pragnesh/ec2selector

Interactive prompt to select an AMI (Amazon Machine Image) from Amazon EC2

Abandoned

ec2selector is a tool to select an EC2 AMI (Amazon Machine Image) from those available on Amazon. There are three basic ways in which you can use this tool.

Selecting an image knowing its ID

If you already know the image ID, you can use ec2selector to make sure that such an instance is still available on Amazon for a given region, e.g.:

from ec2selector import EC2Selector
EC2Selector().select(region_id='us-east-1', image_id='ami-ed16f984')

will return Image:ami-ed16f984 if such an image exists on US East 1.

Selecting an image specifying a set of properties

If you don't know the image ID, you can specify a set of properties to match one of the images available on Amazon. The properties you can filter on are:

• name
• owner_id
• virtualization_type
• image_type
• architecture

You can also use the properties followed by '_contains' to specify a partial match. For instance, the following commands will return an AMI in US East, owned by the user 'alestic' with a 'machine' image at 64bit, and whose name includes the string 'ubuntu-10':

from ec2selector import EC2Selector
EC2Selector().select(region_id='us-east-1', filters={
    'name_contains': 'ubuntu-10', 
    'owner_id':'063491364108',     # corresponds to 'alestic'
    'image_type': 'machine', 
    'virtualization_type':'paravirtual',
    'architecture': 'x86_64', 
})

Selecting an image with an interactive prompt

Finally, if you do not specify enough filters to get exactly one image, you will get an interactive shell to add more filters or pick one of the available images. For instance, this is a prompt session:

>>> from ec2selector import EC2Selector
>>> EC2Selector().select()

5 regions available
Please pick one of the available options:
 [0] eu-west-1
 [1] us-east-1
 [2] ap-northeast-1
 [3] us-west-1
 [4] ap-southeast-1

> 0
Loading list of available EC2 images for eu-west-1 with filters: None, might take a while...
1717 images available.
Please pick one of the available options:
 [0] name
 [1] owner_id
 [2] architecture
 [3] type
 [4] virtualization_type
 
 > 2
 Please pick one of the available options:
  [0] x86_64
  [1] i386

 > 0
 577 images available.
 Please pick one of the available options:
  [0] name
  [1] owner_id
  [2] architecture
  [3] type
  [4] virtualization_type

 > 3
 Please pick one of the available options:
  [0] machine
  [1] kernel
  [2] ramdisk

 > 0
 519 images available.
 Please pick one of the available options:
  [0] name
  [1] owner_id
  [2] architecture
  [3] type
  [4] virtualization_type

 > 4
 Please pick one of the available options:
  [0] hvm
  [1] paravirtual

 > 1
 386 images available.
 Please pick one of the available options:
  [0] name
  [1] owner_id
  [2] architecture
  [3] type
  [4] virtualization_type

 > 0
 Please pick one of the available options:
  [0] Ubuntu 10.10 64-bit Server Python Erlang Java
  [1] RightImage_Ubuntu_8.04_x64_v5.5.9.1_EBS
  ...
  [383] ubuntu-9.10-karmic-server-amd64-20100121
  [384] xarch-core-image
  [385] radiant-0.9.1_64_0.2_ami-75d4e101

 > 385
 Image selected: radiant-0.9.1_64_0.2_ami-75d4e101
 Image:ami-033d0977

Requirements

In order to use ec2selector you need login credentials to Amazon EC2. You can either pass these credentials to EC2Selector, e.g.:

from ec2selector import EC2Selector
EC2Selector().select(access_key='AKIA...', secret_key='xWD3...')

or you can set them in environment variables, e.g.:

import os
os.environ['AWS_ACCESS_KEY_ID'] = 'AKIA...'
os.environ['AWS_SECRET_ACCESS_KEY'] = 'xWD3...'

If you don't specify them, the interactive prompt will ask for them.

Motivation

ec2selector is based on boto, a great Python library to interact with Amazon EC2 instances. The only way to select an AMI with boto is by specifying its image ID. ec2selector expands this behavior by letting you pick an image based on its attributes, not on its ID.

Installing

pip install ec2selector

Contributing

1. fork and clone the project
2. install the dependencies above
3. run the tests with python tests.py
4. hack at will
5. commit and push
6. send a pull request

License

See LICENSE.txt