1. id3c
  2. dev
  3. doc
  4. design-2019-summer
  5. 09-modularization

Data logistics system enabling real-time pathogen surveillance. Built for the Seattle Flu Study.
Source code
github.com/seattleflu/id3c
Contributors
tsibley
joverlee521
kairstenfay
jccraft-uw
kschwabefry
Latest commits
Pages

Modularization

Refactor parts of ID3C which are generic from parts which are specific to the Seattle Flu Study, making a core component which is extended by customizations and plugins. This is a standard design approach and will probably be easier to do sooner than later.

The common core (epiphyla/id3c?) would contain:

  • base CLI
  • base web API
  • ETL framework
  • sqitch project with warehouse schema and some of receiving (FHIR, RDML, sequence read sets, genomes, etc)

Extension packages (seattleflu/id3c?) would contain:

  • custom ETL routines, loaded as command plugins (via setuptools entrypoints) and using id3c Python libraries

  • custom database schema (receiving, shipping) as a sqitch project with cross-project dependencies (these are supported!)

  • code to handle the "edges", the places where the core meets the outside world

Motivation

  1. Replace fauna and power the next generation of Nextstrain work.

  2. Be a re-usable data system like Augur is a re-usable bioinformatics toolkit and Auspice is a re-usable visualization tool.

Challenges

  1. Adds complexity during development with an additional repo to touch.

  2. After doing the work, there will be a one-time cutover/flag day where we go through the incantations to make it so in production. This will involve things like making use of sqitch's log-only deploys to pick up new names for already-deployed changes.