Phenoscape web UI planning

Discussion, 2008-10-15

Cartik, Jim, and Hilmar met to discuss web service requirements for the Phenoscape web application.

Services

Term info

• Could use OLS, but makes more sense to retrieve exact data we're working off of in the database
• Input: term ID
• Output: name, parents & children (relation, ID, name), definition, other properties?
• Could return numerous formats (JSON, OBO, OBD-XML, OWL/RDF)
  • JSON is most useful for web application
  • OWL/RDF may be most useful for a standards-based alternative

"Anatomy search"

Didn't really cover this much in the discussion. What format should summary counts be returned in? These are then used to navigate to further data services.

Publication detail

This could potentially be handled by the term info service, if the client knew which properties to traverse from the publication to get the information it needed. This is a trade-off between general and more specific services.

Short-term plans

Cartik

• By Oct. 30
  • Develop script for loading data files into database
    • should be run both automated nightly and on-demand
    • loading tool should check out latest data files from Phenoscape SVN, and load latest versions of all needed ontologies
    • this could be a collection of distinct tools which are chained together in a shell script
• By Nov. 6
  • Determine how to distinguish between pre- and post-composed compositional descriptions/relations - without hacks

Jim

• By Oct. 30
  • Initiate OBD-WS project in OBO repository
  • Define JSON schema for term info service
  • Prepare formal architecture draft of web application
  • Validate result of data loading script (identify missing data not yet handled by OBDModelBridge.java)
  • More comprehensively define data service requirements