Bounded-memory serverless distributed N-dimensional array processing
Project Links
Meta
Author: Tom White
Requires Python: >=3.8
Classifiers
Development Status
- 3 - Alpha
License
- OSI Approved :: Apache Software License
Operating System
- OS Independent
Programming Language
- Python
- Python :: 3
- Python :: 3.8
Cubed
Note: this is a proof-of-concept, and many things are incomplete or don't work.
Bounded-memory serverless distributed N-dimensional array processing
Cubed is a distributed N-dimensional array library implemented in Python using bounded-memory serverless processing and Zarr for storage.
- Implements the Python Array API standard (see coverage status)
- Guaranteed maximum memory usage for standard array functions
- Follows Dask Array's chunked array API (
map_blocks,rechunk,apply_gufunc, etc) - Zarr for persistent and intermediate storage
- Multiple serverless runtimes: Python (in-process), Lithops, Modal, Apache Beam
0.21.0
Mar 13, 2025
0.20.0
Jan 15, 2025
0.19.0
Dec 02, 2024
0.18.0
Oct 08, 2024
0.17.0
Sep 13, 2024
0.16.0
Aug 08, 2024
0.15.0
Jun 06, 2024
0.14.3
Apr 29, 2024
0.14.2
Apr 17, 2024
0.14.1
Apr 12, 2024
0.14.0
Mar 21, 2024
0.13.0
Feb 27, 2024
0.12.0
Dec 17, 2023
0.11.0
Sep 29, 2023
0.10.0
Aug 03, 2023
0.9.0
Jul 07, 2023
0.8.0
Jun 16, 2023
0.7.0
May 25, 2023
0.6.3
May 19, 2023
0.6.2
May 12, 2023
0.6.1
Mar 27, 2023
0.6.0
Mar 21, 2023
0.5.0
Feb 21, 2023
0.4.1
Oct 21, 2022
0.4.0
Sep 29, 2022
0.3.0
Sep 02, 2022
0.2.0
Jul 22, 2022
0.1.0
Jul 08, 2022
0.0.1
May 22, 2022
Wheel compatibility matrix
Files in release
Extras:
Dependencies:
aiostream
fsspec
mypy-extensions
networkx
(<2.8.3)
numpy
(>=1.22)
tenacity
toolz
zarr