---
license: apache-2.0
authors:
- Brett Peters
task_categories:
- robotics
tags:
- tsfile
- timeseries
- tabular
modality:
- timeseries
- tabular
pretty_name: BAP13 SO101 Test 3 TsFile
configs:
- config_name: default
  data_files:
  - split: train
    path: data/bap13_so101_test_3_train.tsfile
size_categories:
- 10K<n<100K
---

# BAP13 SO101 Test 3 TsFile

This is an Apache TsFile conversion of
[bap13/so101_test_3](https://huggingface.co/datasets/bap13/so101_test_3), a LeRobot v2.1 SO101
robot-manipulation dataset. The task is: Grasp a blue block and put it in
the bowl.

## Source Dataset and Provenance

- Original dataset: [bap13/so101_test_3](https://huggingface.co/datasets/bap13/so101_test_3)
- Original repository owner and uploader: [Brett Peters (bap13)](https://huggingface.co/bap13)
- License: Apache-2.0
- Robot type: so101
- LeRobot codebase version: v2.1
- Split: train
- Sampling rate: 30 fps
- Scale: 53 episodes, 30,617 frame rows, 1 task, 53 source Parquet files, and 106 source videos
- Source frame layout: data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet
- Source video layout: videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4
- Paper/citation: not supplied by the source dataset card

The source README contains an obsolete 3-episode example. The pinned repository
files, meta/info.json, meta/episodes.jsonl, and all 53 Parquet shards agree on
the current 53-episode, 30,617-row dataset.

## Schema

| Column | TsFile type | Role | Encoding | Compression |
|---|---|---|---|---|
| Time | TIMESTAMP | TIME | TS_2DIFF | LZ4 |
| episode_index | STRING | TAG | Device/tag storage | Device/tag storage |
| task_index | STRING | TAG | Device/tag storage | Device/tag storage |
| frame_index | INT64 | FIELD | TS_2DIFF | LZ4 |
| sample_index | INT64 | FIELD | TS_2DIFF | LZ4 |
| action_0 through action_5 | FLOAT | FIELD | GORILLA | LZ4 |
| observation_state_0 through observation_state_5 | FLOAT | FIELD | GORILLA | LZ4 |

The source index column is renamed to sample_index. frame_index is retained.
episode_index and task_index are stored as TsFile table-model TAG/device columns. 

## Videos

Videos are not included in this TsFile dataset. They remain at the pinned
source revision in two frame-aligned streams:

- [observation.images.context](https://huggingface.co/datasets/bap13/so101_test_3/tree/a904f6e6558b910eee1cac6b75e6c0bf0bb7d0f1/videos/chunk-000/observation.images.context): 53 per-episode MP4 files, 202,287,802 bytes
- [observation.images.wrist](https://huggingface.co/datasets/bap13/so101_test_3/tree/a904f6e6558b910eee1cac6b75e6c0bf0bb7d0f1/videos/chunk-000/observation.images.wrist): 53 per-episode MP4 files, 312,978,469 bytes

Together the 106 AV1 videos occupy 515,266,271 bytes (about 491.40
MiB). Both streams are 640 by 480 at 30 fps without audio. Numeric rows align
with the source videos through episode_index and frame_index.

## Minimal Read Example

    from tsfile import TsFileReader

    reader = TsFileReader("data/bap13_so101_test_3_train.tsfile")
    table_name = "bap13_so101_test_3_train"
    columns = [
        "episode_index",
        "task_index",
        "frame_index",
        "sample_index",
        "action_0",
        "observation_state_0",
    ]

    with reader.query_table(table_name, columns, batch_size=65536) as result:
        batch = result.read_arrow_batch()
        print(batch.to_pandas().head())
    reader.close()

## Citation

The source card provides no paper or completed citation. Cite the original
Hugging Face dataset and Brett Peters (bap13) when using this converted
artifact.