Disable dataset viewer, add sample data preview section

README.md

@@ -19,6 +19,12 @@ size_categories:
 language:
   - en
 pretty_name: "ADAS-TO: A Large-Scale Multimodal Naturalistic Dataset and Empirical Characterization of Human Takeovers during ADAS Engagement"
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: "dummy_no_autoload"
+viewer: false
 ---
 
 <div align="center">
@@ -71,6 +77,88 @@ ADAS-TO captures the critical moment of control transition — the exact instant
 
 ---
 
+## 🔍 Sample Data Preview
+
+<div align="center">
+
+*Example clip: On-coming Traffic scenario (HONDA CIVIC)*
+
+<img src="https://huggingface.co/datasets/HenryYHW/ADAS-TO/resolve/main/assets/takeover_1.gif" width="480"/>
+
+</div>
+
+<details>
+<summary>📋 <b>meta.json</b> — Clip Metadata</summary>
+
+```json
+{
+  "car_model": "HONDA_CIVIC",
+  "dongle_id": "driver_085",
+  "route_id": "route_001",
+  "log_kind": "qlog",
+  "log_hz": 10,
+  "vid_kind": "qcamera",
+  "camera_fps": 20,
+  "clip_id": 0,
+  "event_mono": 289529060357,
+  "video_time_s": 256.0,
+  "clip_start_s": 246.0,
+  "clip_dur_s": 20.0
+}
+```
+</details>
+
+<details>
+<summary>🚗 <b>carState.csv</b> — Vehicle dynamics & driver inputs (200 rows × 16 cols @ 10 Hz)</summary>
+
+| time_s | vEgo (m/s) | aEgo (m/s²) | steeringAngleDeg | steeringTorque | gasPressed | brakePressed | cruiseState.enabled |
+|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|
+| 242.80 | 13.74 | 0.30 | 0.4 | -132.0 | False | False | **True** |
+| 242.90 | 13.80 | 0.52 | 0.5 | -119.0 | False | False | **True** |
+| 243.00 | 13.82 | 0.32 | 0.7 | -115.0 | False | False | **True** |
+| ... | ... | ... | ... | ... | ... | ... | ... |
+| ~256.0 | — | — | — | — | — | — | **True → False** |
+| ... | ... | ... | ... | ... | ... | ... | ... |
+| 261.80 | — | — | — | — | — | — | **False** |
+
+*The `cruiseState.enabled` column transitions from True to False at the takeover moment (~10s into the clip).*
+
+</details>
+
+<details>
+<summary>🤖 <b>controlsState.csv</b> — ADAS controller state (200 rows × 12 cols @ 10 Hz)</summary>
+
+Columns: `logMonoTime`, `time_s`, `enabled`, `active`, `curvature`, `desiredCurvature`, `vCruise`, `vCruiseCluster`, `forceDecel`, `longControlState`, `alertText1`, `alertText2`
+
+</details>
+
+<details>
+<summary>📡 <b>radarState.csv</b> — Lead vehicle detection (80 rows × 19 cols @ 4 Hz)</summary>
+
+Columns: `logMonoTime`, `time_s`, `leadOne.status`, `leadOne.dRel`, `leadOne.vRel`, `leadOne.vLead`, `leadOne.aLeadK`, `leadTwo.dRel`, `leadTwo.vRel`, `leadTwo.vLead`, `leadTwo.aLeadK`, ...
+
+</details>
+
+<details>
+<summary>📂 <b>All 10 files in each clip</b></summary>
+
+| File | Rows × Cols | Freq | Description |
+|:---|:---|:---|:---|
+| `takeover.mp4` | 400 frames | 20 fps | Front-camera video |
+| `meta.json` | — | — | Clip metadata |
+| `carState.csv` | 200 × 16 | 10 Hz | Vehicle dynamics & driver inputs |
+| `controlsState.csv` | 200 × 12 | 10 Hz | ADAS controller state |
+| `carControl.csv` | 200 × 8 | 10 Hz | Control commands |
+| `carOutput.csv` | 200 × 9 | 10 Hz | Actuator outputs |
+| `drivingModelData.csv` | 200 × 7 | 10 Hz | Model predictions |
+| `radarState.csv` | 80 × 19 | 4 Hz | Radar / lead vehicle |
+| `accelerometer.csv` | ~200 × 4 | ~10 Hz | IMU data |
+| `longitudinalPlan.csv` | 200 × 8 | 10 Hz | Planner outputs |
+
+</details>
+
+---
+
 ## 📊 Dataset at a Glance
 
 <div align="center">