Hi everyone,
I am using label studio (version 1.21.0) and I want to import labels for time series and videos from a csv file containing time stamp and labels. Is there any practical way to do that? I tried with importing that as JSON format like this:
{
“id”: 20981,
“data”: {
“video”: “/data/local-files/?d=21112025/video_2.mp4”,
“acc_IMU”: “/data/local-files/?d=21112025/2025-11-21 12_50_24Z_253830002029_accele.csv”,
“gyr_IMU”: “/data/local-files/?d=21112025/2025-11-21 12_50_24Z_253830002029_gyrosc.csv”,
“mag_IMU”: “/data/local-files/?d=21112025/2025-11-21 12_50_24Z_253830002029_magnet.csv”,
“ECG”: “/data/local-files/?d=21112025/2025-11-21 12_50_24Z_253830002019_ecg_uu.csv”
},
“annotations”: [
{
“id”: 20,
“result”: [
{
“value”: {
“start”: “2025-11-21 12:50:36.315000”,
“end”: “2025-11-21 12:53:37.691000”,
“instant”: false,
“timeserieslabels”: [
“baseline”
]
},
“id”: “rT-AV1muuP”,
“from_name”: “bl_labels”,
“to_name”: “ecg_ts”,
“type”: “timeserieslabels”,
“origin”: “manual”
},
{
“value”: {
“start”: “2025-11-21 12:53:45.143000”,
“end”: “2025-11-21 12:54:38.655000”,
“instant”: false,
“timeserieslabels”: [
“ordering”
]
},
“id”: “Nd2T7v_MrT”,
“from_name”: “bl_labels”,
“to_name”: “ecg_ts”,
“type”: “timeserieslabels”,
“origin”: “manual”
}
],
“created_username”: " xxx, 1",
“created_ago”: “5 days, 20 hours”,
“completed_by”: {
“id”: 1,
“first_name”: “”,
“last_name”: “”,
“avatar”: null,
“email”: “xxx”,
“initials”: “xx”
},
“was_cancelled”: false,
“ground_truth”: false,
“created_at”: “2026-02-03T13:03:58.015452Z”,
“updated_at”: “2026-02-03T13:03:58.015477Z”,
“draft_created_at”: null,
“lead_time”: 67469.847,
“import_id”: 19,
“last_action”: null,
“bulk_created”: false,
“task”: 20981,
“project”: 29,
“updated_by”: null,
“parent_prediction”: null,
“parent_annotation”: null,
“last_created_by”: null
}
],
“predictions”: []
}
However I was wondering if there is any practical ways of importing labels directly from the csv file