Private Artifact Manager using Amazon S3

• Easy & lightweight artifact manager for Java/Scala
  • Especially optimized for all-in-one package (e.g. the result of sbt dist or sbt assembly)
• No database or extra servers needed except Amazon S3
• Quite simple command line interface (CLI)
• Integrate your daily build & deploy tasks

• Python 2.6 / 2.7
• pytz
• python-dateutil
• GitPython >= 0.3.5
• boto >= 2.38.0
• moto (for testing)

• pip command may need sudo

The responsibility of the AWS's charge is your own.

Manage your own access key ID and secret access key to call the APIs.

Set permissions to the IAM user.

Create ~/.artifact-cli file and write credentials for AWS like this:

[default]
aws_access_key_id = XXXXXXXXXXXXXXXXXXXX
aws_secret_access_key = XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
bucket = your-bucket-name
region = your-region (e.g. ap-northeast-1, us-east-1)

You can override these settings by specifying command line options or environment variables.

• Command Line Options (will override the settings from environment variables)

• Environment Variables

Now, you are ready for using art command in the shell. Just list your artifacts:

$ art list GROUP_ID
[INFO] No artifacts.

Of course, there are no artifacts!

Building is outside the reach of this tool. In other words, you can build as you like.

In the builder's environment, you can upload the artifact to Amazon S3:

$ art upload GROUP_ID /path/to/your-artifact-0.0.1.jar

Specify group id and your local file path.

To view the index data, run art list or art info.

Login to the deployer's environment, then download the artifact from Amazon S3:

$ art download GROUP_ID /path/to/deployers/your-artifact-0.0.1.jar 1

To download the latest revision, use latest keyword. (case-sensitive):

$ art download GROUP_ID /path/to/deployers/your-artifact-0.0.1.jar latest

Deploy the artifact any way you like!

For further information, type art -h.

The structure of the paths is the following:

your-bucket-name
├── group.id.1                          // group ID
│   ├── .meta                           // meta data directory for each group
│   │   ├── index-awesome-project.json  // index data is written as JSON for each artifact ID
│   │   └── index-play-project.json
│   ├── awesome-project                 // artifact ID
│   │   ├── 0.0.1                       // version
│   │   │   ├── 1                       // revision (auto assigned, starting from 1)
|   │   │   │   └── awesome-project-0.0.1.jar
|   │   │   ├── 2
|   │   │   │   └── awesome-project-0.0.1.jar
|   │   │   ├── 3
|   │   │   │   └── awesome-project-0.0.1.jar
|   │   │   └── 4
|   │   │       └── awesome-project-0.0.1.jar
│   │   └── 0.0.2-SNAPSHOT
│   │       ├── 1
|   │       │   └── awesome-project-0.0.2-SNAPSHOT.jar
|   │       └── 2
|   │           └── awesome-project-0.0.2-SNAPSHOT.jar
│   └── play-project
│       └── 0.0.1
│           └── 1
|               └── play-project-0.0.1.zip
└── group.id.2
    ├── .meta
    │   └── index-awesome-project.json 
    └── awesome-project                 // completely separated to the group.id.1's artifact
        └── 0.0.1
            └── 1
                └── awesome-project-0.0.1.zip

• This tool supports only artifact-id-level concurrency.
  • Simultaneous uploading of the artifacts with same artifact id could let repository broken.