Utils API

このページでは、Robopyのユーティリティ機能に関連するAPIを説明します。

実験ハンドラー

RakudaExpHandler

RakudaExpHandler

RakudaExpHandler(rakuda_config: RakudaConfig, metadata_config: MetaDataConfig, fps: int = 10)

Bases: ExpHandler

This class handles the experimental interface for the Rakuda robot.

Sensors: 1x Realsense, 2x Digit,

Example:

from robopy.utils.exp_interface import RakudaExpHandler
handler = RakudaExpHandler(
    leader_port="/dev/ttyUSB0",
    follower_port="/dev/ttyUSB1",
    left_digit_serial="D20542",
    right_digit_serial="D20537",
    fps=20,
)
handler.record_save(max_frames=150, save_path="test_01", if_async=True)

init initialize Rakuda experimental handler

PARAMETER	DESCRIPTION
leader_port	leader serial port TYPE: str
follower_port	follower serial port TYPE: str
left_digit_serial	left digit serial number TYPE: str
right_digit_serial	right digit serial number TYPE: str
fps	The frequency to capture obs. Defaults to 10 TYPE: int DEFAULT: 10

RAISES	DESCRIPTION
ValueError	fps must be between 1 and 20
RuntimeError	failed to connect to Rakuda robot

METHOD	DESCRIPTION
record_save	record and save data from Rakuda robot

record_save

record_save(max_frames: int, save_path: str, if_async: bool = True, save_gif: bool = True, warmup_time: int = 5) -> None

record and save data from Rakuda robot

PARAMETER	DESCRIPTION
max_frames	maximum number of frames to record TYPE: int
save_path	path to save the recorded data: TYPE: str
if_async	If use parallel. Defaults to True. TYPE: bool DEFAULT: True
save_gif	if save gif. Defaults to True. TYPE: bool DEFAULT: True
warmup_time	warm up time before recording. Defaults to 5. TYPE: int DEFAULT: 5

RAISES	DESCRIPTION
RuntimeError	failed to record from Rakuda robot
RuntimeError	failed to save data

データ処理

H5Handler

H5Handler

Handler for saving and loading data using HDF5 format with h5py.

METHOD	DESCRIPTION
get_info	Get information about HDF5 file structure.
load_hierarchical	Load hierarchical data structure from HDF5 file.

get_info staticmethod

get_info(file_path: str) -> Dict[str, Any]

Get information about HDF5 file structure.

PARAMETER	DESCRIPTION
file_path	Path to the HDF5 file. TYPE: str

RETURNS	DESCRIPTION
Dict[str, Any]	Dict[str, Any]: Information about file structure and sizes.

load_hierarchical staticmethod

load_hierarchical(file_path: str) -> Dict[str, Any]

Load hierarchical data structure from HDF5 file.

PARAMETER	DESCRIPTION
file_path	Path to the HDF5 file. TYPE: str

RETURNS	DESCRIPTION
Dict[str, Any]	Dict[str, Any]: Hierarchical dictionary containing loaded data.

Example

data = H5Handler.load_hierarchical('output.h5') camera_data = data['camera']['main'] # Access nested data

BlocsHandler

BLOSCHandler

BLOSCHandler()

METHOD	DESCRIPTION
load	Load data dictionary from a Blosc2 file.
save	Save data dictionary to a Blosc2 file.

load staticmethod

load(path: str) -> NDArray

Load data dictionary from a Blosc2 file.

save staticmethod

save(data: NDArray, path: str) -> None

Save data dictionary to a Blosc2 file.

ユーティリティ関数

find_usb_port

find_usb_port

FUNCTION	DESCRIPTION
find_available_ports	Find available USB ports on the system.

find_available_ports

find_available_ports() -> list[str]

Find available USB ports on the system.

RETURNS	DESCRIPTION
list[str]	list[str]: A list of available USB port names.

使用例

実験データの記録と保存

from robopy.config import RakudaConfig, RakudaSensorParams, TactileParams
from robopy.utils.exp_interface import RakudaExpHandler

# 設定
config = RakudaConfig(
    leader_port="/dev/ttyUSB0",
    follower_port="/dev/ttyUSB1",
    sensors=RakudaSensorParams(
        tactile=[
            TactileParams(serial_num="D20542", name="left"),
            TactileParams(serial_num="D20537", name="right"),
        ],
    ),
)

# ハンドラー作成
handler = RakudaExpHandler(
    rakuda_config=config,
    fps=20
)

# データ記録と保存
handler.record_save(
    max_frames=1000,
    save_path="experiment_001",
    if_async=True
)

H5ファイルの読み込みと保存

from robopy.utils import H5Handler

# H5ファイルの情報取得
file_info = H5Handler.get_info("path/to/data.h5")

# 階層的データの読み込み
hierarchical_data = H5Handler.load_hierarchical("path/to/data.h5")

# データの保存
import h5py
with h5py.File("output.h5", "w") as f:
    f.create_dataset("array_data", data=your_data)