aws_s3_uri.rb

require "uri"

# Parses an S3 URI for region, bucket and key
#
# Based on the java sdk:
#
#   https://github.com/aws/aws-sdk-java/blob/master/aws-java-sdk-s3/src/main/java/com/amazonaws/services/s3/AmazonS3URI.java
#
# Reference:
#
#   https://docs.aws.amazon.com/general/latest/gr/s3.html
#   https://docs.aws.amazon.com/AmazonS3/latest/userguide/VirtualHosting.html
#
class AwsS3URI
  class InvalidURI < ArgumentError; end

  ENDPOINT_PATTERN = /\A(?:(?<bucket>.+)\.)?s3(?:-[a-z0-9-]+)?\.(?:dualstack\.)?(?:(?<region>[a-z0-9-]+?)\.)?amazonaws\.com\z/i

  attr_reader :uri, :region, :bucket, :key

  def initialize(url)
    @uri = URI.parse(url)

    if @uri.scheme == "s3"
      unless @uri.host
        raise InvalidURI.new("Missing bucket")
      end

      @region = nil
      @bucket = @uri.host

      unless @uri.path.empty?
        @key = @uri.path.delete_prefix("/")
      end
    elsif @uri.scheme == "http" || @uri.scheme == "https"
      if @uri.host.nil? || @uri.host.empty?
        raise InvalidURI.new("Missing hostname")
      end

      if match = ENDPOINT_PATTERN.match(@uri.host)
        @region = match[:region] || "us-east-1"

        if match[:bucket]
          # Virtual host bucket name
          @bucket = match[:bucket]

          unless @uri.path.empty?
            @key = @uri.path.delete_prefix("/")
          end
        else
          # Path-style bucket name
          prefix, rest = @uri.path.delete_prefix("/").split("/", 2)

          if prefix.nil? || prefix.empty?
            raise InvalidURI.new("Missing bucket")
          end

          @bucket = prefix

          unless rest.nil? || rest.empty?
            @key = rest
          end
        end
      else
        raise InvalidURI.new("Not a recognized S3 endpoint")
      end
    else
      raise InvalidURI.new("Unknown scheme")
    end
  end
end

aws_s3_uri_spec.rb

require "spec_helper"

require "aws_s3_uri"

RSpec.describe AwsS3URI do
  it "parses s3://BUCKET" do
    expect(AwsS3URI.new("s3://foo")).to have_attributes(region: nil, bucket: "foo", key: nil)
  end

  it "parses s3://BUCKET/" do
    expect(AwsS3URI.new("s3://foo")).to have_attributes(region: nil, bucket: "foo", key: nil)
  end

  it "parses s3://BUCKET/KEY" do
    expect(AwsS3URI.new("s3://foo/bar/baz")).to have_attributes(region: nil, bucket: "foo", key: "bar/baz")
  end

  it "parses http://BUCKET.s3.amazonaws.com" do
    expect(AwsS3URI.new("http://foo.s3.amazonaws.com")).to have_attributes(region: "us-east-1", bucket: "foo", key: nil)
  end

  it "parses http://BUCKET.s3.amazonaws.com/KEY" do
    expect(AwsS3URI.new("http://foo.s3.amazonaws.com/bar/baz")).to have_attributes(region: "us-east-1", bucket: "foo", key: "bar/baz")
  end

  it "parses http://BUCKET.s3-fips.amazonaws.com" do
    expect(AwsS3URI.new("http://foo.s3-fips.amazonaws.com")).to have_attributes(region: "us-east-1", bucket: "foo", key: nil)
  end

  it "parses http://BUCKET.s3-fips.amazonaws.com/KEY" do
    expect(AwsS3URI.new("http://foo.s3-fips.amazonaws.com/bar/baz")).to have_attributes(region: "us-east-1", bucket: "foo", key: "bar/baz")
  end

  it "parses http://BUCKET.s3.dualstack.amazonaws.com" do
    expect(AwsS3URI.new("http://foo.s3.dualstack.amazonaws.com")).to have_attributes(region: "us-east-1", bucket: "foo", key: nil)
  end

  it "parses http://BUCKET.s3.dualstack.amazonaws.com/KEY" do
    expect(AwsS3URI.new("http://foo.s3.dualstack.amazonaws.com/bar/baz")).to have_attributes(region: "us-east-1", bucket: "foo", key: "bar/baz")
  end

  it "parses http://BUCKET.s3.REGION.amazonaws.com" do
    expect(AwsS3URI.new("http://foo.s3.us-west-1.amazonaws.com")).to have_attributes(region: "us-west-1", bucket: "foo", key: nil)
  end

  it "parses http://BUCKET.s3.REGION.amazonaws.com/KEY" do
    expect(AwsS3URI.new("http://foo.s3.us-west-1.amazonaws.com/bar/baz")).to have_attributes(region: "us-west-1", bucket: "foo", key: "bar/baz")
  end

  it "parses http://BUCKET.s3-fips.REGION.amazonaws.com" do
    expect(AwsS3URI.new("http://foo.s3-fips.us-west-1.amazonaws.com")).to have_attributes(region: "us-west-1", bucket: "foo", key: nil)
  end

  it "parses http://BUCKET.s3-fips.REGION.amazonaws.com/KEY" do
    expect(AwsS3URI.new("http://foo.s3-fips.us-west-1.amazonaws.com/bar/baz")).to have_attributes(region: "us-west-1", bucket: "foo", key: "bar/baz")
  end

  it "parses http://BUCKET.s3.dualstack.REGION.amazonaws.com" do
    expect(AwsS3URI.new("http://foo.s3.dualstack.us-west-1.amazonaws.com")).to have_attributes(region: "us-west-1", bucket: "foo", key: nil)
  end

  it "parses http://BUCKET.s3.dualstack.REGION.amazonaws.com/KEY" do
    expect(AwsS3URI.new("http://foo.s3.dualstack.us-west-1.amazonaws.com/bar/baz")).to have_attributes(region: "us-west-1", bucket: "foo", key: "bar/baz")
  end

  it "parses http://s3.amazonaws.com/BUCKET" do
    expect(AwsS3URI.new("http://s3.amazonaws.com/foo")).to have_attributes(region: "us-east-1", bucket: "foo", key: nil)
  end

  it "parses http://s3.amazonaws.com/BUCKET/KEY" do
    expect(AwsS3URI.new("http://s3.amazonaws.com/foo/bar/baz")).to have_attributes(region: "us-east-1", bucket: "foo", key: "bar/baz")
  end
end