s3workers

Helper to simplify concurrent access to object scanning in AWS S3 buckets.

Free software: MIT license
Documentation: https://s3workers.readthedocs.io.

Simplest install method is via pip install s3workers (see installation for other methods).

Features

S3workers provides faster list and delete operations on S3 buckets by opening up simultaneous connections to issue distinct sets of shared prefix queries. Effectively, this splits up the query space into 36 independent queries (26 alpha and 10 numeric prefixes). For example, a request to list all objects in the myfancy/ bucket would result in concurrent list queries to S3 for everything from myfancy/a... through myfancy/b... and everything from myfancy/0... through myfancy/9..., all at the same time, reporting and collating the results locally.

Selection

The default output of s3workers is to simply list (or delete) all objects found at the prefix requested. However, often it is advantageous to restrict the output to only those matching certain criteria. The --select option provides the ability for evaluating matches using any normal Python operators or builtins against one or more of the following variables provided to the selector for each object found:

name: The full S3 key name, everything except the bucket name (string)
size: The number of bytes as used by the S3 object (integer).
md5: The MD5 hash of the S3 object (string).
last_modified: The timestamp indicating the last time the S3 object was changed (string).

Reduction

In cases where aggregation of some kind is desired, s3workers provides the ability to execute reduction logic against an accumulator value. For example, to produce a sum of the size of all selected S3 objects or to even group the size according to MD5 values. See the usage output for examples. In all cases, the same variables provided by selection are also provided when reducing.

Usage

For more help and examples of selection and reduction abilities, check out the usage.

Credits