s5cmdpy 0.1.0

python binding for using s5cmd to download and upload files to s3 efficiently

s5cmd-python

python binding for using s5cmd to download and upload files to s3 efficiently

The S5CmdRunner class provides a Python interface for interacting with s5cmd, a command-line tool designed for efficient data transfer to and from Amazon S3.

For more information about s5cmd, please refer to the original s5cmd repository.

Features

• Check for the presence of s5cmd and download it if necessary.
• Execute s5cmd commands cp, mv, and run.
• Handle file downloads from URLs and S3 URIs.
• Generate command files for batch operations with s5cmd.
• Simplify operations like copying and moving files between local paths and S3 URIs.

Installation

To use S5CmdRunner, ensure that Python 3.10 or higher is installed. The project itself can be installed from pip:

pip install s5cmdpy

or from source:

git clone https://github.com/troph-team/s5cmd-python
cd s5cmd-python
pip install -e .

Usage

Here are some examples of how to use the S5CmdRunner class:

Initialize S5CmdRunner

from s5cmdpy import S5CmdRunner
runner = S5CmdRunner()

Run s5cmd with a Local Command File

# local_txt: `cp s3://dataset-artstation-uw2/artists/__andrey__/1841730##GZGgW.json .`
local_txt_path = "s5cmd_test.txt"
runner.run(local_txt_path)

Run s5cmd with a Command File from S3

# Useful in environments like SageMaker or for reproducibility; 
# Extends `s5cmd run something.txt` to support command files stored in S3
txt_s3_uri = "s3://dataset-artstation-uw2/s5cmd_test.txt"
runner.run(txt_s3_uri)

Download Multiple Files from S3

# Input a series of S3 URIs to create the necessary commands.txt for `s5cmd run`, 
# then execute `s5cmd run <commands.txt>`

s3_uris = [
    's3://dataset-artstation-uw2/artists/__andrey__/1841730##GZGgW.json', 
    's3://dataset-artstation-uw2/artists/__andrey__/2249992##q5Y22.json'
]
destination_dir = '/home/ubuntu/datasets/s5cmd_test'
runner.download_from_s3_list(s3_uris, destination_dir)

License

S5cmd itself is MIT licensed. This project is also MIT licensed.