",
+  "source_selection_manifest_sha256": "5fa39c05688213b7117e8840d154a2f170d7732cb19bf4e0151e1513ec1b305f",
+  "task": "shake_bottle"
+}

real_data/shakebottle/README.md

@@ -0,0 +1,39 @@
+# ShakeBottle selected 16,000 trajectories
+
+This release contains exactly 16,000 selected simulated ShakeBottle trajectories,
+8,000 using the left arm and 8,000 using the right arm, collected with the Piper
+user real-robot URDF (piper_ros@ac41fcbc). These are simulated trajectories;
+the parent directory name `real_data` does not mean real-world recordings.
+
+Each shard contains byte-identical original HDF5 episodes under
+`shard_*/data/episode*.hdf5`, their original configurations under
+`shard_*/configs/episode*.json`, and episode metadata under `shard_*/manifests/`.
+Three original RGB views, measured joint states, absolute joint control targets,
+the raw 250 Hz stream, timestamps, calibration, and scene metadata are preserved.
+Physics runs at 250 Hz with save_freq=15 (nominal observations at 16.6667 Hz).
+No image conversion, resampling, normalization, or action realignment is applied.
+The 5,256 unselected successful candidates and the processed NumPy training cache
+are excluded. This matches the selected-trajectory publication used for ClickBell.
+
+During publication, UPLOAD_STATUS.json says `uploading`. Wait until it says
+`complete`, then check SHA256SUMS before extracting all shards into one directory:
+
+```python
+from pathlib import Path
+import tarfile
+from huggingface_hub import snapshot_download
+
+root = Path(snapshot_download("INTACT-JEPA/INTACT",
+    allow_patterns="real_data/shakebottle/*")) / "real_data/shakebottle"
+for archive in sorted(root.glob("shakebottle-*-of-*.tar.gz")):
+    with tarfile.open(archive, "r:gz") as f:
+        f.extractall("shakebottle_episodes", filter="data")
+```
+
+The final selection_manifest.jsonl maps all 16,000 selections to their published
+archive, HDF5 member, byte count, and SHA-256, as well as the paired config hash.
+source_selection_manifest.jsonl and source_summary.json preserve the original
+selection. DATASET_INFO.json describes the release. SHA256SUMS covers all other
+published files. Source completion markers are prefixed `source_` and describe
+the source dataset, not the upload progress. Downloaded HDF5 SHA-256 hashes can
+be checked against selection_manifest.jsonl to verify exact byte preservation.

real_data/shakebottle/UPLOAD_STATUS.json

@@ -0,0 +1,8 @@
+{
+  "expected_archives": 112,
+  "expected_episodes": 16000,
+  "status": "uploading",
+  "updated_at": "2026-10-22T20:22:39.821970+08:00",
+  "uploaded_archives": 0,
+  "uploaded_episodes": 0
+}

real_data/shakebottle/collection_contract.json

@@ -0,0 +1,11 @@
+{
+  "task": "shake-bottle",
+  "original_source": "/mnt/nas-h/junhan/10_datasets/robotwin-2.0/piper/shake-bottle/raw/piper-shake-bottle-official-default-250hz-3view-realized-high-entropy-official250-v1",
+  "robot_urdf_sha256": "7b40446de9af155391a4c4efe5a254fb2812c56ef8b66551e3a9f4a9ea58fded",
+  "source_conditions_unchanged": true,
+  "target_formal_episodes": 16000,
+  "excluded_worker_hosts": [
+    "4090-13",
+    "4090-14"
+  ]
+}

real_data/shakebottle/source_dataset_origin.json

@@ -0,0 +1,12 @@
+{
+  "task": "shake-bottle",
+  "collection_robot": "user_real_urdf",
+  "collected_with_user_real_urdf": true,
+  "robot_urdf_sha256": "7b40446de9af155391a4c4efe5a254fb2812c56ef8b66551e3a9f4a9ea58fded",
+  "formal_manifest_sha256": "5fa39c05688213b7117e8840d154a2f170d7732cb19bf4e0151e1513ec1b305f",
+  "source_dataset_provenance": "/mnt/nas-h/junhan/10_datasets/robotwin-2.0/piper/shake-bottle/raw/piper-shake-bottle-official-default-250hz-3view-realized-high-entropy-official250-v1",
+  "inherited_environment": "Junhan task/camera/action/success environment retained; only the robot URDF was replaced with the user realURDF.",
+  "provenance_note": "The source_dataset field is inherited environment provenance, not a live training input path. Training reads the user-owned dataset directory.",
+  "training_input_root": "/data/huangzengxiang/intact-flash-sim2real/dataset/robotwin-high-entropy/single-arm/shake-bottle/processed/formal-16000",
+  "evaluation_requirement": "Same-distribution evaluation must use the user realURDF and inherited task/camera/action contract; original-robot evaluation is separate."
+}

real_data/shakebottle/source_delivery_verified.json

@@ -0,0 +1,15 @@
+{
+  "all_shard_sha256_verified": true,
+  "candidate_episodes": 21256,
+  "completed_unix": 1791285489.7807014,
+  "coverage_support": "original fixed support; entropy recomputed on new trajectories",
+  "destination": "/data/huangzengxiang/intact-flash-sim2real/dataset/robotwin-high-entropy/single-arm/shake-bottle",
+  "formal_episodes": 16000,
+  "formal_manifest_sha256": "5fa39c05688213b7117e8840d154a2f170d7732cb19bf4e0151e1513ec1b305f",
+  "robot_urdf_sha256": "7b40446de9af155391a4c4efe5a254fb2812c56ef8b66551e3a9f4a9ea58fded",
+  "selection_override": "control/selection_override.json",
+  "source_dataset": "/mnt/nas-h/junhan/10_datasets/robotwin-2.0/piper/shake-bottle/raw/piper-shake-bottle-official-default-250hz-3view-realized-high-entropy-official250-v1",
+  "status": "VERIFIED",
+  "success_criteria": "original task criteria unchanged",
+  "task": "shake-bottle"
+}

real_data/shakebottle/source_summary.json

@@ -0,0 +1,42 @@
+{
+  "accuracy_support_verified": false,
+  "all_observed_cells": 28,
+  "arm_balanced_image_cell_share_lower_bound": 0.08125,
+  "audited_projection_records": 21256,
+  "candidate_records": 21256,
+  "cell_max": 1299,
+  "cell_median_nonzero": 650,
+  "cell_min_feasible": 192,
+  "cell_min_nonzero": 192,
+  "coverage_cell_source": "hdf5_critical_point_projection",
+  "excluded_unstable_boundary_cells": {},
+  "feasible_image_cells": 18,
+  "formal_stable_cell_minimum": 64,
+  "image_8x8_occupied": 18,
+  "image_cell_share_gate": 0.08125,
+  "in_grid_surplus_not_selected": 5256,
+  "left": 8000,
+  "master_candidate_records": 21256,
+  "max_image_cell_share": 0.08125,
+  "normalized_entropy_8x8_image": 0.9692660370543952,
+  "normalized_entropy_8x8_image_all64": 0.6736277801603389,
+  "normalized_entropy_8x8x3_all192": 0.6149921391916243,
+  "normalized_entropy_8x8x3_empirical_feasible": 0.9703240949828214,
+  "official_baseline_subset_verified": true,
+  "optimized_cell_floor": 192,
+  "optimized_cell_peak": 1299,
+  "projection_outside_fixed_roi": 0,
+  "realized_cells": 28,
+  "right": 8000,
+  "schema_version": "robotwin-piper-official-default-250hz-3view-formal16000-v1",
+  "selected_records": 16000,
+  "selection_manifest": "selection_manifest.jsonl",
+  "source_audit_reports": 112,
+  "stable_feasible_min_success_count": 8,
+  "stable_support_mask": "/data/huangzengxiang/intact-flash-sim2real/dataset/robotwin-high-entropy/single-arm/shake-bottle/master/replay_metadata/stable_support_mask.json",
+  "stable_support_mask_sha256": "d9c0aba79dd506bb9169a0b9951edef9dd17c5e231f189a90913aeed76199346",
+  "stable_support_verified": true,
+  "surplus_not_selected": 5256,
+  "task_name": "shake_bottle",
+  "uniform_image_cell_share_lower_bound": 0.05555555555555555
+}