Details

    • Type: Improvement Improvement
    • Status: Resolved
    • Priority: Major Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None
    • Lucene Fields:
      New

      Description

      This is an approach to indexing bounding boxes using 4 numeric fields (xmin,ymin,xmax,ymax) and a flag to say if it crosses the dateline.

      This is a modification from the Apache 2.0 code from the ESRI Geoportal:
      http://geoportal.svn.sourceforge.net/svnroot/geoportal/Geoportal/trunk/src/com/esri/gpt/catalog/lucene/SpatialClauseAdapter.java

        Issue Links

          Activity

          Hide
          Ryan McKinley added a comment -

          Here is a BBox strategy with some simple tests that touch each of the supported operations.

          This also passes all the other (weak) spatial tests

          Show
          Ryan McKinley added a comment - Here is a BBox strategy with some simple tests that touch each of the supported operations. This also passes all the other (weak) spatial tests
          Hide
          Ryan McKinley added a comment -

          I just added this to trunk in r1354841

          I think we should ignore (remove?) 4.x for now... there will be some heavy committing in this area soon and keeping two branches is not very helpful.

          I'll take a stab at how this could look with LUCENE-4167 (remove SpatialOperation)

          Show
          Ryan McKinley added a comment - I just added this to trunk in r1354841 I think we should ignore (remove?) 4.x for now... there will be some heavy committing in this area soon and keeping two branches is not very helpful. I'll take a stab at how this could look with LUCENE-4167 (remove SpatialOperation)
          Hide
          Chris Male added a comment -

          I think we should ignore (remove?) 4.x for now... there will be some heavy committing in this area soon and keeping two branches is not very helpful.

          what?

          Show
          Chris Male added a comment - I think we should ignore (remove?) 4.x for now... there will be some heavy committing in this area soon and keeping two branches is not very helpful. what?
          Hide
          Steve Rowe added a comment -

          Ryan,

          Jenkins builds are failing. e.g. http://jenkins.sd-datasolutions.de/job/Lucene-Solr-trunk-Linux-Java6-64/1082/, as a result of a missing package javadocs file for org.apache.lucene.spatial.bbox:

          javadocs-lint:
               [exec] 
               [exec] Crawl/parse...
          ...
               [exec] 
               [exec] Verify...
               [exec] 
               [exec] build/docs/spatial/overview-summary.html
               [exec]   missing: org.apache.lucene.spatial.bbox
               [exec] 
               [exec] Missing javadocs were found!
          
          Show
          Steve Rowe added a comment - Ryan, Jenkins builds are failing. e.g. http://jenkins.sd-datasolutions.de/job/Lucene-Solr-trunk-Linux-Java6-64/1082/ , as a result of a missing package javadocs file for org.apache.lucene.spatial.bbox : javadocs-lint: [exec] [exec] Crawl/parse... ... [exec] [exec] Verify... [exec] [exec] build/docs/spatial/overview-summary.html [exec] missing: org.apache.lucene.spatial.bbox [exec] [exec] Missing javadocs were found!
          Hide
          David Smiley added a comment -

          Cool!

          Perhaps the (small) infrastructure changes to SpatialTestCase & StrategyTestCase should be done to 4x. (an aside: Doing development against two branches is a pain. I'm glad I don't deal with this overhead for any other projects)

          Show
          David Smiley added a comment - Cool! Perhaps the (small) infrastructure changes to SpatialTestCase & StrategyTestCase should be done to 4x. (an aside: Doing development against two branches is a pain. I'm glad I don't deal with this overhead for any other projects)
          Hide
          Robert Muir added a comment -

          Is it still by intention that this is not backported to 4.x? I just want to be sure.

          Show
          Robert Muir added a comment - Is it still by intention that this is not backported to 4.x? I just want to be sure.
          Hide
          David Smiley added a comment - - edited

          The BBoxStrategy is tested insofar as search filtering and returning the standard center-point distance as a ValueSource. But it contains a lot of code pertaining to an area overlap similarity as a ValueSource (very cool stuff) but that isn't tested at all, and that's my only concern. If we want to be cautious about avoiding releasing untested code, then I suggest back-porting BBoxStrategy to 4x (pretty easy) but not the AreaSimilarity class. The reference to it in BBoxStrategy is the method makeBBoxAreaSimilarityValueSource() which can be commented out in 4x.

          Sound good?

          Show
          David Smiley added a comment - - edited The BBoxStrategy is tested insofar as search filtering and returning the standard center-point distance as a ValueSource. But it contains a lot of code pertaining to an area overlap similarity as a ValueSource (very cool stuff) but that isn't tested at all, and that's my only concern. If we want to be cautious about avoiding releasing untested code, then I suggest back-porting BBoxStrategy to 4x (pretty easy) but not the AreaSimilarity class. The reference to it in BBoxStrategy is the method makeBBoxAreaSimilarityValueSource() which can be commented out in 4x. Sound good?
          Hide
          Chris Male added a comment -

          With the very near release of 4.0, I don't think we should backport anything untested. I also don't think we're in any immediate hurry for this since we've got other options in 4.0. But we should definitely work on the testing and push it for 4.1.

          Show
          Chris Male added a comment - With the very near release of 4.0, I don't think we should backport anything untested. I also don't think we're in any immediate hurry for this since we've got other options in 4.0. But we should definitely work on the testing and push it for 4.1.
          Hide
          David Smiley added a comment -

          FYI the improved testing of this strategy and 4x port is on LUCENE-5714. Oh, what a coincidence – the issue number is reversed from this issue, LUCENE-4175. I swear I wasn't waiting till it would happen

          Show
          David Smiley added a comment - FYI the improved testing of this strategy and 4x port is on LUCENE-5714 . Oh, what a coincidence – the issue number is reversed from this issue, LUCENE-4175 . I swear I wasn't waiting till it would happen

            People

            • Assignee:
              Ryan McKinley
              Reporter:
              Ryan McKinley
            • Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development