Browse everything s3 (#532)

* config for be-s3

* set bucket name and region direct in deploy yaml

* use existing iam user so no need for these here

* use existing defaults for s3 access

* 🐛 Override BrowseEverything for S3 file size

This commit will bring in an override for the BrowseEverything gem to
retrieve the file size from an S3 object and pass it into the params so
that the Bulkrax importer can use it.

* remove file-system option from BE providers

---------

Co-authored-by: Kirk Wang <[email protected]>

config/application.rb

@@ -36,6 +36,10 @@ class Application < Rails::Application
  Dir.glob(File.join(File.dirname(__FILE__), "../app/**/*_decorator*.rb")).sort.each do |c|
  Rails.configuration.cache_classes ? require(c) : load(c)
  end
+
+ Dir.glob(File.join(File.dirname(__FILE__), "../lib/**/*_decorator*.rb")).sort.each do |c|
+ Rails.configuration.cache_classes ? require(c) : load(c)
+ end
  end
 
  # resolve reloading issue in dev mode

config/browse_everything_providers.yml

@@ -2,8 +2,8 @@
 # To make browse-everything aware of a provider, uncomment the info for that provider and add your API key information.
 # The file_system provider can be a path to any directory on the server where your application is running.
 #
-file_system:
- home: /app/samvera/hyrax-webapp/data
+#file_system:
+# home: /app/samvera/hyrax-webapp/data
 # dropbox:
 # client_id: YOUR_DROPBOX_APP_KEY
 # client_secret: YOUR_DROPBOX_APP_SECRET
@@ -14,11 +14,11 @@ file_system:
 # google_drive:
 # client_id: YOUR_GOOGLE_API_CLIENT_ID
 # client_secret: YOUR_GOOGLE_API_CLIENT_SECRET
-# s3:
-# bucket: YOUR_AWS_S3_BUCKET
-#  response_type: signed_url # set to :public_url for public urls or :s3_uri for an s3://BUCKET/KEY uri
-#  expires_in: 14400 # for signed_url response_type, number of seconds url will be valid for.
-# app_key: YOUR_AWS_S3_KEY  # :app_key, :app_secret, and :region can be specified
-# app_secret: YOUR_AWS_S3_SECRET # explicitly here, or left out to use system-configured
-# region: YOUR_AWS_S3_REGION  # defaults.
-#  See https://aws.amazon.com/blogs/security/a-new-and-standardized-way-to-manage-credentials-in-the-aws-sdks/
+s3:
+ bucket: temp-bl-bucket-for-browse-everything # .s3.amazonaws.com #arn:aws:s3:::temp-bl-bucket-for-browse-everything
+ response_type: signed_url # set to :public_url for public urls or :s3_uri for an s3://BUCKET/KEY uri
+ expires_in: 14400 # for signed_url response_type, number of seconds url will be valid for.
+ app_key: <%= ENV['AWS_ACCESS_KEY_ID'] %> # :app_key, :app_secret, and :region can be specified
+ app_secret: <%= ENV['AWS_SECRET_ACCESS_KEY'] %> # explicitly here, or left out to use system-configured
+ region: eu-west-1 # defaults.
+ # See https://aws.amazon.com/blogs/security/a-new-and-standardized-way-to-manage-credentials-in-the-aws-sdks/

lib/browse_everything/driver/s3_decorator.rb

@@ -0,0 +1,31 @@
+# frozen_string_literal: true
+
+# OVERRIDE BrowseEverything v1.1.2 to add file_size for S3 files
+
+module BrowseEverything
+ module Driver
+ module S3Decorator
+ def link_for(path)
+ obj = bucket.object(full_path(path))
+ obj_head = obj.head
+ file_size = obj_head.content_length
+
+ extras = {
+ file_name: File.basename(path),
+ file_size: file_size,
+ expires: (config[:expires_in] if config[:response_type] == :signed_url)
+ }.compact
+
+ url = case config[:response_type].to_sym
+ when :signed_url then obj.presigned_url(:get, expires_in: config[:expires_in])
+ when :public_url then obj.public_url
+ when :s3_uri then "s3://#{obj.bucket_name}/#{obj.key}"
+ end
+
+ [url, extras]
+ end
+ end
+ end
+end
+
+BrowseEverything::Driver::S3.prepend(BrowseEverything::Driver::S3Decorator)