Home

This is the wiki for Caesar, a backend automation service for Panoptes. Caesar subscribes to the classification stream produced by Panoptes by use of an Amazon Lambda connected to Kinesis that posts classifications into the REST API Caesar provides.

Features

1. Configurable automation of simple tasks (subject retirement, promotion, and sorting) in response to classification events
2. A RESTful API for workflow configuration and inspection of results
3. An HTML front-end for workflow configuration and inspection of results
4. A queue-based system for performing work asynchronously
5. Hooks to push intermediate values to external systems for highly custom configurations

Concepts

Caesar separates the process of responding to classifications into three distinct phases: feature extraction, combining classifications (reduction), and rule application.

Technologies

1. Rails for ORM / API
2. Sidekiq for job queueing
3. Faraday for HTTPS/REST