make overlap threshold for area reviews configurable (#4318)

conf/core/ConfigOptions.asciidoc

@@ -289,6 +289,14 @@ input as part of the query. This is more inefficient for memory consumption but
 overall for significantly large datasets, as the cropped database query performs slowly at larger
 input sizes.
 
+=== area.overlap.review.threshold
+
+* Data Type: double
+* Default Value: `0.835`
+
+Areas that do not match and have a percentage overlap at or above this value will be marked for
+review.
+
 === area.tag.threshold
 
 * Data Type: double
@@ -318,7 +326,7 @@ option may result in some false positive road joining.
 * Data Type: bool
 * Default Value: `true`
 
-If true, allows Attribute Conflation to modify the geometry of a reference road layer to accomodate
+If true, allows Attribute Conflation to modify the geometry of a reference road layer to accommodate
 bridge features.  If false, any bridge features represented in the secondary dataset, but not in the
 reference dataset will be lost.
 

conf/services/conflationTypes.json

@@ -42,6 +42,7 @@
         "matcher": "hoot::ScriptMatchCreator,Area.js",
         "merger": "hoot::ScriptMergerCreator",
         "members": {
+            "area.overlap.review.threshold": "Overlap review threshold; 0.0 to 1.0",
             "search.radius.area": "Search radius (m); -1 for default",
             "area.tag.threshold": "Tag match threshold; 0.0 to 1.0"
         }

hoot-core/src/main/cpp/hoot/core/algorithms/extractors/SmallerOverlapExtractor.h

@@ -40,10 +40,10 @@ class Element;
 /**
  * Based on the JCS OverlapExtractor.
  *
- * This extractor uses the building with more overlap to calculate the percentage overlap. For
- * instance if you have a small building contained within a larger building the result will be 1.
+ * This extractor uses the feature with more overlap to calculate the percentage overlap. For
+ * instance if you have a small feature contained within a larger feature, the result will be 1.0.
  *
- * If either building has an area of zero then a zero is returned.
+ * If either feature has an area of zero, then zero is returned.
  *
  * @author RoadMatcher
  * @copyright GPL

rules/Area.js

@@ -46,6 +46,9 @@ nodes and polygons or a school polygon which encloses school buildings on the ca
  */
 exports.isMatchCandidate = function(map, e)
 {
+  hoot.trace("e: " + e.getElementId());
+  hoot.trace("isNonBuildingArea: " + isNonBuildingArea(map, e));
+
   return isNonBuildingArea(map, e);
 };
 
@@ -200,7 +203,8 @@ exports.matchScore = function(map, e1, e2)
       smallerOverlap = smallerOverlapExtractor.extract(map, e1, e2);
       hoot.trace("smallerOverlap: " + smallerOverlap);
     }
-    if (smallerOverlap > 0.835)
+    var overlapReviewThreshold = parseFloat(hoot.get("area.overlap.review.threshold"));
+    if (smallerOverlap >= overlapReviewThreshold)
     {
       hoot.trace("review");
       result = { match: 0.0, miss: 0.0, review: 1.0 };

rules/River.js

@@ -162,7 +162,7 @@ function geometryMismatch(map, e1, e2)
   hoot.trace(sublines);
   if (sublines && sublines == "RecursiveComplexityException")
   {
-    // If receive the specfic string above from the matching routine, we know our subline matcher
+    // If we receive the specfic string above from the matching routine, we know our subline matcher
     // hit the cap on the number of recursive calls we allow for it 
     // (see waterway.maximal.subline.max.recursive.complexity above; A little kludgy, but not sure 
     // how to handle hoot exceptions in a js script at this point). So, now we'll try a backup matcher