This is a proof-of-concept for an application capable of publishing a "micro-archive" as a bare-bones website. A micro-archive here is basically a set of photographs or document scans in some folder structure. The output is a website published on AWS Cloudfront that includes a IIIF Viewer for browsing the images, and the generation of EAD 2002 (XML) metadata.
This app, as it currently stands, also requires a fair bit of configuration and additional infrastructure, namely:
- an AWS S3 bucket, accessible to the app, where the image collections are located
- a IIIF image server (e.g. Cantaloupe) that can also access the source images
There is a fair bit of configuration needed so the Streamlit secrets TOML file looks something like this:
[s3_credentials]
access_key = "..." # The AWS access key
secret_key = "..." # The AWS secret
region = "..." # e.g. eu-west-1
bucket = "..." # The AWS S3 bucket name
[iiif]
server_url = "https://www.example.com/iiif/3/" # replace with actual IIIF URL
[datasets]
# Dataset key should be an AWS S3 prefix to the source data
# including a trailing '/'
[datasets."collection-1/"]
name = "Collection 1"
format = ".jpg"
[datasets."collection-2/"]
name = "Collection 2"
format = ".png"
To work correctly the AWS permissions need to be set up so that in addition to having read access to the files on S3, the IAM user can also create Cloudfront distributions.