My notes from the meeting: Part A: temporal indexing.
The slides are available at https://docs.google.com/presentation/d/1d9f2AcAnjkVLJJfcYJbkAFcOJxEvZnO0iiLWW3n_CaU/edit?usp=sharing The temporal index is based on the object part of the RDF triple. There are many useful technical details in the slides. Part B: GSoC project The slides are available at https://docs.google.com/presentation/d/1c5Bk_oedGzcaYS4MzER-KONrBB6pGxGBVi0K-jRCuB4/edit?usp=sharing My take-away from the discussion: 1. For the pre-commit integration, many people recommend starting with Apache Yetus. Checks to do: * (required)make sure the project builds with the incoming pull-request, the tests pass * (nice to have, if time permits) an Intellectual Property check to ensure the correct Apache headers are in the files in the PR Once Apache Yetus is integrated with Jenkins, maybe we can try integrating SonarQube, FindBugs or PMD (this was an email after the discussion, but I included here since it fits) 2. Building a testing framework that will make it easier for people to run performance testing will be more useful than integrating such a framework with the automatic build. Integrating would be very hard and potentially not very useful/meaningful, as a Rya deployment is needed for testing. One of the goals for the GSoC project will then be to create a testing framework that is easy to deploy. Different people using Rya can use the framework on their own deployed environment, with their own data and queries. Requirements for the framework: - easy to deploy - maybe just a main is enough, no need for a web interface - loading data in Rya/sending queries to Rya should be done in a fashion similar with typical usage, which is through a Java client, not through the web interface - configurable parameters for the framework: -Accumulo vs MongoDB backend -Rya/Accumulo or Rya/MongoDB connection parameters -location of and format of data to load -queries to run -performance metrics to collect. Examples are: -Queries per second -total query time -time to first answer -time to first 10 answers -time to first 100 answers -load time JMeter http://jmeter.apache.org/ and Metrics https://github.com/dropwizard/metrics might be useful tools for this task. Thank you all for the discussion, comments, and suggestions On Wed, May 18, 2016 at 12:01 PM, David Lotts <[email protected]> wrote: > The Rya working group is starting now! > Here is a link to the Temporal slides: > > https://docs.google.com/presentation/d/1d9f2AcAnjkVLJJfcYJbkAFcOJxEvZnO0iiLWW3n_CaU/edit?usp=sharing > Also see Adina's slides > david lotts. > > On Mon, May 16, 2016 at 1:41 PM, David Lotts <[email protected]> wrote: > > > All, > > > > The Rya working group is this Wednesday at noon. This week we will be > > going over: > > > > - Temporal indexing and queries, > > - a little bit about Rya's Geo indexing, and > > - a mentioning of Rya's full-text indexing. > > > > I am the implementer of the original temporal index code, so if you have > > some ideas about an index you'd like to create, using or making tweaks to > > these indexes, this session might be what you need to get started. > > > > No screen sharing this week, but we can use the following phone > conference > > for the discussion: > > > > Call-in toll-free number: 1-888-598-1409 <1-888-5981409> > > Attendee access code: 7483409 > > > > I'll send out slides at some point before the working group. > > > > Talk to you Wednesday! > > > > David Lotts > > > > > -- Dr. Adina Crainiceanu Associate Professor, Computer Science Department United States Naval Academy 410-293-6822 [email protected] http://www.usna.edu/Users/cs/adina/
