NIH | National Cancer Institute | NCI Wiki  

Error rendering macro 'rw-search'

null

Versions Compared

Old Version 2

changes.mady.by.user Unknown User (endlecm)

Saved on

compared with

New Version 3

changes.mady.by.user Unknown User (endlecm)

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Immediate Priorities and Problems (Mayo Team)

 Scott

  • Finish up OWL 2 Jira items/Junits/documentation
  • Support for SNOMED loader issue - Passed the issue back to Brian Carlsen
  • External Support - helped out with Windows user
  • Browser APPSCAN support - Scott is going to review the report
  • Monthly Reports

  Cory

  •    LexEVS Active/Inactive issue - The fix was put into 6.3
    • Latest 6.3 is out on NCI DEV
    • TODO - put note in 6.2 to mention the change in behavior in 6.2.
  •    LexEVS CTS2 active/inactive issue - The fix was put into 6.3
  •    API Docs issues - Waiting for results from app scan

   Craig

    ▪    Monthly Reports, Lucene Planning - To discuss the Lucene options below.

Doc API UpdateWaiting to hear back from app scan
OWL2

Gilberto is verifying Scott's latest updates

Kim found issues with some Junits and Scott has fixed.

All OWL2 issues have been resolved.

Gilberto - Consistency with querying - nested qualifiers (to deal with the data types).  This may not be a good idea.  This forces people to do additional queries to get back all of the data.

  • Scott mentioned that there are some limitations in LexEVS and that is why the nesting is done this way.
  • Scott suggested that we write some sample code and see if it is complicated and should be simplified.
  • It is confusing seeing object wrappers wrapping other object wrappers.  This should be fixed.
  • We should consider if the changes needed from above would create a backwards compatibility issue in 6.4.

Gilberto needs to review the OWL2 Loader to determine if it is complete

AppScan for Term Browser 
SNOMED Load

Scott reviewed the load.

Brian and Joanne are looking at it now and hope to have a fix for it today.

Gilberto suggested that we document all of the dependencies on specific structures that we have in the META

  • Scott to look at the RRF loader and determine what is hard-coded and needs to be tracked.  This can be put on the wiki.
Lucene Discussion
  • Options:
    • Lucene - complete post-September
    • Lucene - complete by September - reduced scope
      • Minimal options - (This may not save a lot of time)
        • NOT to redo the simple search.
          • We were hoping to not have to do this anyway.
          • Initial testing and prototyping would determine if this is needed.
            • Build and test separate code scheme search and see if it performs well enough so we wouldn't need to redo the simple search.
          • Scope Reduction - potentially neutral
        • NOT do the modularization of the API - leave as integrated. 
          • Build out the modeling to decouple the API or build out the new Lucene API in the code.
          • Without the modularization, there is a lot of technical debt in place.  The search API is tied tightly to Lucene.
          • Scope Reduction - potentially neutral
        • Reduced testing, and prototyping.
          • This could reduce 2 - 4 weeks, assuming we don't run into any issues. 
          • Without prototyping we will have to deal with any problems as the come up.
          • Scope Reduction - potentially 2 - 4 weeks
        • Keep existing code (lbindex) - minimal code cleanup.
          • This could reduce the time by approximately 2 weeks
          • Leaving the code in is more "dead" code (technical debt).
          • Scope Reduction - potentially 2 weeks
    • Postponed Lucene - complete other JIRA items by September and push Lucene until next year.
      • We could do some prep (prototype and code decoupling) by September and then implement post-September

NCI would need to see what the issues are involved if we do option 2 (minimal implementation)

  • How can we make this work modular to complete some portion of work by end of September?  And then continue on if the contract gets extended.
    • Modularization - Our API/bridge to any search engine in the future.
    • Suggested that we do the modularization and then work on the different components as we go.
  • Do we need to decide now?  If so, NCI would need to request our work to be extended now rather than later.

NCI would like a clear statement to quantify how the Lucene modularization should be done.

Craig to send updated Lucene plan to the whole group.

...