Paul is the "data grabber" application that prepares data and metadata
files for ingestion into the CMM MIRAGE repository as follows:
- It receives file create / modify / delete events for files in the directory
trees that correspond to the different instruments.
- It aggregates these based on filename patterns to work out sets of files
that constitute Datasets.
- When a Dataset stabilizes, it "grabs" the files by copying them to a separate
queue area.
- It uses the information recorded by Eccles to figure out who was logged into
the facility. If this is known, the Dataset is tagged with the user and
marked as ready to be ingested.
- Ingestible Datasets are then exposed to the downstream system (e.g. MyTardis)
via an Atom feed.
Paul also provides a web UI that supports ACLS session control, queue management
integrity checking and repair, and configuration of instruments.
Eccles is a proxy for an ACLS service that tracks sessions, and will
insulate the ACLS clients from server outages. Eventually, it may evolve into
a replacement for ACLS.
Benny is simple HTTP service for authenticating against a backend ACLS instance.