rkale 1.0.3

Rclone wrapper to manage multiple datasets in a project

rkale

Install

Install rkale in your project using poetry:

poetry add rkale

Use pip if you want a global installation:

pip install rkale

Configuration

Global

~/.config/rkale/rkale.conf:

[data]
root = "path to data folder where datasets are stored"

[aliases]
wasabi = "optional alias for remote in rclone.conf"

[rclone] # global flags for rclone
flags = ["--transfers 32", "--checkers 32"]

If aliases are empty the remote name from the project config is used in the rclone lookup.

Project

Configure project datasets in the pyproject.toml file:

<project path>/pyproject.toml:

[[tool.rkale.dataset]]
name = "dataset_1"
remote = "remote_1"

[[tool.rkale.dataset]]
name = "dataset_2"
remote = "remote_2"

The remote specified for the dataset must match a remote in the rclone.conf or an alias in the global rkale configuration.

Usage

Python interface

from rkale.config import dataset_paths


def dataset_path():
    return dataset_paths()["dataset_1"]

Syncing datasets

Syncs the local datasets to be identical to the remote

rkale psync

Syncs the remote datasets to be identical to the local

rkale psync --upstream

Same as rclone sync but checks differences first and asks for confirmation

rkale sync <source> <destination>