catheter_force_depth_eef

This dataset is a TsFile conversion of the Hugging Face dataset giakhuyendihoc/catheter_force_depth_eef. The source dataset was created using LeRobot and is stored in the episode-split LeRobot v2.1 layout used by OpenPI.

Modalities: Time-series. The original dataset also contains robot camera videos; those videos are not included in this converted repository and remain available from the original Hugging Face dataset.

Source Dataset

• Original dataset: giakhuyendihoc/catheter_force_depth_eef
• License: apache-2.0
• Source format: LeRobot v2.1 episode Parquet files plus metadata and videos
• Robot type: ur5
• Source scale: 100 episodes, 45,255 frames, 1 task, 1 data chunk, 600 videos
• Sampling rate: 30 fps
• Split: train (0:100)

Task: Insert the gray cable into the rubber tube and push it down until the cable tip reaches the end.

Converted Files

• TsFile data: data/catheter_force_depth_eef.tsfile
• Mirrored metadata: meta/
• Table name: catheter_force_depth_eef
• Time precision: milliseconds
• Converted row count: 45,255
• Source episode shard count: 100 Parquet files

The original source data path is data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet. The converted TsFile path is data/catheter_force_depth_eef.tsfile.

Schema

Time is synthesized as round(timestamp * 1000) and restarts within each episode, matching the source LeRobot timestamp semantics.

TAG columns:

• episode_index
• task_index

FIELD columns:

• frame_index
• sample_index (renamed from source index)
• observation_state_0 to observation_state_6
• action_0 to action_6
• observation_ee_pose_0 to observation_ee_pose_6
• teleop_target_pose_0 to teleop_target_pose_6
• observation_force_torque_0 to observation_force_torque_5

The source motor names for both observation.state and action are:

• shoulder_pan.pos
• shoulder_lift.pos
• elbow.pos
• wrist_1.pos
• wrist_2.pos
• wrist_3.pos
• gripper.pos

The source pose names for observation.ee_pose and teleop.target_pose are:

• translation_x
• translation_y
• translation_z
• rotation_x
• rotation_y
• rotation_z
• rotation_w

The source force/torque names are:

• fx
• fy
• fz
• tx
• ty
• tz

Conversion Notes

The source vector columns are flattened into scalar FLOAT fields:

• observation.state[7] -> observation_state_0..6
• action[7] -> action_0..6
• observation.ee_pose[7] -> observation_ee_pose_0..6
• teleop.target_pose[7] -> teleop_target_pose_0..6
• observation.force_torque[6] -> observation_force_torque_0..5

Dots in source column names are replaced with underscores before appending the vector element index.

The source timestamp column is not retained as a separate field because it is redundant with Time / 1000 seconds. The source index column is renamed to sample_index.

The source video features are omitted from this repository:

• observation.images.wrist
• observation.images.front
• observation.images.base
• observation.images.wrist_depth
• observation.images.front_depth
• observation.images.base_depth

Videos remain in the original dataset at: https://huggingface.co/datasets/giakhuyendihoc/catheter_force_depth_eef/tree/main/videos

The meta/info.json file is mirrored and updated with a tsfile_conversion object that records the source dataset, converted data path, table name, time mapping, TAG columns, flattened features, omitted video features, and row count.

Minimal Read Example

from tsfile import TsFileReader

path = "data/catheter_force_depth_eef.tsfile"

with TsFileReader(path) as reader:
    schemas = reader.get_all_table_schemas()
    print(schemas.keys())