bucket-dir 1.12.0

Generate directory listings for S3 statically hosted content.

bucket-dir

bucket-dir is a utility for generating a browsable directory tree for an AWS S3 bucket.

It was built in order to host Maven and Ivy repositories in S3 and serve them via CloudFront, but it could meet other needs too.

Installation

pip install bucket-dir

Usage

Run bucket-dir with the name of the bucket you wish to index as a parameter:

bucket-dir foo-bucket

If you only want to upload indexes for a particular part of the bucket, use --target-path. This will generate indexes for folders that lead to the path, and everything under the path:

# These all update the root index, foo-folder's index, and everything underneath foo-folder
bucket-dir foo-bucket --target-path '/foo-folder/foo-object'
bucket-dir foo-bucket --target-path '/foo-folder/'
bucket-dir foo-bucket --target-path 'foo-folder/foo-object'
bucket-dir foo-bucket --target-path 'foo-folder/'

If you need to exclude objects with certain names from the index use --exclude-object. This will hide any objects that match this name. index.html objects are ignored for free:

bucket-dir foo-bucket --exclude-object 'error.html' --exclude-object 'foo-object'

Use bucket-dir --help for all arguments.

Be sure to provide the command with credentials that allow it to perform ListBucket and PutObject calls against the bucket. E.g. with aws-vault:

aws-vault exec foo-profile -- bucket-dir foo-bucket

IAM requirements

This example demonstrates the most restrictive policy you can apply to the principal (e.g. an IAM user or role) that is going to run bucket-dir. Substitute foo-bucket for the name of your bucket:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": "s3:ListBucket",
            "Resource": "arn:aws:s3:::foo-bucket"
        },
        {
            "Effect": "Allow",
            "Action": "s3:PutObject",
            "Resource": "arn:aws:s3:::foo-bucket/*"
        }
    ]
}

• s3:ListBucket is required for bucket-dir to be able to map out the folders and objects that the bucket contains.
• s3:PutObject is required for bucket-dir to be able to upload the generated index.html documents.

Example AWS configuration

For examples on how you can configure an S3 bucket to serve static site content indexed by bucket-dir, see:

• Configuring a public S3 Bucket for use with bucket-dir.

Examples of how you can front public and private buckets with CloudFront, and how bucket-dir can be run in a lambda, will be added in due course.

Using bucket-dir as a library

bucket-dir can also be used as a dependency of your own python applications.

from bucket_dir import BucketDirGenerator

bucket_dir_generator = BucketDirGenerator()
bucket_dir_generator.generate(bucket="foo-bucket", site_name="my static site")

Character support

bucket-dir supports objects using any of the Safe characters listed in the S3 object key naming guidelines.

The exception to the above rule is using forward slashes consecutively (e.g. my-folder//my-object). This results in a folder called /, which breaks hyperlinks.

Use of characters in the Characters that might require special handling list is currently unsupported but is theoretically ok.

Some characters in Characters to avoid may also work, but you're on your own.

Development

Start with make init. This will install prerequisties and set up a poetry managed virtual environment containing all the required runtime and development dependencies.

Unit testing can be performed with make test. If you want to run pytest with other options, use poetry run pytest ....

You can execute the source code directly with poetry run bucket-dir.

Finally, you can build with make build. This will update dependencies, run security checks and analysis and finally package the code into a wheel and archive.

Publishing can be performed with make publish, but this is only intended to run in CI on commit to the main branch. If running locally, you need to have PyPI credentials set as env vars.

For other rules, see the Makefile.

If you are a collaborator, feel free to make changes directly to the main branch. Otherwise, please raise a PR. Don't forget to bump the version in pyproject.toml.

Profiling

To get a performance profile, use:

make profile

You must have the graphviz library installed.

A combined.svg image will be generated in the prof directory which you can use to find bottlenecks and potential enhancements.

License

This code is open source software licensed under the Apache 2.0 License.