---
task_categories:
- robotics
tags:
- LeRobot
- robotics
- so100
- manipulation
- tsfile
- modality:timeseries
pretty_name: SO-100 Grab Pen (record-test-11, TsFile)
size_categories:
- 10K<n<100K
configs:
- config_name: default
  data_files:
  - split: train
    path: data/alexisbo_record_test_11.tsfile
modality:
- timeseries
license: apache-2.0
---

# SO-100 Grab Pen — record-test-11 (TsFile)

Apache TsFile version of [`Alexisbo/record-test-11`](https://huggingface.co/datasets/Alexisbo/record-test-11).

## Overview

Teleoperated demonstrations on an **SO-100 follower arm** (`robot_type =
so100_follower`, 6 DoF) of a single task: **"Grab pen and put in orange box"**.
Each episode records the arm's joint positions (state) and commanded joint
targets (action) over time.

- **Episodes:** 46 demonstrations of one task.
- **Frames:** 33,461 timesteps total, recorded at **30 fps**.
- **Robot:** SO-100 follower arm (6 joints incl. gripper).
- **Source format:** LeRobot v2.1.

> **Episode count note:** the source `data/` directory ships **46** episode files
> (indices 0–45). The source `meta/` declares only **42** episodes
> (`info.json` `total_episodes=42` / `total_frames=28964`; `episodes.jsonl` has
> 42 lines). Episodes 42–45 (4,497 frames) are present and valid on disk but
> missing from `meta`. This TsFile includes **all 46** episodes; the mirrored
> `meta/` is left as-is (still describing 42), so prefer the TsFile contents as
> the source of truth for episode/frame counts.

## Schema (TsFile structure)

- **Time** (INT64, milliseconds) — per-frame timestamp, restarts at 0 for each
  episode (`round(timestamp * 1000)`; ~33 ms apart at 30 fps).
- **episode_index** (TAG) — demonstration id (0–45). Query a single episode
  with `WHERE episode_index=0`.
- **task_index** (TAG) — task id (always `0`; the dataset has one task).
- **frame_index** (FIELD, INT64) — frame number within the episode.
- **sample_index** (FIELD, INT64) — the source global `index` column.
- **observation_state_0 .. observation_state_5** (FIELD, FLOAT) — measured joint
  positions.
- **action_0 .. action_5** (FIELD, FLOAT) — commanded joint targets.

The 6 channels of each vector are the SO-100 joints, in order:
`shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper`
(positions, in degrees).

The source `timestamp` column is dropped because it equals `Time / 1000`.
Camera videos are not included in this repository (see below).

## Usage

Read the `.tsfile` file with the Apache TsFile Java or Python SDK.

```python
# pip install tsfile
from tsfile import TsFileReader
reader = TsFileReader("data/alexisbo_record_test_11.tsfile")
# query one episode: WHERE episode_index=0
```

## Videos

The original dataset records two camera streams per episode
(`observation.images.front`, `observation.images.side`). These MP4 videos are
**not** included here — they are not time-series data. Retrieve them from the
original dataset under
[`Alexisbo/record-test-11/videos`](https://huggingface.co/datasets/Alexisbo/record-test-11/tree/main/videos).

## Source & license

- Original dataset: https://huggingface.co/datasets/Alexisbo/record-test-11
- Author / publisher: Alexisbo
- License: not declared by the original dataset; please defer to the original.