This is a small Ruby (not Rails) project meant to run on Heroku, whose only purpose is to connect to our ArchivesSpace's API once a week and export .ead.xml files to a location on S3. They can be harvested there by various partners. All the important code can be found in export_archivesspace_xml/lib/exporter.rb.
More documentation can be found in the wiki.
The files are uploaded to an s3 bucket; this is publicly accessible via a cname record at http://ead.sciencehistory.org/ .
index_page.rb creates a very simple index.html file in the bucket. The file allows our partners to use a variation on the following command to download all our EAD files:
wget -r http://ead.sciencehistory.org -A *.ead.xml
Because the Heroku dyno is assigned an arbitrary IP address when it spins up, we proxy outgoing traffic through a static address using the Proximo Heroku add-on. This allows us to tell the ArchivesSpace API to accept connections only from the one whitelisted static IP address. See the Proximo docs for more details.
We maintain a description of the app's infrastructure, such as S3 buckets, in Terraform (details).
This is done via environment variables set on the Heroku project. Here are some of the important ones:
These allow the code to contact ArchivesSpace and download the EADs.
ARCHIVESSPACE_URLARCHIVESSPACE_EXPORT_USERNAMEARCHIVESSPACE_EXPORT_PASSWORD
These are needed so the code knows where to put the files.
AWS_BUCKETAWS_REGIONAWS_ACCESS_KEY_IDAWS_SECRET_ACCESS_KEYNote: The IAM permissions associated with this key pair in S3 are minimal: the code can only write files to the ead bucket.
We don't manage these- they're set by Heroku for our add-ons.
PROXIMO_URLPAPERTRAIL_API_TOKEN
The project does not include a web dyno, and relies on the Heroku Scheduler to spin up a nightly process, as follows:
$ bin/proximo bundle exec ruby run.rb