Add files using upload-large-folder tool

DATA_EXPLORER_ANALYSIS.md

@@ -0,0 +1,42 @@
+# Ropedia Xperience-10M Data Explorer Analysis
+
+Generated: 2026-06-23T09:35:08Z
+
+This report summarizes three data scopes without mixing them: the official public sample episode, the selected 128-episode public-safe feature surface, and authenticated metadata for the full gated Hugging Face dataset.
+
+## Scope Summary
+
+| Scope | Episodes | Rows / windows | Storage view | Notes |
+|---|---:|---:|---:|---|
+| Public sample | 1 | 1,161 | 4.76 GiB | Raw sample files are playable or source-linked. |
+| Selected 128 | 128 | 34,269 | 277.71 GiB | Public-safe matrices and window manifests, not raw redistribution. |
+| Full HF dataset | 12,103 episode-like folders | 3,098,112 projected rows at 256/episode | 24.63 TiB | Gated upstream file metadata only. |
+
+## Public Sample
+
+- 5,821 frames at about 20.00 fps.
+- 1,161 aligned 20-frame windows with 5-frame stride.
+- 8,546 model-input dimensions across 7 modality groups.
+- 35 action segments and 34 object labels in the derived explorer.
+
+## Selected 128 Episodes
+
+- Split: train 96, val 16, test 16 episodes.
+- Size bands: short 32, lower_mid 32, upper_mid 32, long 32.
+- Qwen3-Omni v6 multiscale export: 34,269 rows.
+- Dense multiscale compact export: 106,095 rows.
+
+## Full Gated Dataset Metadata
+
+- Repo: `ropedia-ai/xperience-10m` at `ce943cf271a758b60240084892d05cf6dc12dd90`.
+- 85,257 files excluding `.gitattributes`.
+- 12,102 complete episode folders (99.9917%).
+- 72,612 MP4 files and 12,103 `annotation.hdf5` files.
+
+## Generated Charts
+
+- Scope ladder: `assets/charts/data_explorer_scope_ladder.svg`
+- Public sample feature dimensions: `assets/charts/data_explorer_sample_feature_modalities.svg`
+- Public sample action distribution: `assets/charts/data_explorer_sample_action_distribution.svg`
+- Selected-128 split rows: `assets/charts/data_explorer_selected128_split_rows.svg`
+- Full dataset file composition: `assets/charts/data_explorer_full_file_composition.svg`

data/data_explorer_analysis.json

@@ -0,0 +1,782 @@
+{
+  "charts": [
+    {
+      "path": "assets/charts/data_explorer_scope_ladder.svg",
+      "question": "How do the public sample, selected 128 episodes, and full gated dataset differ in scale?",
+      "title": "Scope ladder"
+    },
+    {
+      "path": "assets/charts/data_explorer_sample_feature_modalities.svg",
+      "question": "Which modality groups dominate the one-sample task input?",
+      "title": "Public sample feature dimensions"
+    },
+    {
+      "path": "assets/charts/data_explorer_sample_action_distribution.svg",
+      "question": "Which action labels occupy the most 20-frame windows in the sample?",
+      "title": "Public sample action distribution"
+    },
+    {
+      "path": "assets/charts/data_explorer_selected128_split_rows.svg",
+      "question": "How many rows are available per selected-128 export and split?",
+      "title": "Selected-128 split rows"
+    },
+    {
+      "path": "assets/charts/data_explorer_full_file_composition.svg",
+      "question": "What file types dominate the gated full-dataset metadata?",
+      "title": "Full dataset file composition"
+    }
+  ],
+  "full_hf_dataset": {
+    "annotation_file_size_summary": {
+      "count": 12103,
+      "max_bytes": 2000432556,
+      "max_human": "1.86 GiB",
+      "mean_bytes": 1830083732,
+      "mean_human": "1.70 GiB",
+      "median_bytes": 1969457752,
+      "median_human": "1.83 GiB",
+      "min_bytes": 6687192,
+      "min_human": "6.38 MiB",
+      "p25_bytes": 1869556164,
+      "p75_bytes": 1989975784
+    },
+    "basename_counts": {
+      "README.md": 1,
+      "annotation.hdf5": 12103,
+      "fisheye_cam0.mp4": 12102,
+      "fisheye_cam1.mp4": 12102,
+      "fisheye_cam2.mp4": 12102,
+      "fisheye_cam3.mp4": 12102,
+      "stereo_left.mp4": 12102,
+      "stereo_right.mp4": 12102,
+      "visualization.rrd": 541
+    },
+    "card_data": {
+      "extra_gated_button_content": "I have signed the access control agreement",
+      "extra_gated_description": "Access is reviewed manually and is limited to approved non-commercial use. Completion of an external agreement-signing step may be required before approval. Please sign the agreement via [DocuSign](https://ropedia.docsend.com/view/ra7ej7gs6s98sw87)",
+      "extra_gated_heading": "Request controlled access to Xperience-10M",
+      "language": [
+        "en"
+      ],
+      "license": "other",
+      "pretty_name": "Xperience-10M",
+      "size_categories": [
+        "1M<n<10M"
+      ],
+      "tags": [
+        "egocentric",
+        "first-person",
+        "multimodal",
+        "3d",
+        "4d",
+        "embodied-ai",
+        "robotics",
+        "human-motion",
+        "mocap",
+        "imu",
+        "audio",
+        "depth",
+        "captions",
+        "video"
+      ],
+      "task_categories": [
+        "video-classification",
+        "image-to-text",
+        "depth-estimation",
+        "robotics"
+      ]
+    },
+    "complete_episode_training_size_summary": {
+      "count": 12102,
+      "max_bytes": 2721499520,
+      "max_human": "2.53 GiB",
+      "mean_bytes": 2237801505,
+      "mean_human": "2.08 GiB",
+      "median_bytes": 2365869284,
+      "median_human": "2.20 GiB",
+      "min_bytes": 8158968,
+      "min_human": "7.78 MiB",
+      "p25_bytes": 2286228606,
+      "p75_bytes": 2412829381
+    },
+    "episode_count_per_session_summary": {
+      "count": 803,
+      "max": 51,
+      "mean": 15.07,
+      "median": 11,
+      "min": 1,
+      "p25": 8,
+      "p75": 13
+    },
+    "episode_size_summary": {
+      "count": 12103,
+      "max_bytes": 5108045286,
+      "max_human": "4.76 GiB",
+      "mean_bytes": 2318233840,
+      "mean_human": "2.16 GiB",
+      "median_bytes": 2369826102,
+      "median_human": "2.21 GiB",
+      "min_bytes": 8158968,
+      "min_human": "7.78 MiB",
+      "p25_bytes": 2292986495,
+      "p75_bytes": 2420801956
+    },
+    "file_type_counts": {
+      ".hdf5": 12103,
+      ".md": 1,
+      ".mp4": 72612,
+      ".rrd": 541
+    },
+    "gated": "manual",
+    "incomplete_episode_records": [
+      {
+        "episode_id": "ep1",
+        "episode_path": "dc3f4139-f499-4de7-b057-e25b7dfb2d83/ep1",
+        "file_count": 1,
+        "has_all_six_videos": false,
+        "has_annotation": true,
+        "has_fisheye_cam0": false,
+        "has_visualization_rrd": false,
+        "is_complete": false,
+        "is_degraded_valid": false,
+        "missing_required_files": [
+          "fisheye_cam0.mp4",
+          "fisheye_cam1.mp4",
+          "fisheye_cam2.mp4",
+          "fisheye_cam3.mp4",
+          "stereo_left.mp4",
+          "stereo_right.mp4"
+        ],
+        "top_level_session": "dc3f4139-f499-4de7-b057-e25b7dfb2d83",
+        "total_bytes": 1418232696,
+        "training_bytes_excluding_visualization_rrd": 1418232696,
+        "video_count": 0
+      }
+    ],
+    "last_modified": "2026-04-21T05:03:45+00:00",
+    "metadata_note": "The official dataset is gated. These full-corpus figures use authenticated Hugging Face Hub file metadata only; they do not inspect private row content or redistribute raw MP4/HDF5/RRD files.",
+    "pilot_scale_estimates": {
+      "all_complete_episodes_windows_at_256_each": 3098112,
+      "episode_100_windows_at_256_each": 25600,
+      "episode_32_windows_at_256_each": 8192,
+      "episode_500_windows_at_256_each": 128000,
+      "mean_based_32_episode_training_bytes": 71609648160,
+      "mean_based_32_episode_training_human": "66.69 GiB",
+      "median_based_32_episode_training_bytes": 75707817088,
+      "median_based_32_episode_training_human": "70.51 GiB",
+      "windows_per_episode": 256
+    },
+    "repo_id": "ropedia-ai/xperience-10m",
+    "repo_sha": "ce943cf271a758b60240084892d05cf6dc12dd90",
+    "summary": {
+      "all_complete_episode_training_bytes_excluding_visualization_rrd": 27081873816281,
+      "all_complete_episode_training_human_excluding_visualization_rrd": "24.63 TiB",
+      "annotation_hdf5_count": 12103,
+      "complete_episode_count": 12102,
+      "complete_episode_pct": 99.9917,
+      "complete_sessions": 802,
+      "degraded_valid_episode_count": 12102,
+      "degraded_valid_episode_pct": 99.9917,
+      "degraded_valid_sessions": 802,
+      "episode_like_folder_count": 12103,
+      "file_count_excluding_gitattributes": 85257,
+      "mp4_count": 72612,
+      "sibling_count": 85258,
+      "top_level_session_count": 804,
+      "total_bytes_from_file_metadata": 28057584187079,
+      "total_human_from_file_metadata": "25.52 TiB",
+      "training_bytes_excluding_visualization_rrd": 27083292060675,
+      "training_human_excluding_visualization_rrd": "24.63 TiB",
+      "visualization_rrd_bytes": 974292126404,
+      "visualization_rrd_count": 541,
+      "visualization_rrd_human": "907.38 GiB"
+    },
+    "video_count_histogram": {
+      "0": 1,
+      "6": 12102
+    }
+  },
+  "generated_at_utc": "2026-06-23T09:35:08Z",
+  "public_sample": {
+    "dataset": {
+      "episode_id": "xperience-10m-sample",
+      "hosting_policy": "Raw files are streamed or downloaded from the official public sample dataset. This task-suite repo mirrors derived artifacts and metadata only.",
+      "license": "cc-by-nc-4.0",
+      "repo_commit": "2d80b5af8085a18ffe702e1420968308007f6204",
+      "repo_id": "ropedia-ai/xperience-10m-sample",
+      "repo_url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m-sample"
+    },
+    "feature_blocks": [
+      {
+        "dim": 1092,
+        "display": "Body Joints",
+        "modality": "motion_capture",
+        "name": "body_joints"
+      },
+      {
+        "dim": 980,
+        "display": "Depth + Confidence",
+        "modality": "depth",
+        "name": "depth_confidence"
+      },
+      {
+        "dim": 896,
+        "display": "Language Text",
+        "modality": "language",
+        "name": "caption_objects_interaction_text"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam0",
+        "modality": "video",
+        "name": "video_fisheye_cam0"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam1",
+        "modality": "video",
+        "name": "video_fisheye_cam1"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam2",
+        "modality": "video",
+        "name": "video_fisheye_cam2"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam3",
+        "modality": "video",
+        "name": "video_fisheye_cam3"
+      },
+      {
+        "dim": 686,
+        "display": "Video Stereo Left",
+        "modality": "video",
+        "name": "video_stereo_left"
+      },
+      {
+        "dim": 686,
+        "display": "Video Stereo Right",
+        "modality": "video",
+        "name": "video_stereo_right"
+      },
+      {
+        "dim": 441,
+        "display": "Left Hand",
+        "modality": "motion_capture",
+        "name": "hand_left_joints"
+      },
+      {
+        "dim": 441,
+        "display": "Right Hand",
+        "modality": "motion_capture",
+        "name": "hand_right_joints"
+      },
+      {
+        "dim": 168,
+        "display": "Audio",
+        "modality": "audio",
+        "name": "audio_fisheye_cam0_aac"
+      },
+      {
+        "dim": 147,
+        "display": "Body Contacts",
+        "modality": "motion_capture",
+        "name": "body_contacts"
+      },
+      {
+        "dim": 117,
+        "display": "Calibration",
+        "modality": "pose_slam",
+        "name": "calibration"
+      },
+      {
+        "dim": 63,
+        "display": "Camera Rotation",
+        "modality": "pose_slam",
+        "name": "camera_rotation_matrix"
+      },
+      {
+        "dim": 42,
+        "display": "IMU Accel/Gyro",
+        "modality": "inertial",
+        "name": "imu_accel_gyro"
+      },
+      {
+        "dim": 22,
+        "display": "SLAM Point Cloud",
+        "modality": "pose_slam",
+        "name": "slam_point_cloud"
+      },
+      {
+        "dim": 21,
+        "display": "Camera Translation",
+        "modality": "pose_slam",
+        "name": "camera_translation"
+      }
+    ],
+    "feature_dim_by_modality": {
+      "audio": 168,
+      "depth": 980,
+      "inertial": 42,
+      "language": 896,
+      "motion_capture": 2121,
+      "pose_slam": 223,
+      "video": 4116
+    },
+    "file_bytes_by_kind": {
+      "hdf5_annotation_container": {
+        "bytes": 1931496028,
+        "count": 1,
+        "human": "1.80 GiB"
+      },
+      "rerun_viewer_recording": {
+        "bytes": 2702924036,
+        "count": 1,
+        "human": "2.52 GiB"
+      },
+      "video": {
+        "bytes": 383606257,
+        "count": 5,
+        "human": "365.84 MiB"
+      },
+      "video_with_audio": {
+        "bytes": 89842251,
+        "count": 1,
+        "human": "85.68 MiB"
+      }
+    },
+    "file_count": 8,
+    "hdf5_groups": [
+      {
+        "description": "Camera intrinsics/extrinsics and static calibration values used to align sensor streams.",
+        "group": "calibration"
+      },
+      {
+        "description": "JSON text with task config, action segments, object mentions, interaction descriptions, and global summary.",
+        "group": "caption"
+      },
+      {
+        "description": "Depth maps and confidence channels aligned to the episode timeline.",
+        "group": "depth"
+      },
+      {
+        "description": "Full-body joint and contact signals for human motion modeling.",
+        "group": "full_body_mocap"
+      },
+      {
+        "description": "Left and right hand joint trajectories used by hand-motion and forecast tasks.",
+        "group": "hand_mocap"
+      },
+      {
+        "description": "Accelerometer and gyroscope streams sampled at higher frequency than the video frames.",
+        "group": "imu"
+      },
+      {
+        "description": "Episode metadata, frame indexing, and source bookkeeping.",
+        "group": "metadata"
+      },
+      {
+        "description": "Camera trajectory, pose, and sparse SLAM point-cloud information.",
+        "group": "slam"
+      },
+      {
+        "description": "Video stream metadata and per-frame alignment information.",
+        "group": "video"
+      }
+    ],
+    "object_vocab_count": 34,
+    "segment_count": 35,
+    "source_policy": "Window-level labels, features, predictions, and diagnostics are embedded here. Official raw MP4/HDF5/RRD files are linked from the Raw Sample Browser, with compact browser-preview clips for immediate MP4/audio playback.",
+    "top_actions": [
+      {
+        "count": 159,
+        "name": "Hold gooseneck kettle"
+      },
+      {
+        "count": 159,
+        "name": "Pour coffee"
+      },
+      {
+        "count": 127,
+        "name": "Position kettle to pour"
+      },
+      {
+        "count": 95,
+        "name": "Lift gooseneck kettle"
+      },
+      {
+        "count": 95,
+        "name": "Close bottle cap"
+      },
+      {
+        "count": 95,
+        "name": "Wait/Prepare for pouring"
+      },
+      {
+        "count": 87,
+        "name": "Transfer coffee to dripper"
+      },
+      {
+        "count": 63,
+        "name": "Grasp coffee scoop"
+      },
+      {
+        "count": 47,
+        "name": "Hold coffee carafe"
+      },
+      {
+        "count": 39,
+        "name": "Move kettle"
+      },
+      {
+        "count": 31,
+        "name": "Move kettle away"
+      },
+      {
+        "count": 30,
+        "name": "Pick up kettle"
+      }
+    ],
+    "top_objects": [
+      {
+        "count": 636,
+        "name": "coffee dripper"
+      },
+      {
+        "count": 517,
+        "name": "scale"
+      },
+      {
+        "count": 446,
+        "name": "gooseneck kettle"
+      },
+      {
+        "count": 333,
+        "name": "digital scale"
+      },
+      {
+        "count": 238,
+        "name": "water bottle"
+      },
+      {
+        "count": 217,
+        "name": "white mug"
+      },
+      {
+        "count": 203,
+        "name": "white bottle"
+      },
+      {
+        "count": 170,
+        "name": "coffee mug"
+      },
+      {
+        "count": 158,
+        "name": "bottle"
+      },
+      {
+        "count": 145,
+        "name": "dripper"
+      },
+      {
+        "count": 126,
+        "name": "glass carafe"
+      },
+      {
+        "count": 123,
+        "name": "kettle"
+      }
+    ],
+    "top_subtasks": [
+      {
+        "count": 166,
+        "name": "Handle gooseneck kettle"
+      },
+      {
+        "count": 151,
+        "name": "Pour coffee"
+      },
+      {
+        "count": 150,
+        "name": "Pick up and position kettle"
+      },
+      {
+        "count": 119,
+        "name": "Prepare for pouring"
+      },
+      {
+        "count": 111,
+        "name": "Prepare coffee equipment and scoop grounds"
+      },
+      {
+        "count": 111,
+        "name": "Lift gooseneck kettle"
+      },
+      {
+        "count": 111,
+        "name": "Pour and close white bottle"
+      },
+      {
+        "count": 79,
+        "name": "Transfer coffee grounds to dripper"
+      },
+      {
+        "count": 79,
+        "name": "Set down kettle and retrieve white bottle"
+      },
+      {
+        "count": 39,
+        "name": "Move kettle"
+      },
+      {
+        "count": 14,
+        "name": ""
+      },
+      {
+        "count": 10,
+        "name": "Pour milk into coffee"
+      }
+    ],
+    "total_bytes": 5107868572,
+    "total_human": "4.76 GiB",
+    "windowization": {
+      "audio_sample_rate": 16000,
+      "duration_human": "4.85 min",
+      "duration_sec": 291.03000343524565,
+      "feature_dim": 8546,
+      "fps_observed": 20.00137419266181,
+      "note": "The public task suite converts the raw episode into overlapping 20-frame windows with a 5-frame stride.",
+      "num_frames": 5821,
+      "num_windows": 1161,
+      "stride_frames": 5,
+      "window_frames": 20
+    }
+  },
+  "selected_128": {
+    "bytes_by_size_band": {
+      "long": {
+        "bytes": 75928334176,
+        "human": "70.71 GiB"
+      },
+      "lower_mid": {
+        "bytes": 75730496234,
+        "human": "70.53 GiB"
+      },
+      "short": {
+        "bytes": 70693191518,
+        "human": "65.84 GiB"
+      },
+      "upper_mid": {
+        "bytes": 75837737335,
+        "human": "70.63 GiB"
+      }
+    },
+    "bytes_by_split": {
+      "test": {
+        "bytes": 37243648205,
+        "human": "34.69 GiB"
+      },
+      "train": {
+        "bytes": 223733297061,
+        "human": "208.37 GiB"
+      },
+      "val": {
+        "bytes": 37212813997,
+        "human": "34.66 GiB"
+      }
+    },
+    "exports": [
+      {
+        "episodes": 119,
+        "key": "sparse_export",
+        "label": "Sparse selected-128 export",
+        "samples": 3808,
+        "scale_counts": {},
+        "split_counts": {
+          "test": 448,
+          "train": 2848,
+          "val": 512
+        }
+      },
+      {
+        "episodes": 119,
+        "key": "qwen_v6_multiscale_export",
+        "label": "Qwen3-Omni v6 multiscale JSONL",
+        "samples": 34269,
+        "scale_counts": {},
+        "split_counts": {
+          "test": 4032,
+          "train": 25629,
+          "val": 4608
+        }
+      },
+      {
+        "episodes": 119,
+        "key": "dense_multiscale_compact_export",
+        "label": "Dense multiscale compact export",
+        "samples": 106095,
+        "scale_counts": {
+          "dense_20f_stride10": 60725,
+          "long_80f_stride40": 15067,
+          "medium_40f_stride20": 30303
+        },
+        "split_counts": {
+          "test": 11795,
+          "train": 79593,
+          "val": 14707
+        }
+      }
+    ],
+    "metadata_matrix_sparse": {
+      "bytes": 217477,
+      "feature_dim": 394,
+      "human": "212.38 KiB",
+      "row_count": 3808,
+      "sha256": "cdaa55ea3bb229d70bee845f05628ed7a984dd46f95080358d51943b4b8975df",
+      "split_counts": {
+        "test": 448,
+        "train": 2848,
+        "val": 512
+      }
+    },
+    "metadata_matrix_v2": {
+      "bytes": 3375193,
+      "feature_dim": 394,
+      "human": "3.22 MiB",
+      "row_count": 34269,
+      "sha256": "3c2a23b681c9f257a4c04710c72441650844ede6279d3e67be57ccc3abd4fa90",
+      "split_counts": {
+        "test": 4032,
+        "train": 25629,
+        "val": 4608
+      }
+    },
+    "official_dataset": {
+      "access_note": "Episode directory pages are browsable on Hugging Face. Raw annotation/video file downloads require access to the gated official dataset and are not redistributed here.",
+      "gated_resolve_base_url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m/resolve/main",
+      "repo_id": "ropedia-ai/xperience-10m",
+      "repo_sha": "ce943cf271a758b60240084892d05cf6dc12dd90",
+      "tree_base_url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m/tree/main"
+    },
+    "raw20_proxy_tasks": [
+      "camera_view_sync_retrieval",
+      "interaction_text_prediction"
+    ],
+    "raw20_result_records": 40,
+    "selection_summary": {
+      "one_episode_per_top_level_session": true,
+      "selected_download_size_excluding_visualization_rrd_bytes": 298189759263,
+      "selected_episode_count": 128,
+      "selected_split_counts": {
+        "test": 16,
+        "train": 96,
+        "val": 16
+      },
+      "size_band_counts": {
+        "long": 32,
+        "lower_mid": 32,
+        "short": 32,
+        "upper_mid": 32
+      }
+    },
+    "size_band_counts": {
+      "long": 32,
+      "lower_mid": 32,
+      "short": 32,
+      "upper_mid": 32
+    },
+    "sparse_main_task_counts": [
+      {
+        "count": 96,
+        "name": "Stocking grocery shelves"
+      },
+      {
+        "count": 96,
+        "name": "Stocking grocery store shelves"
+      },
+      {
+        "count": 96,
+        "name": "Measuring and marking cardboard"
+      },
+      {
+        "count": 64,
+        "name": "Stocking supermarket shelves"
+      },
+      {
+        "count": 64,
+        "name": "Stocking retail store shelves"
+      },
+      {
+        "count": 64,
+        "name": "Assembling a jigsaw puzzle"
+      },
+      {
+        "count": 64,
+        "name": "Crafting cardboard shapes"
+      },
+      {
+        "count": 64,
+        "name": "Stocking shelves in a retail store"
+      },
+      {
+        "count": 64,
+        "name": "Playing Mahjong"
+      },
+      {
+        "count": 64,
+        "name": "Stocking canned goods on a supermarket shelf"
+      },
+      {
+        "count": 32,
+        "name": "Sorting and organizing colorful buttons on a table"
+      },
+      {
+        "count": 32,
+        "name": "Crafting cardboard shapes and preparing workspace"
+      }
+    ],
+    "sparse_windows_by_split": {
+      "test": 448,
+      "train": 2848,
+      "val": 512
+    },
+    "split_episode_counts": {
+      "test": 16,
+      "train": 96,
+      "val": 16
+    }
+  },
+  "sources": [
+    {
+      "path": "docs/data/raw_sample_files.json",
+      "scope": "public_sample"
+    },
+    {
+      "path": "docs/data/single_episode_explorer.json",
+      "scope": "public_sample"
+    },
+    {
+      "path": "results/episode_task_suite/windows.csv",
+      "scope": "public_sample"
+    },
+    {
+      "path": "docs/data/xperience10m_128_episode_feature_index.json",
+      "scope": "selected_128"
+    },
+    {
+      "path": "results/omni_finetune/xperience10m_128_episode_selection.csv",
+      "scope": "selected_128"
+    },
+    {
+      "path": "results/omni_finetune/multi_episode_128_task_baselines/windows.csv",
+      "scope": "selected_128"
+    },
+    {
+      "path": "results/omni_finetune/full_dataset_metadata_audit.json",
+      "scope": "full_hf_dataset"
+    }
+  ],
+  "status": "pass"
+}

data/mirror_parity.json

@@ -1,9 +1,9 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-23T09:40:30+00:00",
+  "generated_at_utc": "2026-06-23T10:02:51+00:00",
   "hf_root": "hf_publish",
   "summary": {
-    "group_count": 1308,
+    "group_count": 1316,
     "failure_count": 0,
     "failures_by_surface": {}
   },
@@ -230,6 +230,55 @@
       },
       "failures": []
     },
+    {
+      "name": "data/data_explorer_analysis.json",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/data/data_explorer_analysis.json",
+        "exists": true,
+        "bytes": 21206,
+        "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_artifacts_data": {
+          "path": "hf_artifacts:data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:docs/data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_model_data": {
+          "path": "hf_model:data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_model_docs_data": {
+          "path": "hf_model:docs/data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_model": {
+          "path": "hf_model:metrics/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "data/evidence_contract.json",
       "status": "pass",
@@ -1021,44 +1070,44 @@
         "path": "repo:docs/data/public_surface_qa.json",
         "exists": true,
         "bytes": 7690,
-        "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+        "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_artifacts_data": {
           "path": "hf_artifacts:data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_model_data": {
           "path": "hf_model:data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_model_docs_data": {
           "path": "hf_model:docs/data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_model": {
           "path": "hf_model:metrics/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         }
       },
       "failures": []
@@ -1217,44 +1266,44 @@
         "path": "repo:docs/data/quality_gates.json",
         "exists": true,
         "bytes": 8640,
-        "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+        "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_artifacts_data": {
           "path": "hf_artifacts:data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_model_data": {
           "path": "hf_model:data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_model_docs_data": {
           "path": "hf_model:docs/data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_model": {
           "path": "hf_model:metrics/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         }
       },
       "failures": []
@@ -2766,6 +2815,191 @@
       },
       "failures": []
     },
+    {
+      "name": "assets/charts/data_explorer_full_file_composition.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_full_file_composition.svg",
+        "exists": true,
+        "bytes": 49697,
+        "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_sample_action_distribution.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_sample_action_distribution.svg",
+        "exists": true,
+        "bytes": 59497,
+        "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_sample_feature_modalities.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_sample_feature_modalities.svg",
+        "exists": true,
+        "bytes": 48782,
+        "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_scope_ladder.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_scope_ladder.svg",
+        "exists": true,
+        "bytes": 94443,
+        "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_selected128_split_rows.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_selected128_split_rows.svg",
+        "exists": true,
+        "bytes": 50637,
+        "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "assets/charts/two_evidence_line_map.svg",
       "status": "pass",
@@ -6518,6 +6752,31 @@
       },
       "failures": []
     },
+    {
+      "name": "scripts/build_data_explorer_analysis.py",
+      "status": "pass",
+      "local": {
+        "path": "repo:scripts/build_data_explorer_analysis.py",
+        "exists": true,
+        "bytes": 26652,
+        "sha256": "10ce20f5e3dcefbd3a81bb649de0f2cea335a3ffc15fe58334a3f20b8007ac9d"
+      },
+      "mirrors": {
+        "hf_artifacts": {
+          "path": "hf_artifacts:scripts/build_data_explorer_analysis.py",
+          "exists": true,
+          "bytes": 26652,
+          "sha256": "10ce20f5e3dcefbd3a81bb649de0f2cea335a3ffc15fe58334a3f20b8007ac9d"
+        },
+        "hf_model": {
+          "path": "hf_model:scripts/build_data_explorer_analysis.py",
+          "exists": true,
+          "bytes": 26652,
+          "sha256": "10ce20f5e3dcefbd3a81bb649de0f2cea335a3ffc15fe58334a3f20b8007ac9d"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "scripts/build_evaluation_protocol.py",
       "status": "pass",
@@ -7124,21 +7383,21 @@
       "local": {
         "path": "repo:scripts/validate_mirror_parity.py",
         "exists": true,
-        "bytes": 38964,
-        "sha256": "c9cf04089e331142fa95950258cb5124089c37ddff49ce43d3db06e2aafa0045"
+        "bytes": 39342,
+        "sha256": "1be1241cac3e43ee79946db68bf74bbac188b637cfc55fb1ff2ddd4225960ecd"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 38964,
-          "sha256": "c9cf04089e331142fa95950258cb5124089c37ddff49ce43d3db06e2aafa0045"
+          "bytes": 39342,
+          "sha256": "1be1241cac3e43ee79946db68bf74bbac188b637cfc55fb1ff2ddd4225960ecd"
         },
         "hf_model": {
           "path": "hf_model:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 38964,
-          "sha256": "c9cf04089e331142fa95950258cb5124089c37ddff49ce43d3db06e2aafa0045"
+          "bytes": 39342,
+          "sha256": "1be1241cac3e43ee79946db68bf74bbac188b637cfc55fb1ff2ddd4225960ecd"
         }
       },
       "failures": []
@@ -32496,6 +32755,37 @@
       },
       "failures": []
     },
+    {
+      "name": "docs/DATA_EXPLORER_ANALYSIS.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:DATA_EXPLORER_ANALYSIS.md",
+        "exists": true,
+        "bytes": 1925,
+        "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:DATA_EXPLORER_ANALYSIS.md",
+          "exists": true,
+          "bytes": 1925,
+          "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:DATA_EXPLORER_ANALYSIS.md",
+          "exists": true,
+          "bytes": 1925,
+          "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+        },
+        "hf_model": {
+          "path": "hf_model:DATA_EXPLORER_ANALYSIS.md",
+          "exists": true,
+          "bytes": 1925,
+          "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "docs/OMNI_MODEL_EXTENSION_CONTRACT.md",
       "status": "pass",

data/public_surface_qa.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Public Project Surface",
   "status": "pass",
-  "generated_at_utc": "2026-06-23T09:41:03+00:00",
+  "generated_at_utc": "2026-06-23T10:03:30+00:00",
   "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
   "checks": [
     {
@@ -48,7 +48,7 @@
         "mirror_parity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-23T09:40:30+00:00"
+          "generated_at_utc": "2026-06-23T10:02:51+00:00"
         }
       },
       "failures": {}

data/quality_gates.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Release Checks",
   "status": "pass",
-  "generated_at_utc": "2026-06-23T09:41:03+00:00",
+  "generated_at_utc": "2026-06-23T10:03:30+00:00",
   "rule": "A release is current when the automated reports pass and the live GitHub/Hugging Face mirrors are verified after publishing.",
   "automated_gates": [
     {

docs/data/data_explorer_analysis.json

@@ -0,0 +1,782 @@
+{
+  "charts": [
+    {
+      "path": "assets/charts/data_explorer_scope_ladder.svg",
+      "question": "How do the public sample, selected 128 episodes, and full gated dataset differ in scale?",
+      "title": "Scope ladder"
+    },
+    {
+      "path": "assets/charts/data_explorer_sample_feature_modalities.svg",
+      "question": "Which modality groups dominate the one-sample task input?",
+      "title": "Public sample feature dimensions"
+    },
+    {
+      "path": "assets/charts/data_explorer_sample_action_distribution.svg",
+      "question": "Which action labels occupy the most 20-frame windows in the sample?",
+      "title": "Public sample action distribution"
+    },
+    {
+      "path": "assets/charts/data_explorer_selected128_split_rows.svg",
+      "question": "How many rows are available per selected-128 export and split?",
+      "title": "Selected-128 split rows"
+    },
+    {
+      "path": "assets/charts/data_explorer_full_file_composition.svg",
+      "question": "What file types dominate the gated full-dataset metadata?",
+      "title": "Full dataset file composition"
+    }
+  ],
+  "full_hf_dataset": {
+    "annotation_file_size_summary": {
+      "count": 12103,
+      "max_bytes": 2000432556,
+      "max_human": "1.86 GiB",
+      "mean_bytes": 1830083732,
+      "mean_human": "1.70 GiB",
+      "median_bytes": 1969457752,
+      "median_human": "1.83 GiB",
+      "min_bytes": 6687192,
+      "min_human": "6.38 MiB",
+      "p25_bytes": 1869556164,
+      "p75_bytes": 1989975784
+    },
+    "basename_counts": {
+      "README.md": 1,
+      "annotation.hdf5": 12103,
+      "fisheye_cam0.mp4": 12102,
+      "fisheye_cam1.mp4": 12102,
+      "fisheye_cam2.mp4": 12102,
+      "fisheye_cam3.mp4": 12102,
+      "stereo_left.mp4": 12102,
+      "stereo_right.mp4": 12102,
+      "visualization.rrd": 541
+    },
+    "card_data": {
+      "extra_gated_button_content": "I have signed the access control agreement",
+      "extra_gated_description": "Access is reviewed manually and is limited to approved non-commercial use. Completion of an external agreement-signing step may be required before approval. Please sign the agreement via [DocuSign](https://ropedia.docsend.com/view/ra7ej7gs6s98sw87)",
+      "extra_gated_heading": "Request controlled access to Xperience-10M",
+      "language": [
+        "en"
+      ],
+      "license": "other",
+      "pretty_name": "Xperience-10M",
+      "size_categories": [
+        "1M<n<10M"
+      ],
+      "tags": [
+        "egocentric",
+        "first-person",
+        "multimodal",
+        "3d",
+        "4d",
+        "embodied-ai",
+        "robotics",
+        "human-motion",
+        "mocap",
+        "imu",
+        "audio",
+        "depth",
+        "captions",
+        "video"
+      ],
+      "task_categories": [
+        "video-classification",
+        "image-to-text",
+        "depth-estimation",
+        "robotics"
+      ]
+    },
+    "complete_episode_training_size_summary": {
+      "count": 12102,
+      "max_bytes": 2721499520,
+      "max_human": "2.53 GiB",
+      "mean_bytes": 2237801505,
+      "mean_human": "2.08 GiB",
+      "median_bytes": 2365869284,
+      "median_human": "2.20 GiB",
+      "min_bytes": 8158968,
+      "min_human": "7.78 MiB",
+      "p25_bytes": 2286228606,
+      "p75_bytes": 2412829381
+    },
+    "episode_count_per_session_summary": {
+      "count": 803,
+      "max": 51,
+      "mean": 15.07,
+      "median": 11,
+      "min": 1,
+      "p25": 8,
+      "p75": 13
+    },
+    "episode_size_summary": {
+      "count": 12103,
+      "max_bytes": 5108045286,
+      "max_human": "4.76 GiB",
+      "mean_bytes": 2318233840,
+      "mean_human": "2.16 GiB",
+      "median_bytes": 2369826102,
+      "median_human": "2.21 GiB",
+      "min_bytes": 8158968,
+      "min_human": "7.78 MiB",
+      "p25_bytes": 2292986495,
+      "p75_bytes": 2420801956
+    },
+    "file_type_counts": {
+      ".hdf5": 12103,
+      ".md": 1,
+      ".mp4": 72612,
+      ".rrd": 541
+    },
+    "gated": "manual",
+    "incomplete_episode_records": [
+      {
+        "episode_id": "ep1",
+        "episode_path": "dc3f4139-f499-4de7-b057-e25b7dfb2d83/ep1",
+        "file_count": 1,
+        "has_all_six_videos": false,
+        "has_annotation": true,
+        "has_fisheye_cam0": false,
+        "has_visualization_rrd": false,
+        "is_complete": false,
+        "is_degraded_valid": false,
+        "missing_required_files": [
+          "fisheye_cam0.mp4",
+          "fisheye_cam1.mp4",
+          "fisheye_cam2.mp4",
+          "fisheye_cam3.mp4",
+          "stereo_left.mp4",
+          "stereo_right.mp4"
+        ],
+        "top_level_session": "dc3f4139-f499-4de7-b057-e25b7dfb2d83",
+        "total_bytes": 1418232696,
+        "training_bytes_excluding_visualization_rrd": 1418232696,
+        "video_count": 0
+      }
+    ],
+    "last_modified": "2026-04-21T05:03:45+00:00",
+    "metadata_note": "The official dataset is gated. These full-corpus figures use authenticated Hugging Face Hub file metadata only; they do not inspect private row content or redistribute raw MP4/HDF5/RRD files.",
+    "pilot_scale_estimates": {
+      "all_complete_episodes_windows_at_256_each": 3098112,
+      "episode_100_windows_at_256_each": 25600,
+      "episode_32_windows_at_256_each": 8192,
+      "episode_500_windows_at_256_each": 128000,
+      "mean_based_32_episode_training_bytes": 71609648160,
+      "mean_based_32_episode_training_human": "66.69 GiB",
+      "median_based_32_episode_training_bytes": 75707817088,
+      "median_based_32_episode_training_human": "70.51 GiB",
+      "windows_per_episode": 256
+    },
+    "repo_id": "ropedia-ai/xperience-10m",
+    "repo_sha": "ce943cf271a758b60240084892d05cf6dc12dd90",
+    "summary": {
+      "all_complete_episode_training_bytes_excluding_visualization_rrd": 27081873816281,
+      "all_complete_episode_training_human_excluding_visualization_rrd": "24.63 TiB",
+      "annotation_hdf5_count": 12103,
+      "complete_episode_count": 12102,
+      "complete_episode_pct": 99.9917,
+      "complete_sessions": 802,
+      "degraded_valid_episode_count": 12102,
+      "degraded_valid_episode_pct": 99.9917,
+      "degraded_valid_sessions": 802,
+      "episode_like_folder_count": 12103,
+      "file_count_excluding_gitattributes": 85257,
+      "mp4_count": 72612,
+      "sibling_count": 85258,
+      "top_level_session_count": 804,
+      "total_bytes_from_file_metadata": 28057584187079,
+      "total_human_from_file_metadata": "25.52 TiB",
+      "training_bytes_excluding_visualization_rrd": 27083292060675,
+      "training_human_excluding_visualization_rrd": "24.63 TiB",
+      "visualization_rrd_bytes": 974292126404,
+      "visualization_rrd_count": 541,
+      "visualization_rrd_human": "907.38 GiB"
+    },
+    "video_count_histogram": {
+      "0": 1,
+      "6": 12102
+    }
+  },
+  "generated_at_utc": "2026-06-23T09:35:08Z",
+  "public_sample": {
+    "dataset": {
+      "episode_id": "xperience-10m-sample",
+      "hosting_policy": "Raw files are streamed or downloaded from the official public sample dataset. This task-suite repo mirrors derived artifacts and metadata only.",
+      "license": "cc-by-nc-4.0",
+      "repo_commit": "2d80b5af8085a18ffe702e1420968308007f6204",
+      "repo_id": "ropedia-ai/xperience-10m-sample",
+      "repo_url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m-sample"
+    },
+    "feature_blocks": [
+      {
+        "dim": 1092,
+        "display": "Body Joints",
+        "modality": "motion_capture",
+        "name": "body_joints"
+      },
+      {
+        "dim": 980,
+        "display": "Depth + Confidence",
+        "modality": "depth",
+        "name": "depth_confidence"
+      },
+      {
+        "dim": 896,
+        "display": "Language Text",
+        "modality": "language",
+        "name": "caption_objects_interaction_text"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam0",
+        "modality": "video",
+        "name": "video_fisheye_cam0"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam1",
+        "modality": "video",
+        "name": "video_fisheye_cam1"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam2",
+        "modality": "video",
+        "name": "video_fisheye_cam2"
+      },
+      {
+        "dim": 686,
+        "display": "Video Fisheye Cam3",
+        "modality": "video",
+        "name": "video_fisheye_cam3"
+      },
+      {
+        "dim": 686,
+        "display": "Video Stereo Left",
+        "modality": "video",
+        "name": "video_stereo_left"
+      },
+      {
+        "dim": 686,
+        "display": "Video Stereo Right",
+        "modality": "video",
+        "name": "video_stereo_right"
+      },
+      {
+        "dim": 441,
+        "display": "Left Hand",
+        "modality": "motion_capture",
+        "name": "hand_left_joints"
+      },
+      {
+        "dim": 441,
+        "display": "Right Hand",
+        "modality": "motion_capture",
+        "name": "hand_right_joints"
+      },
+      {
+        "dim": 168,
+        "display": "Audio",
+        "modality": "audio",
+        "name": "audio_fisheye_cam0_aac"
+      },
+      {
+        "dim": 147,
+        "display": "Body Contacts",
+        "modality": "motion_capture",
+        "name": "body_contacts"
+      },
+      {
+        "dim": 117,
+        "display": "Calibration",
+        "modality": "pose_slam",
+        "name": "calibration"
+      },
+      {
+        "dim": 63,
+        "display": "Camera Rotation",
+        "modality": "pose_slam",
+        "name": "camera_rotation_matrix"
+      },
+      {
+        "dim": 42,
+        "display": "IMU Accel/Gyro",
+        "modality": "inertial",
+        "name": "imu_accel_gyro"
+      },
+      {
+        "dim": 22,
+        "display": "SLAM Point Cloud",
+        "modality": "pose_slam",
+        "name": "slam_point_cloud"
+      },
+      {
+        "dim": 21,
+        "display": "Camera Translation",
+        "modality": "pose_slam",
+        "name": "camera_translation"
+      }
+    ],
+    "feature_dim_by_modality": {
+      "audio": 168,
+      "depth": 980,
+      "inertial": 42,
+      "language": 896,
+      "motion_capture": 2121,
+      "pose_slam": 223,
+      "video": 4116
+    },
+    "file_bytes_by_kind": {
+      "hdf5_annotation_container": {
+        "bytes": 1931496028,
+        "count": 1,
+        "human": "1.80 GiB"
+      },
+      "rerun_viewer_recording": {
+        "bytes": 2702924036,
+        "count": 1,
+        "human": "2.52 GiB"
+      },
+      "video": {
+        "bytes": 383606257,
+        "count": 5,
+        "human": "365.84 MiB"
+      },
+      "video_with_audio": {
+        "bytes": 89842251,
+        "count": 1,
+        "human": "85.68 MiB"
+      }
+    },
+    "file_count": 8,
+    "hdf5_groups": [
+      {
+        "description": "Camera intrinsics/extrinsics and static calibration values used to align sensor streams.",
+        "group": "calibration"
+      },
+      {
+        "description": "JSON text with task config, action segments, object mentions, interaction descriptions, and global summary.",
+        "group": "caption"
+      },
+      {
+        "description": "Depth maps and confidence channels aligned to the episode timeline.",
+        "group": "depth"
+      },
+      {
+        "description": "Full-body joint and contact signals for human motion modeling.",
+        "group": "full_body_mocap"
+      },
+      {
+        "description": "Left and right hand joint trajectories used by hand-motion and forecast tasks.",
+        "group": "hand_mocap"
+      },
+      {
+        "description": "Accelerometer and gyroscope streams sampled at higher frequency than the video frames.",
+        "group": "imu"
+      },
+      {
+        "description": "Episode metadata, frame indexing, and source bookkeeping.",
+        "group": "metadata"
+      },
+      {
+        "description": "Camera trajectory, pose, and sparse SLAM point-cloud information.",
+        "group": "slam"
+      },
+      {
+        "description": "Video stream metadata and per-frame alignment information.",
+        "group": "video"
+      }
+    ],
+    "object_vocab_count": 34,
+    "segment_count": 35,
+    "source_policy": "Window-level labels, features, predictions, and diagnostics are embedded here. Official raw MP4/HDF5/RRD files are linked from the Raw Sample Browser, with compact browser-preview clips for immediate MP4/audio playback.",
+    "top_actions": [
+      {
+        "count": 159,
+        "name": "Hold gooseneck kettle"
+      },
+      {
+        "count": 159,
+        "name": "Pour coffee"
+      },
+      {
+        "count": 127,
+        "name": "Position kettle to pour"
+      },
+      {
+        "count": 95,
+        "name": "Lift gooseneck kettle"
+      },
+      {
+        "count": 95,
+        "name": "Close bottle cap"
+      },
+      {
+        "count": 95,
+        "name": "Wait/Prepare for pouring"
+      },
+      {
+        "count": 87,
+        "name": "Transfer coffee to dripper"
+      },
+      {
+        "count": 63,
+        "name": "Grasp coffee scoop"
+      },
+      {
+        "count": 47,
+        "name": "Hold coffee carafe"
+      },
+      {
+        "count": 39,
+        "name": "Move kettle"
+      },
+      {
+        "count": 31,
+        "name": "Move kettle away"
+      },
+      {
+        "count": 30,
+        "name": "Pick up kettle"
+      }
+    ],
+    "top_objects": [
+      {
+        "count": 636,
+        "name": "coffee dripper"
+      },
+      {
+        "count": 517,
+        "name": "scale"
+      },
+      {
+        "count": 446,
+        "name": "gooseneck kettle"
+      },
+      {
+        "count": 333,
+        "name": "digital scale"
+      },
+      {
+        "count": 238,
+        "name": "water bottle"
+      },
+      {
+        "count": 217,
+        "name": "white mug"
+      },
+      {
+        "count": 203,
+        "name": "white bottle"
+      },
+      {
+        "count": 170,
+        "name": "coffee mug"
+      },
+      {
+        "count": 158,
+        "name": "bottle"
+      },
+      {
+        "count": 145,
+        "name": "dripper"
+      },
+      {
+        "count": 126,
+        "name": "glass carafe"
+      },
+      {
+        "count": 123,
+        "name": "kettle"
+      }
+    ],
+    "top_subtasks": [
+      {
+        "count": 166,
+        "name": "Handle gooseneck kettle"
+      },
+      {
+        "count": 151,
+        "name": "Pour coffee"
+      },
+      {
+        "count": 150,
+        "name": "Pick up and position kettle"
+      },
+      {
+        "count": 119,
+        "name": "Prepare for pouring"
+      },
+      {
+        "count": 111,
+        "name": "Prepare coffee equipment and scoop grounds"
+      },
+      {
+        "count": 111,
+        "name": "Lift gooseneck kettle"
+      },
+      {
+        "count": 111,
+        "name": "Pour and close white bottle"
+      },
+      {
+        "count": 79,
+        "name": "Transfer coffee grounds to dripper"
+      },
+      {
+        "count": 79,
+        "name": "Set down kettle and retrieve white bottle"
+      },
+      {
+        "count": 39,
+        "name": "Move kettle"
+      },
+      {
+        "count": 14,
+        "name": ""
+      },
+      {
+        "count": 10,
+        "name": "Pour milk into coffee"
+      }
+    ],
+    "total_bytes": 5107868572,
+    "total_human": "4.76 GiB",
+    "windowization": {
+      "audio_sample_rate": 16000,
+      "duration_human": "4.85 min",
+      "duration_sec": 291.03000343524565,
+      "feature_dim": 8546,
+      "fps_observed": 20.00137419266181,
+      "note": "The public task suite converts the raw episode into overlapping 20-frame windows with a 5-frame stride.",
+      "num_frames": 5821,
+      "num_windows": 1161,
+      "stride_frames": 5,
+      "window_frames": 20
+    }
+  },
+  "selected_128": {
+    "bytes_by_size_band": {
+      "long": {
+        "bytes": 75928334176,
+        "human": "70.71 GiB"
+      },
+      "lower_mid": {
+        "bytes": 75730496234,
+        "human": "70.53 GiB"
+      },
+      "short": {
+        "bytes": 70693191518,
+        "human": "65.84 GiB"
+      },
+      "upper_mid": {
+        "bytes": 75837737335,
+        "human": "70.63 GiB"
+      }
+    },
+    "bytes_by_split": {
+      "test": {
+        "bytes": 37243648205,
+        "human": "34.69 GiB"
+      },
+      "train": {
+        "bytes": 223733297061,
+        "human": "208.37 GiB"
+      },
+      "val": {
+        "bytes": 37212813997,
+        "human": "34.66 GiB"
+      }
+    },
+    "exports": [
+      {
+        "episodes": 119,
+        "key": "sparse_export",
+        "label": "Sparse selected-128 export",
+        "samples": 3808,
+        "scale_counts": {},
+        "split_counts": {
+          "test": 448,
+          "train": 2848,
+          "val": 512
+        }
+      },
+      {
+        "episodes": 119,
+        "key": "qwen_v6_multiscale_export",
+        "label": "Qwen3-Omni v6 multiscale JSONL",
+        "samples": 34269,
+        "scale_counts": {},
+        "split_counts": {
+          "test": 4032,
+          "train": 25629,
+          "val": 4608
+        }
+      },
+      {
+        "episodes": 119,
+        "key": "dense_multiscale_compact_export",
+        "label": "Dense multiscale compact export",
+        "samples": 106095,
+        "scale_counts": {
+          "dense_20f_stride10": 60725,
+          "long_80f_stride40": 15067,
+          "medium_40f_stride20": 30303
+        },
+        "split_counts": {
+          "test": 11795,
+          "train": 79593,
+          "val": 14707
+        }
+      }
+    ],
+    "metadata_matrix_sparse": {
+      "bytes": 217477,
+      "feature_dim": 394,
+      "human": "212.38 KiB",
+      "row_count": 3808,
+      "sha256": "cdaa55ea3bb229d70bee845f05628ed7a984dd46f95080358d51943b4b8975df",
+      "split_counts": {
+        "test": 448,
+        "train": 2848,
+        "val": 512
+      }
+    },
+    "metadata_matrix_v2": {
+      "bytes": 3375193,
+      "feature_dim": 394,
+      "human": "3.22 MiB",
+      "row_count": 34269,
+      "sha256": "3c2a23b681c9f257a4c04710c72441650844ede6279d3e67be57ccc3abd4fa90",
+      "split_counts": {
+        "test": 4032,
+        "train": 25629,
+        "val": 4608
+      }
+    },
+    "official_dataset": {
+      "access_note": "Episode directory pages are browsable on Hugging Face. Raw annotation/video file downloads require access to the gated official dataset and are not redistributed here.",
+      "gated_resolve_base_url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m/resolve/main",
+      "repo_id": "ropedia-ai/xperience-10m",
+      "repo_sha": "ce943cf271a758b60240084892d05cf6dc12dd90",
+      "tree_base_url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m/tree/main"
+    },
+    "raw20_proxy_tasks": [
+      "camera_view_sync_retrieval",
+      "interaction_text_prediction"
+    ],
+    "raw20_result_records": 40,
+    "selection_summary": {
+      "one_episode_per_top_level_session": true,
+      "selected_download_size_excluding_visualization_rrd_bytes": 298189759263,
+      "selected_episode_count": 128,
+      "selected_split_counts": {
+        "test": 16,
+        "train": 96,
+        "val": 16
+      },
+      "size_band_counts": {
+        "long": 32,
+        "lower_mid": 32,
+        "short": 32,
+        "upper_mid": 32
+      }
+    },
+    "size_band_counts": {
+      "long": 32,
+      "lower_mid": 32,
+      "short": 32,
+      "upper_mid": 32
+    },
+    "sparse_main_task_counts": [
+      {
+        "count": 96,
+        "name": "Stocking grocery shelves"
+      },
+      {
+        "count": 96,
+        "name": "Stocking grocery store shelves"
+      },
+      {
+        "count": 96,
+        "name": "Measuring and marking cardboard"
+      },
+      {
+        "count": 64,
+        "name": "Stocking supermarket shelves"
+      },
+      {
+        "count": 64,
+        "name": "Stocking retail store shelves"
+      },
+      {
+        "count": 64,
+        "name": "Assembling a jigsaw puzzle"
+      },
+      {
+        "count": 64,
+        "name": "Crafting cardboard shapes"
+      },
+      {
+        "count": 64,
+        "name": "Stocking shelves in a retail store"
+      },
+      {
+        "count": 64,
+        "name": "Playing Mahjong"
+      },
+      {
+        "count": 64,
+        "name": "Stocking canned goods on a supermarket shelf"
+      },
+      {
+        "count": 32,
+        "name": "Sorting and organizing colorful buttons on a table"
+      },
+      {
+        "count": 32,
+        "name": "Crafting cardboard shapes and preparing workspace"
+      }
+    ],
+    "sparse_windows_by_split": {
+      "test": 448,
+      "train": 2848,
+      "val": 512
+    },
+    "split_episode_counts": {
+      "test": 16,
+      "train": 96,
+      "val": 16
+    }
+  },
+  "sources": [
+    {
+      "path": "docs/data/raw_sample_files.json",
+      "scope": "public_sample"
+    },
+    {
+      "path": "docs/data/single_episode_explorer.json",
+      "scope": "public_sample"
+    },
+    {
+      "path": "results/episode_task_suite/windows.csv",
+      "scope": "public_sample"
+    },
+    {
+      "path": "docs/data/xperience10m_128_episode_feature_index.json",
+      "scope": "selected_128"
+    },
+    {
+      "path": "results/omni_finetune/xperience10m_128_episode_selection.csv",
+      "scope": "selected_128"
+    },
+    {
+      "path": "results/omni_finetune/multi_episode_128_task_baselines/windows.csv",
+      "scope": "selected_128"
+    },
+    {
+      "path": "results/omni_finetune/full_dataset_metadata_audit.json",
+      "scope": "full_hf_dataset"
+    }
+  ],
+  "status": "pass"
+}

docs/data/mirror_parity.json

@@ -1,9 +1,9 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-23T09:40:30+00:00",
+  "generated_at_utc": "2026-06-23T10:02:51+00:00",
   "hf_root": "hf_publish",
   "summary": {
-    "group_count": 1308,
+    "group_count": 1316,
     "failure_count": 0,
     "failures_by_surface": {}
   },
@@ -230,6 +230,55 @@
       },
       "failures": []
     },
+    {
+      "name": "data/data_explorer_analysis.json",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/data/data_explorer_analysis.json",
+        "exists": true,
+        "bytes": 21206,
+        "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_artifacts_data": {
+          "path": "hf_artifacts:data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:docs/data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_model_data": {
+          "path": "hf_model:data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_model_docs_data": {
+          "path": "hf_model:docs/data/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        },
+        "hf_model": {
+          "path": "hf_model:metrics/data_explorer_analysis.json",
+          "exists": true,
+          "bytes": 21206,
+          "sha256": "84fce408007f34bf4574fdfd39d14e738e32e243654634419259074c8db4f81a"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "data/evidence_contract.json",
       "status": "pass",
@@ -1021,44 +1070,44 @@
         "path": "repo:docs/data/public_surface_qa.json",
         "exists": true,
         "bytes": 7690,
-        "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+        "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_artifacts_data": {
           "path": "hf_artifacts:data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_model_data": {
           "path": "hf_model:data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_model_docs_data": {
           "path": "hf_model:docs/data/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         },
         "hf_model": {
           "path": "hf_model:metrics/public_surface_qa.json",
           "exists": true,
           "bytes": 7690,
-          "sha256": "0574b2c35ffb577979f11c2c5e27f1086703cccc2a6e1e91277fb2d4a6376e28"
+          "sha256": "e8a581d2de654cc7b02b293123dcb6ec76fd32d451a3e145d8f2502a63a524ab"
         }
       },
       "failures": []
@@ -1217,44 +1266,44 @@
         "path": "repo:docs/data/quality_gates.json",
         "exists": true,
         "bytes": 8640,
-        "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+        "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_artifacts_data": {
           "path": "hf_artifacts:data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_model_data": {
           "path": "hf_model:data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_model_docs_data": {
           "path": "hf_model:docs/data/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         },
         "hf_model": {
           "path": "hf_model:metrics/quality_gates.json",
           "exists": true,
           "bytes": 8640,
-          "sha256": "515ce3da5b1920be2d5d711dd0fb09802e30ae409450307c1268c7e0ae9c1eb8"
+          "sha256": "9178b10176d4c2157249c483e2739b9d3530abe5e760d2ff9dec4468e868a4ab"
         }
       },
       "failures": []
@@ -2766,6 +2815,191 @@
       },
       "failures": []
     },
+    {
+      "name": "assets/charts/data_explorer_full_file_composition.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_full_file_composition.svg",
+        "exists": true,
+        "bytes": 49697,
+        "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_full_file_composition.svg",
+          "exists": true,
+          "bytes": 49697,
+          "sha256": "d610de2826e3f0e39ba647e50791989fdbca28bd85ceff9edb1e30858493b0cd"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_sample_action_distribution.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_sample_action_distribution.svg",
+        "exists": true,
+        "bytes": 59497,
+        "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_sample_action_distribution.svg",
+          "exists": true,
+          "bytes": 59497,
+          "sha256": "9f4b3caeafbce47f66a9bca2985866f0820feb6e1fe7b9b7aa9dc3b57f3e50ed"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_sample_feature_modalities.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_sample_feature_modalities.svg",
+        "exists": true,
+        "bytes": 48782,
+        "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_sample_feature_modalities.svg",
+          "exists": true,
+          "bytes": 48782,
+          "sha256": "3089e8a3916caf6fcaad348e34d793643c0e211dd3fb93aea67de3324514a955"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_scope_ladder.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_scope_ladder.svg",
+        "exists": true,
+        "bytes": 94443,
+        "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_scope_ladder.svg",
+          "exists": true,
+          "bytes": 94443,
+          "sha256": "35eb717e2299cd8e0499f8a210255ac34edb0c6dcb58b3ced7082c5e2e45e82b"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "assets/charts/data_explorer_selected128_split_rows.svg",
+      "status": "pass",
+      "local": {
+        "path": "repo:docs/assets/charts/data_explorer_selected128_split_rows.svg",
+        "exists": true,
+        "bytes": 50637,
+        "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        },
+        "hf_artifacts_docs": {
+          "path": "hf_artifacts:docs/assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        },
+        "hf_artifacts_card": {
+          "path": "hf_artifacts:assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        },
+        "hf_model": {
+          "path": "hf_model:assets/charts/data_explorer_selected128_split_rows.svg",
+          "exists": true,
+          "bytes": 50637,
+          "sha256": "60a21c4528b5167d2d4856f010d74a86159d9df2f43ee1263240507888a41610"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "assets/charts/two_evidence_line_map.svg",
       "status": "pass",
@@ -6518,6 +6752,31 @@
       },
       "failures": []
     },
+    {
+      "name": "scripts/build_data_explorer_analysis.py",
+      "status": "pass",
+      "local": {
+        "path": "repo:scripts/build_data_explorer_analysis.py",
+        "exists": true,
+        "bytes": 26652,
+        "sha256": "10ce20f5e3dcefbd3a81bb649de0f2cea335a3ffc15fe58334a3f20b8007ac9d"
+      },
+      "mirrors": {
+        "hf_artifacts": {
+          "path": "hf_artifacts:scripts/build_data_explorer_analysis.py",
+          "exists": true,
+          "bytes": 26652,
+          "sha256": "10ce20f5e3dcefbd3a81bb649de0f2cea335a3ffc15fe58334a3f20b8007ac9d"
+        },
+        "hf_model": {
+          "path": "hf_model:scripts/build_data_explorer_analysis.py",
+          "exists": true,
+          "bytes": 26652,
+          "sha256": "10ce20f5e3dcefbd3a81bb649de0f2cea335a3ffc15fe58334a3f20b8007ac9d"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "scripts/build_evaluation_protocol.py",
       "status": "pass",
@@ -7124,21 +7383,21 @@
       "local": {
         "path": "repo:scripts/validate_mirror_parity.py",
         "exists": true,
-        "bytes": 38964,
-        "sha256": "c9cf04089e331142fa95950258cb5124089c37ddff49ce43d3db06e2aafa0045"
+        "bytes": 39342,
+        "sha256": "1be1241cac3e43ee79946db68bf74bbac188b637cfc55fb1ff2ddd4225960ecd"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 38964,
-          "sha256": "c9cf04089e331142fa95950258cb5124089c37ddff49ce43d3db06e2aafa0045"
+          "bytes": 39342,
+          "sha256": "1be1241cac3e43ee79946db68bf74bbac188b637cfc55fb1ff2ddd4225960ecd"
         },
         "hf_model": {
           "path": "hf_model:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 38964,
-          "sha256": "c9cf04089e331142fa95950258cb5124089c37ddff49ce43d3db06e2aafa0045"
+          "bytes": 39342,
+          "sha256": "1be1241cac3e43ee79946db68bf74bbac188b637cfc55fb1ff2ddd4225960ecd"
         }
       },
       "failures": []
@@ -32496,6 +32755,37 @@
       },
       "failures": []
     },
+    {
+      "name": "docs/DATA_EXPLORER_ANALYSIS.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:DATA_EXPLORER_ANALYSIS.md",
+        "exists": true,
+        "bytes": 1925,
+        "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:DATA_EXPLORER_ANALYSIS.md",
+          "exists": true,
+          "bytes": 1925,
+          "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:DATA_EXPLORER_ANALYSIS.md",
+          "exists": true,
+          "bytes": 1925,
+          "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+        },
+        "hf_model": {
+          "path": "hf_model:DATA_EXPLORER_ANALYSIS.md",
+          "exists": true,
+          "bytes": 1925,
+          "sha256": "7a87ca17e7137302aae5e5b70d1bed217734e1144442b8b517ff98f54ec6fc6f"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "docs/OMNI_MODEL_EXTENSION_CONTRACT.md",
       "status": "pass",

docs/data/public_surface_qa.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Public Project Surface",
   "status": "pass",
-  "generated_at_utc": "2026-06-23T09:41:03+00:00",
+  "generated_at_utc": "2026-06-23T10:03:30+00:00",
   "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
   "checks": [
     {
@@ -48,7 +48,7 @@
         "mirror_parity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-23T09:40:30+00:00"
+          "generated_at_utc": "2026-06-23T10:02:51+00:00"
         }
       },
       "failures": {}

docs/data/quality_gates.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Release Checks",
   "status": "pass",
-  "generated_at_utc": "2026-06-23T09:41:03+00:00",
+  "generated_at_utc": "2026-06-23T10:03:30+00:00",
   "rule": "A release is current when the automated reports pass and the live GitHub/Hugging Face mirrors are verified after publishing.",
   "automated_gates": [
     {

scripts/build_data_explorer_analysis.py

@@ -0,0 +1,600 @@
+#!/usr/bin/env python3
+"""Build reader-facing data exploration assets for Xperience-10M.
+
+The script intentionally separates three scopes:
+
+1. The official public sample episode mirrored in this repository.
+2. The selected 128-episode public-safe feature/export surface.
+3. The gated upstream Hugging Face dataset, inspected through Hub file metadata.
+
+It does not download or redistribute gated raw files.
+"""
+
+from __future__ import annotations
+
+import argparse
+import csv
+import json
+import os
+from collections import Counter
+from datetime import datetime, timezone
+from pathlib import Path
+from typing import Any, Iterable
+
+os.environ.setdefault("MPLCONFIGDIR", "/tmp/ropedia-matplotlib")
+os.environ.setdefault("XDG_CACHE_HOME", "/tmp/ropedia-cache")
+
+import matplotlib
+
+matplotlib.use("Agg")
+import matplotlib.pyplot as plt
+
+
+BG = "#020502"
+PANEL = "#071307"
+GRID = "#23341f"
+TEXT = "#f4f7ee"
+MUTED = "#b8c4b4"
+GREEN = "#c6ff92"
+GREEN_DARK = "#6fb03f"
+CYAN = "#67e8d1"
+BLUE = "#9bb8ff"
+GOLD = "#ffd166"
+PINK = "#f472b6"
+PURPLE = "#b084ff"
+
+
+def repo_root() -> Path:
+    return Path(__file__).resolve().parents[1]
+
+
+def read_json(path: Path) -> dict[str, Any]:
+    return json.loads(path.read_text())
+
+
+def read_csv_rows(path: Path) -> list[dict[str, str]]:
+    with path.open(newline="") as handle:
+        return list(csv.DictReader(handle))
+
+
+def write_json(path: Path, payload: dict[str, Any]) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_text(json.dumps(payload, indent=2, sort_keys=True) + "\n")
+
+
+def human_bytes(value: int | float | None) -> str:
+    if value is None:
+        return "n/a"
+    value = float(value)
+    units = ["B", "KiB", "MiB", "GiB", "TiB", "PiB"]
+    index = 0
+    while value >= 1024 and index < len(units) - 1:
+        value /= 1024
+        index += 1
+    if index == 0:
+        return f"{int(value):,} {units[index]}"
+    return f"{value:,.2f} {units[index]}"
+
+
+def pct(numer: float, denom: float) -> float:
+    return 0.0 if denom == 0 else 100.0 * numer / denom
+
+
+def top_items(counter: Counter[str], limit: int = 10) -> list[dict[str, Any]]:
+    return [
+        {"name": name, "count": int(count)}
+        for name, count in counter.most_common(limit)
+    ]
+
+
+def style_axis(ax: plt.Axes) -> None:
+    ax.set_facecolor(PANEL)
+    ax.tick_params(colors=MUTED, labelsize=9)
+    for spine in ax.spines.values():
+        spine.set_color(GRID)
+    ax.grid(True, axis="x", color=GRID, alpha=0.55, linewidth=0.8)
+    ax.set_axisbelow(True)
+
+
+def save_figure(fig: plt.Figure, path: Path) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    fig.savefig(path, format="svg", bbox_inches="tight", facecolor=BG)
+    plt.close(fig)
+    path.write_text(
+        "\n".join(line.rstrip() for line in path.read_text(encoding="utf-8").splitlines()) + "\n",
+        encoding="utf-8",
+    )
+
+
+def add_value_labels(ax: plt.Axes, bars: Iterable[Any], formatter=str, pad: float = 0.02) -> None:
+    xmax = ax.get_xlim()[1]
+    for bar in bars:
+        width = bar.get_width()
+        label = formatter(width)
+        ax.text(
+            width + xmax * pad,
+            bar.get_y() + bar.get_height() / 2,
+            label,
+            va="center",
+            ha="left",
+            color=TEXT,
+            fontsize=9,
+            fontweight="bold",
+        )
+
+
+def build_public_sample(root: Path) -> dict[str, Any]:
+    raw = read_json(root / "docs/data/raw_sample_files.json")
+    explorer = read_json(root / "docs/data/single_episode_explorer.json")
+    windows = read_csv_rows(root / "results/episode_task_suite/windows.csv")
+
+    files = raw.get("files", [])
+    file_bytes_by_kind: Counter[str] = Counter()
+    file_count_by_kind: Counter[str] = Counter()
+    for item in files:
+        kind = str(item.get("kind", "other"))
+        file_count_by_kind[kind] += 1
+        file_bytes_by_kind[kind] += int(item.get("bytes", 0) or 0)
+
+    feature_by_modality: Counter[str] = Counter()
+    feature_blocks = []
+    for block in explorer.get("feature_blocks", []):
+        modality = str(block.get("modality", "other"))
+        dim = int(block.get("dim", 0) or 0)
+        feature_by_modality[modality] += dim
+        feature_blocks.append(
+            {
+                "name": block.get("name"),
+                "display": block.get("display"),
+                "modality": modality,
+                "dim": dim,
+            }
+        )
+
+    action_counts = Counter(row.get("action_label", "unknown") for row in windows)
+    subtask_counts = Counter(row.get("subtask_label", "unknown") for row in windows)
+    object_counts: Counter[str] = Counter()
+    for row in explorer.get("windows", []):
+        for obj in row.get("objects", []) or []:
+            if obj:
+                object_counts[str(obj)] += 1
+
+    windowization = raw.get("windowization", {})
+    frames = int(windowization.get("num_frames", 0) or 0)
+    fps = float(windowization.get("fps_observed", 0.0) or 0.0)
+    duration_sec = frames / fps if fps > 0 else None
+
+    return {
+        "dataset": raw.get("dataset", {}),
+        "windowization": {
+            **windowization,
+            "duration_sec": duration_sec,
+            "duration_human": f"{duration_sec / 60.0:.2f} min" if duration_sec else "n/a",
+        },
+        "file_count": len(files),
+        "total_bytes": sum(int(item.get("bytes", 0) or 0) for item in files),
+        "total_human": human_bytes(sum(int(item.get("bytes", 0) or 0) for item in files)),
+        "file_bytes_by_kind": {
+            key: {"bytes": int(value), "human": human_bytes(value), "count": int(file_count_by_kind[key])}
+            for key, value in sorted(file_bytes_by_kind.items())
+        },
+        "hdf5_groups": raw.get("hdf5_organization", []),
+        "feature_dim_by_modality": dict(sorted(feature_by_modality.items(), key=lambda item: (-item[1], item[0]))),
+        "feature_blocks": sorted(feature_blocks, key=lambda item: (-item["dim"], item["display"] or item["name"] or "")),
+        "top_actions": top_items(action_counts, 12),
+        "top_subtasks": top_items(subtask_counts, 12),
+        "top_objects": top_items(object_counts, 12),
+        "segment_count": len(explorer.get("segments", [])),
+        "object_vocab_count": int(explorer.get("meta", {}).get("object_vocab_count", 0) or 0),
+        "source_policy": explorer.get("meta", {}).get("source_policy"),
+    }
+
+
+def build_selected_128(root: Path) -> dict[str, Any]:
+    feature_index = read_json(root / "docs/data/xperience10m_128_episode_feature_index.json")
+    selection_rows = read_csv_rows(root / "results/omni_finetune/xperience10m_128_episode_selection.csv")
+    sparse_windows = read_csv_rows(root / "results/omni_finetune/multi_episode_128_task_baselines/windows.csv")
+
+    split_episode_counts = Counter(row.get("split", "unknown") for row in selection_rows)
+    band_counts = Counter(row.get("size_band", "unknown") for row in selection_rows)
+    bytes_by_split: Counter[str] = Counter()
+    bytes_by_band: Counter[str] = Counter()
+    for row in selection_rows:
+        value = int(float(row.get("training_bytes_excluding_visualization_rrd", 0) or 0))
+        bytes_by_split[row.get("split", "unknown")] += value
+        bytes_by_band[row.get("size_band", "unknown")] += value
+
+    sparse_windows_by_split = Counter(row.get("split", "unknown") for row in sparse_windows)
+    main_task_counts = Counter(row.get("main_task", "unknown") for row in sparse_windows)
+
+    processed = feature_index.get("processed_summary", {})
+    export_rows = []
+    for key, label in [
+        ("sparse_export", "Sparse selected-128 export"),
+        ("qwen_v6_multiscale_export", "Qwen3-Omni v6 multiscale JSONL"),
+        ("dense_multiscale_compact_export", "Dense multiscale compact export"),
+    ]:
+        item = processed.get(key, {})
+        export_rows.append(
+            {
+                "key": key,
+                "label": label,
+                "episodes": int(item.get("num_episodes", 0) or 0),
+                "samples": int(item.get("num_samples", 0) or 0),
+                "split_counts": {k: int(v) for k, v in (item.get("split_counts", {}) or {}).items()},
+                "scale_counts": {k: int(v) for k, v in (item.get("scale_counts", {}) or {}).items()},
+            }
+        )
+
+    matrix = processed.get("metadata_matrix_v2", {})
+    sparse_matrix = processed.get("metadata_matrix_sparse", {})
+    return {
+        "official_dataset": feature_index.get("official_dataset", {}),
+        "selection_summary": feature_index.get("selection_summary", {}),
+        "split_episode_counts": dict(split_episode_counts),
+        "size_band_counts": dict(band_counts),
+        "bytes_by_split": {
+            key: {"bytes": int(value), "human": human_bytes(value)}
+            for key, value in sorted(bytes_by_split.items())
+        },
+        "bytes_by_size_band": {
+            key: {"bytes": int(value), "human": human_bytes(value)}
+            for key, value in sorted(bytes_by_band.items())
+        },
+        "sparse_windows_by_split": dict(sparse_windows_by_split),
+        "sparse_main_task_counts": top_items(main_task_counts, 12),
+        "exports": export_rows,
+        "metadata_matrix_v2": {
+            "row_count": int(matrix.get("row_count", 0) or 0),
+            "feature_dim": int(matrix.get("feature_dim", 0) or 0),
+            "bytes": int(matrix.get("bytes", 0) or 0),
+            "human": human_bytes(int(matrix.get("bytes", 0) or 0)),
+            "split_counts": {k: int(v) for k, v in (matrix.get("split_counts", {}) or {}).items()},
+            "sha256": matrix.get("sha256"),
+        },
+        "metadata_matrix_sparse": {
+            "row_count": int(sparse_matrix.get("row_count", 0) or 0),
+            "feature_dim": int(sparse_matrix.get("feature_dim", 0) or 0),
+            "bytes": int(sparse_matrix.get("bytes", 0) or 0),
+            "human": human_bytes(int(sparse_matrix.get("bytes", 0) or 0)),
+            "split_counts": {k: int(v) for k, v in (sparse_matrix.get("split_counts", {}) or {}).items()},
+            "sha256": sparse_matrix.get("sha256"),
+        },
+        "raw20_result_records": int(processed.get("raw20_result_records", 0) or 0),
+        "raw20_proxy_tasks": processed.get("raw20_proxy_tasks", []),
+    }
+
+
+def build_full_hf_dataset(root: Path) -> dict[str, Any]:
+    audit = read_json(root / "results/omni_finetune/full_dataset_metadata_audit.json")
+    summary = audit.get("summary", {})
+    return {
+        "repo_id": audit.get("repo_id"),
+        "repo_sha": audit.get("repo_sha"),
+        "gated": audit.get("gated"),
+        "last_modified": audit.get("last_modified"),
+        "card_data": audit.get("card_data", {}),
+        "summary": summary,
+        "file_type_counts": audit.get("file_type_counts", {}),
+        "basename_counts": audit.get("basename_counts", {}),
+        "video_count_histogram": audit.get("video_count_histogram", {}),
+        "episode_count_per_session_summary": audit.get("episode_count_per_session_summary", {}),
+        "episode_size_summary": audit.get("episode_size_summary", {}),
+        "annotation_file_size_summary": audit.get("annotation_file_size_summary", {}),
+        "complete_episode_training_size_summary": audit.get("complete_episode_training_size_summary", {}),
+        "incomplete_episode_records": audit.get("incomplete_episode_records", []),
+        "pilot_scale_estimates": audit.get("pilot_scale_estimates", {}),
+        "metadata_note": (
+            "The official dataset is gated. These full-corpus figures use authenticated "
+            "Hugging Face Hub file metadata only; they do not inspect private row content "
+            "or redistribute raw MP4/HDF5/RRD files."
+        ),
+    }
+
+
+def plot_scope_ladder(payload: dict[str, Any], out: Path) -> None:
+    sample = payload["public_sample"]
+    selected = payload["selected_128"]
+    full = payload["full_hf_dataset"]
+
+    labels = ["Sample", "Selected-128", "Full dataset"]
+    episodes = [
+        1,
+        selected["selection_summary"].get("selected_episode_count", 0),
+        full["summary"].get("episode_like_folder_count", 0),
+    ]
+    windows = [
+        sample["windowization"].get("num_windows", 0),
+        selected["metadata_matrix_v2"].get("row_count", 0),
+        full["pilot_scale_estimates"].get("all_complete_episodes_windows_at_256_each", 0),
+    ]
+    storage = [
+        sample["total_bytes"],
+        selected["selection_summary"].get("selected_download_size_excluding_visualization_rrd_bytes", 0),
+        full["summary"].get("training_bytes_excluding_visualization_rrd", 0),
+    ]
+
+    fig, axes = plt.subplots(1, 3, figsize=(14.5, 4.8), facecolor=BG)
+    specs = [
+        ("Episodes", episodes, lambda v: f"{int(v):,}"),
+        ("Window rows", windows, lambda v: f"{int(v):,}"),
+        ("Training bytes", storage, lambda v: human_bytes(v)),
+    ]
+    colors = [GREEN, CYAN, BLUE]
+    for ax, (title, values, formatter) in zip(axes, specs):
+        style_axis(ax)
+        safe_values = [max(float(v), 1.0) for v in values]
+        bars = ax.barh(labels, safe_values, color=colors, edgecolor=TEXT, linewidth=0.4)
+        ax.set_xscale("log")
+        ax.set_title(title, color=TEXT, fontsize=15, fontweight="bold", loc="left", pad=10)
+        ax.tick_params(axis="y", colors=TEXT, labelsize=10)
+        xmax = max(safe_values) * 3.8
+        ax.set_xlim(0.8, xmax)
+        for bar, raw_value in zip(bars, values):
+            raw_value = max(float(raw_value), 1.0)
+            if raw_value > 20:
+                label_x = raw_value / 1.16
+                ha = "right"
+                color = BG
+            else:
+                label_x = raw_value * 1.18
+                ha = "left"
+                color = TEXT
+            ax.text(
+                label_x,
+                bar.get_y() + bar.get_height() / 2,
+                formatter(raw_value),
+                va="center",
+                ha=ha,
+                color=color,
+                fontsize=9,
+                fontweight="bold",
+            )
+    fig.suptitle(
+        "Xperience-10M scope ladder",
+        color=TEXT,
+        fontsize=18,
+        fontweight="bold",
+        x=0.02,
+        y=0.99,
+        ha="left",
+    )
+    fig.subplots_adjust(left=0.08, right=0.985, top=0.79, bottom=0.18, wspace=0.34)
+    fig.text(
+        0.02,
+        0.02,
+        "Log-scale bars compare the one public sample, the selected-128 surface, and authenticated full-corpus file metadata.",
+        color=MUTED,
+        fontsize=10,
+    )
+    save_figure(fig, out)
+
+
+def plot_feature_breakdown(payload: dict[str, Any], out: Path) -> None:
+    values = payload["public_sample"]["feature_dim_by_modality"]
+    labels = list(values.keys())[::-1]
+    dims = [values[label] for label in labels]
+    colors = [GREEN, CYAN, BLUE, GOLD, PINK, PURPLE, GREEN_DARK, MUTED][: len(labels)]
+
+    fig, ax = plt.subplots(figsize=(11, 6.2), facecolor=BG)
+    style_axis(ax)
+    bars = ax.barh(labels, dims, color=colors[::-1], edgecolor=TEXT, linewidth=0.35)
+    ax.set_title("Public sample feature dimensions by modality", color=TEXT, fontsize=18, fontweight="bold", loc="left", pad=14)
+    ax.set_xlabel("Feature dimensions in the 8,546-D task input", color=MUTED)
+    ax.tick_params(axis="y", colors=TEXT, labelsize=10)
+    add_value_labels(ax, bars, lambda v: f"{int(v):,}")
+    save_figure(fig, out)
+
+
+def plot_action_distribution(payload: dict[str, Any], out: Path) -> None:
+    items = payload["public_sample"]["top_actions"][:10]
+    labels = [item["name"] for item in items][::-1]
+    counts = [item["count"] for item in items][::-1]
+
+    fig, ax = plt.subplots(figsize=(11.5, 6.4), facecolor=BG)
+    style_axis(ax)
+    bars = ax.barh(labels, counts, color=GREEN, edgecolor=TEXT, linewidth=0.35)
+    ax.set_title("Public sample action-window distribution", color=TEXT, fontsize=18, fontweight="bold", loc="left", pad=14)
+    ax.set_xlabel("20-frame windows carrying each action label", color=MUTED)
+    ax.tick_params(axis="y", colors=TEXT, labelsize=9)
+    add_value_labels(ax, bars, lambda v: f"{int(v):,}")
+    save_figure(fig, out)
+
+
+def plot_selected_split_windows(payload: dict[str, Any], out: Path) -> None:
+    exports = payload["selected_128"]["exports"]
+    split_order = ["train", "val", "test"]
+    split_colors = {"train": GREEN, "val": CYAN, "test": BLUE}
+    labels = [item["label"] for item in exports]
+    y_positions = range(len(exports))
+
+    fig, ax = plt.subplots(figsize=(12.5, 5.8), facecolor=BG)
+    style_axis(ax)
+    left = [0] * len(exports)
+    for split in split_order:
+        values = [int(item.get("split_counts", {}).get(split, 0) or 0) for item in exports]
+        bars = ax.barh(
+            list(y_positions),
+            values,
+            left=left,
+            color=split_colors[split],
+            label=split,
+            edgecolor=BG,
+            linewidth=0.4,
+        )
+        for index, (bar, value) in enumerate(zip(bars, values)):
+            if value:
+                ax.text(
+                    left[index] + value / 2,
+                    bar.get_y() + bar.get_height() / 2,
+                    f"{value:,}",
+                    va="center",
+                    ha="center",
+                    color=BG,
+                    fontsize=8,
+                    fontweight="bold",
+                )
+        left = [left_value + value for left_value, value in zip(left, values)]
+    for y, total in zip(y_positions, left):
+        ax.text(total * 1.01, y, f"{total:,}", va="center", ha="left", color=TEXT, fontsize=9, fontweight="bold")
+    ax.set_yticks(list(y_positions), labels)
+    ax.tick_params(axis="y", colors=TEXT, labelsize=9)
+    ax.set_xlabel("Rows / samples", color=MUTED)
+    ax.set_title("Selected-128 processed rows by split", color=TEXT, fontsize=18, fontweight="bold", loc="left", pad=14)
+    leg = ax.legend(loc="lower right", frameon=True, facecolor=PANEL, edgecolor=GRID, labelcolor=TEXT)
+    for text in leg.get_texts():
+        text.set_color(TEXT)
+    save_figure(fig, out)
+
+
+def plot_full_file_composition(payload: dict[str, Any], out: Path) -> None:
+    counts = payload["full_hf_dataset"]["basename_counts"]
+    ordered = [
+        ("annotation.hdf5", counts.get("annotation.hdf5", 0)),
+        ("all MP4 streams", payload["full_hf_dataset"]["summary"].get("mp4_count", 0)),
+        ("visualization.rrd", counts.get("visualization.rrd", 0)),
+        ("README.md", counts.get("README.md", 0)),
+    ]
+    labels = [name for name, _ in ordered][::-1]
+    values = [value for _, value in ordered][::-1]
+
+    fig, ax = plt.subplots(figsize=(10.5, 5.1), facecolor=BG)
+    style_axis(ax)
+    bars = ax.barh(labels, values, color=[MUTED, BLUE, CYAN, GREEN], edgecolor=TEXT, linewidth=0.35)
+    ax.set_xscale("log")
+    ax.set_xlabel("File count, log scale", color=MUTED)
+    ax.set_title("Full gated dataset file composition", color=TEXT, fontsize=18, fontweight="bold", loc="left", pad=14)
+    ax.tick_params(axis="y", colors=TEXT, labelsize=10)
+    ax.set_xlim(0.8, max(values) * 5)
+    for bar, value in zip(bars, values):
+        ax.text(max(value, 1) * 1.08, bar.get_y() + bar.get_height() / 2, f"{int(value):,}", va="center", ha="left", color=TEXT, fontsize=9, fontweight="bold")
+    save_figure(fig, out)
+
+
+def render_markdown(payload: dict[str, Any]) -> str:
+    sample = payload["public_sample"]
+    selected = payload["selected_128"]
+    full = payload["full_hf_dataset"]
+    lines = [
+        "# Ropedia Xperience-10M Data Explorer Analysis",
+        "",
+        f"Generated: {payload['generated_at_utc']}",
+        "",
+        "This report summarizes three data scopes without mixing them: the official public sample episode, the selected 128-episode public-safe feature surface, and authenticated metadata for the full gated Hugging Face dataset.",
+        "",
+        "## Scope Summary",
+        "",
+        "| Scope | Episodes | Rows / windows | Storage view | Notes |",
+        "|---|---:|---:|---:|---|",
+        f"| Public sample | 1 | {sample['windowization'].get('num_windows', 0):,} | {sample['total_human']} | Raw sample files are playable or source-linked. |",
+        f"| Selected 128 | {selected['selection_summary'].get('selected_episode_count', 0):,} | {selected['metadata_matrix_v2'].get('row_count', 0):,} | {human_bytes(selected['selection_summary'].get('selected_download_size_excluding_visualization_rrd_bytes', 0))} | Public-safe matrices and window manifests, not raw redistribution. |",
+        f"| Full HF dataset | {full['summary'].get('episode_like_folder_count', 0):,} episode-like folders | {full['pilot_scale_estimates'].get('all_complete_episodes_windows_at_256_each', 0):,} projected rows at 256/episode | {full['summary'].get('training_human_excluding_visualization_rrd', 'n/a')} | Gated upstream file metadata only. |",
+        "",
+        "## Public Sample",
+        "",
+        f"- {sample['windowization'].get('num_frames', 0):,} frames at about {sample['windowization'].get('fps_observed', 0):.2f} fps.",
+        f"- {sample['windowization'].get('num_windows', 0):,} aligned 20-frame windows with {sample['windowization'].get('stride_frames', 0)}-frame stride.",
+        f"- {sample['windowization'].get('feature_dim', 0):,} model-input dimensions across {len(sample['feature_dim_by_modality'])} modality groups.",
+        f"- {sample['segment_count']:,} action segments and {sample['object_vocab_count']:,} object labels in the derived explorer.",
+        "",
+        "## Selected 128 Episodes",
+        "",
+        f"- Split: train {selected['split_episode_counts'].get('train', 0)}, val {selected['split_episode_counts'].get('val', 0)}, test {selected['split_episode_counts'].get('test', 0)} episodes.",
+        f"- Size bands: {', '.join(f'{k} {v}' for k, v in selected['size_band_counts'].items())}.",
+        f"- Qwen3-Omni v6 multiscale export: {next((item['samples'] for item in selected['exports'] if item['key'] == 'qwen_v6_multiscale_export'), 0):,} rows.",
+        f"- Dense multiscale compact export: {next((item['samples'] for item in selected['exports'] if item['key'] == 'dense_multiscale_compact_export'), 0):,} rows.",
+        "",
+        "## Full Gated Dataset Metadata",
+        "",
+        f"- Repo: `{full['repo_id']}` at `{full['repo_sha']}`.",
+        f"- {full['summary'].get('file_count_excluding_gitattributes', 0):,} files excluding `.gitattributes`.",
+        f"- {full['summary'].get('complete_episode_count', 0):,} complete episode folders ({full['summary'].get('complete_episode_pct', 0):.4f}%).",
+        f"- {full['summary'].get('mp4_count', 0):,} MP4 files and {full['summary'].get('annotation_hdf5_count', 0):,} `annotation.hdf5` files.",
+        "",
+        "## Generated Charts",
+        "",
+    ]
+    for chart in payload["charts"]:
+        lines.append(f"- {chart['title']}: `{chart['path']}`")
+    return "\n".join(lines) + "\n"
+
+
+def build_payload(root: Path) -> dict[str, Any]:
+    payload = {
+        "status": "pass",
+        "generated_at_utc": datetime.now(timezone.utc).replace(microsecond=0).isoformat().replace("+00:00", "Z"),
+        "sources": [
+            {"scope": "public_sample", "path": "docs/data/raw_sample_files.json"},
+            {"scope": "public_sample", "path": "docs/data/single_episode_explorer.json"},
+            {"scope": "public_sample", "path": "results/episode_task_suite/windows.csv"},
+            {"scope": "selected_128", "path": "docs/data/xperience10m_128_episode_feature_index.json"},
+            {"scope": "selected_128", "path": "results/omni_finetune/xperience10m_128_episode_selection.csv"},
+            {"scope": "selected_128", "path": "results/omni_finetune/multi_episode_128_task_baselines/windows.csv"},
+            {"scope": "full_hf_dataset", "path": "results/omni_finetune/full_dataset_metadata_audit.json"},
+        ],
+        "public_sample": build_public_sample(root),
+        "selected_128": build_selected_128(root),
+        "full_hf_dataset": build_full_hf_dataset(root),
+    }
+    payload["charts"] = [
+        {
+            "title": "Scope ladder",
+            "path": "assets/charts/data_explorer_scope_ladder.svg",
+            "question": "How do the public sample, selected 128 episodes, and full gated dataset differ in scale?",
+        },
+        {
+            "title": "Public sample feature dimensions",
+            "path": "assets/charts/data_explorer_sample_feature_modalities.svg",
+            "question": "Which modality groups dominate the one-sample task input?",
+        },
+        {
+            "title": "Public sample action distribution",
+            "path": "assets/charts/data_explorer_sample_action_distribution.svg",
+            "question": "Which action labels occupy the most 20-frame windows in the sample?",
+        },
+        {
+            "title": "Selected-128 split rows",
+            "path": "assets/charts/data_explorer_selected128_split_rows.svg",
+            "question": "How many rows are available per selected-128 export and split?",
+        },
+        {
+            "title": "Full dataset file composition",
+            "path": "assets/charts/data_explorer_full_file_composition.svg",
+            "question": "What file types dominate the gated full-dataset metadata?",
+        },
+    ]
+    return payload
+
+
+def build_charts(root: Path, payload: dict[str, Any]) -> None:
+    chart_dir = root / "docs/assets/charts"
+    plot_scope_ladder(payload, chart_dir / "data_explorer_scope_ladder.svg")
+    plot_feature_breakdown(payload, chart_dir / "data_explorer_sample_feature_modalities.svg")
+    plot_action_distribution(payload, chart_dir / "data_explorer_sample_action_distribution.svg")
+    plot_selected_split_windows(payload, chart_dir / "data_explorer_selected128_split_rows.svg")
+    plot_full_file_composition(payload, chart_dir / "data_explorer_full_file_composition.svg")
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--root", type=Path, default=repo_root(), help="Repository root")
+    parser.add_argument("--json-output", type=Path, default=None)
+    parser.add_argument("--report-output", type=Path, default=None)
+    args = parser.parse_args()
+
+    root = args.root.resolve()
+    payload = build_payload(root)
+    build_charts(root, payload)
+
+    json_output = args.json_output or (root / "docs/data/data_explorer_analysis.json")
+    report_output = args.report_output or (root / "DATA_EXPLORER_ANALYSIS.md")
+    write_json(json_output, payload)
+    report_output.write_text(render_markdown(payload))
+    print(f"PASS: wrote {json_output.relative_to(root)}")
+    print(f"PASS: wrote {report_output.relative_to(root)}")
+    for chart in payload["charts"]:
+        print(f"PASS: wrote {chart['path']}")
+
+
+if __name__ == "__main__":
+    main()

scripts/validate_mirror_parity.py

@@ -45,6 +45,7 @@ DATA_FILES = [
     "audio_ablation_summary.json",
     "artifact_index.json",
     "brand_assets.json",
+    "data_explorer_analysis.json",
     "evidence_contract.json",
     "evaluation_protocol.json",
     "figure_index.json",
@@ -100,6 +101,11 @@ DATA_FILES = [
 
 ASSET_FILES = [
     "charts/audio_ablation_delta.svg",
+    "charts/data_explorer_full_file_composition.svg",
+    "charts/data_explorer_sample_action_distribution.svg",
+    "charts/data_explorer_sample_feature_modalities.svg",
+    "charts/data_explorer_scope_ladder.svg",
+    "charts/data_explorer_selected128_split_rows.svg",
     "charts/two_evidence_line_map.svg",
     "charts/single_episode_task_model_radar.svg",
     "charts/episode128_task_model_radar.svg",
@@ -219,6 +225,7 @@ SCRIPT_FILES = [
     "audio_ablation_and_raw_upgrade.py",
     "build_artifact_index.py",
     "build_brand_assets.py",
+    "build_data_explorer_analysis.py",
     "build_evaluation_protocol.py",
     "build_figure_index.py",
     "build_quality_gates.py",
@@ -359,6 +366,7 @@ DOC_FILES = [
     "README.ko.md",
     "README.pt.md",
     "ARTIFACT_GUIDE.md",
+    "DATA_EXPLORER_ANALYSIS.md",
     "OMNI_MODEL_EXTENSION_CONTRACT.md",
     "QUALITY_GATES.md",
     "EVALUATION_PROTOCOL.md",