Skip to main content

Common AWS code for STScI applications

Project description


Utilities for running STScI software on Amazon Web Services.


The ConcurrentS3Client class streams S3 objects into memory (BytesIO) using multiple concurrent HTTP range requests. In tests on an m4.10xlarge with a 10 Gpbs network connection, we have seen ~ 3x speedup compared to boto3's S3 client. Client credentials and region are configured as for boto3 (environment variables, .aws directory, or instance metadata).

AsyncConcurrentS3Client is similar, but with an async interface.

An example of opening a FITS file from S3 using ConcurrentS3Client:

from stsci_aws_utils.s3 import ConcurrentS3Client
from import fits

with ConcurrentS3Client() as client:
    content = client.get_object("some-s3-bucket", "some/key/prefix/some-file.fits")
    hdul =

Note that when making successive requests, it is more efficient to reuse an open client than to create a new one every time. It is possible to create the client without context management, but you'll need to remember to call the close() method yourself:

client = ConcurrentS3Client()

# ...


Otherwise, you'll see a cryptic complaint written to stderr by aiohttp.

If you find yourself in a situation where it's inconvenient to close the client (maybe you're implementing a module method), atexit is at your service:

import atexit

CLIENT = ConcurrentS3Client()

It ensures that close is called before normal program exit.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

stsci_aws_utils-0.1.2.tar.gz (13.7 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page