LUCENE-9152: Improve line intersection detection for polygons (#1187)

lucene/CHANGES.txt

@@ -53,6 +53,8 @@ Improvements
 
 * LUCENE-4702: Better compression of terms dictionaries. (Adrien Grand)
 
+* LUCENE-9152: Improve line intersections with polygons when they are touching from the outside. (Ignacio Vera)
+
 Optimizations
 ---------------------
 

lucene/core/src/java/org/apache/lucene/geo/EdgeTree.java

@@ -261,7 +261,7 @@ public class EdgeTree {
   }
 
   /** Returns true if the line crosses any edge in this edge subtree */
-  protected boolean crossesLine(double minX, double maxX, double minY, double maxY, double a2x, double a2y, double b2x, double b2y) {
+  protected boolean crossesLine(double minX, double maxX, double minY, double maxY, double a2x, double a2y, double b2x, double b2y, boolean includeBoundary) {
     if (minY <= max) {
       double a1x = x1;
       double a1y = y1;
@@ -272,14 +272,21 @@ public class EdgeTree {
           (a1y > maxY && b1y > maxY) ||
           (a1x < minX && b1x < minX) ||
           (a1x > maxX && b1x > maxX);
-      if (outside == false && lineCrossesLineWithBoundary(a1x, a1y, b1x, b1y, a2x, a2y, b2x, b2y)) {
+      if (outside == false) {
+        if (includeBoundary) {
+          if (lineCrossesLineWithBoundary(a1x, a1y, b1x, b1y, a2x, a2y, b2x, b2y)) {
+            return true;
+          }
+        } else {
+          if (lineCrossesLine(a1x, a1y, b1x, b1y, a2x, a2y, b2x, b2y)) {
+            return true;
+          }
+        }
+      }
+      if (left != null && left.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y, includeBoundary)) {
         return true;
       }
-
+      if (right != null && maxY >= low && right.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y, includeBoundary)) {
-      if (left != null && left.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y)) {
-        return true;
-      }
-      if (right != null && maxY >= low && right.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y)) {
         return true;
       }
     }

lucene/core/src/java/org/apache/lucene/geo/Line2D.java

@@ -109,19 +109,19 @@ public final class Line2D implements Component2D {
       }
     } else if (ax == cx && ay == cy) {
       // indexed "triangle" is a line:
-      if (tree.crossesLine(minX, maxX, minY, maxY, ax, ay, bx, by)) {
+      if (tree.crossesLine(minX, maxX, minY, maxY, ax, ay, bx, by, true)) {
         return Relation.CELL_CROSSES_QUERY;
       }
       return Relation.CELL_OUTSIDE_QUERY;
     } else if (ax == bx && ay == by) {
       // indexed "triangle" is a line:
-      if (tree.crossesLine(minX, maxX, minY, maxY, bx, by, cx, cy)) {
+      if (tree.crossesLine(minX, maxX, minY, maxY, bx, by, cx, cy, true)) {
         return Relation.CELL_CROSSES_QUERY;
       }
       return Relation.CELL_OUTSIDE_QUERY;
     } else if (bx == cx && by == cy) {
       // indexed "triangle" is a line:
-      if (tree.crossesLine(minX, maxX, minY, maxY, cx, cy, ax, ay)) {
+      if (tree.crossesLine(minX, maxX, minY, maxY, cx, cy, ax, ay, true)) {
         return Relation.CELL_CROSSES_QUERY;
       }
       return Relation.CELL_OUTSIDE_QUERY;
@@ -149,7 +149,7 @@ public final class Line2D implements Component2D {
     // if any of the edges intersects an the edge belongs to the shape then it cannot be within.
     // if it only intersects edges that do not belong to the shape, then it is a candidate
     // we skip edges at the dateline to support shapes crossing it
-    if (tree.crossesLine(minX, maxX, minY, maxY, ax, ay, bx, by)) {
+    if (tree.crossesLine(minX, maxX, minY, maxY, ax, ay, bx, by, true)) {
       if (ab == true) {
         return WithinRelation.NOTWITHIN;
       } else {
@@ -157,14 +157,14 @@ public final class Line2D implements Component2D {
       }
     }
 
-    if (tree.crossesLine(minX, maxX, minY, maxY, bx, by, cx, cy)) {
+    if (tree.crossesLine(minX, maxX, minY, maxY, bx, by, cx, cy, true)) {
       if (bc == true) {
         return WithinRelation.NOTWITHIN;
       } else {
         relation = WithinRelation.CANDIDATE;
       }
     }
-    if (tree.crossesLine(minX, maxX, minY, maxY, cx, cy, ax, ay)) {
+    if (tree.crossesLine(minX, maxX, minY, maxY, cx, cy, ax, ay, true)) {
       if (ca == true) {
         return WithinRelation.NOTWITHIN;
       } else {

lucene/core/src/java/org/apache/lucene/geo/Polygon2D.java

@@ -187,7 +187,7 @@ public class Polygon2D implements Component2D {
     // if any of the edges intersects an the edge belongs to the shape then it cannot be within.
     // if it only intersects edges that do not belong to the shape, then it is a candidate
     // we skip edges at the dateline to support shapes crossing it
-    if (tree.crossesLine(minX, maxX, minY, maxY, ax, ay, bx, by)) {
+    if (tree.crossesLine(minX, maxX, minY, maxY, ax, ay, bx, by, true)) {
       if (ab == true) {
         return WithinRelation.NOTWITHIN;
       } else {
@@ -195,14 +195,14 @@ public class Polygon2D implements Component2D {
       }
     }
 
-    if (tree.crossesLine(minX, maxX, minY, maxY, bx, by, cx, cy)) {
+    if (tree.crossesLine(minX, maxX, minY, maxY, bx, by, cx, cy, true)) {
       if (bc == true) {
         return WithinRelation.NOTWITHIN;
       } else {
         relation = WithinRelation.CANDIDATE;
       }
     }
-    if (tree.crossesLine(minX, maxX, minY, maxY, cx, cy, ax, ay)) {
+    if (tree.crossesLine(minX, maxX, minY, maxY, cx, cy, ax, ay, true)) {
       if (ca == true) {
         return WithinRelation.NOTWITHIN;
       } else {
@@ -236,12 +236,12 @@ public class Polygon2D implements Component2D {
     }
 
     if (numCorners == 2) {
-      if (tree.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y)) {
+      if (tree.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y, false)) {
         return Relation.CELL_CROSSES_QUERY;
       }
       return Relation.CELL_INSIDE_QUERY;
     } else if (numCorners == 0) {
-      if (tree.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y)) {
+      if (tree.crossesLine(minX, maxX, minY, maxY, a2x, a2y, b2x, b2y, true)) {
         return Relation.CELL_CROSSES_QUERY;
       }
       return Relation.CELL_OUTSIDE_QUERY;
@@ -263,7 +263,7 @@ public class Polygon2D implements Component2D {
       if (Component2D.pointInTriangle(minX, maxX, minY, maxY, tree.x1, tree.y1, ax, ay, bx, by, cx, cy) == true) {
         return Relation.CELL_CROSSES_QUERY;
       }
-      if (tree.crossesTriangle(minX, maxX, minY, maxY, ax, ay, bx, by, cx, cy, false)) {
+      if (tree.crossesTriangle(minX, maxX, minY, maxY, ax, ay, bx, by, cx, cy, true)) {
         return Relation.CELL_CROSSES_QUERY;
       }
       return Relation.CELL_OUTSIDE_QUERY;

lucene/core/src/test/org/apache/lucene/document/TestLatLonShape.java

@@ -19,6 +19,7 @@ package org.apache.lucene.document;
 import com.carrotsearch.randomizedtesting.generators.RandomNumbers;
 import org.apache.lucene.document.ShapeField.QueryRelation;
 import org.apache.lucene.geo.Component2D;
+import org.apache.lucene.geo.GeoEncodingUtils;
 import org.apache.lucene.geo.GeoTestUtil;
 import org.apache.lucene.geo.Line;
 import org.apache.lucene.geo.Line2D;
@@ -699,4 +700,45 @@ public class TestLatLonShape extends LuceneTestCase {
 
     IOUtils.close(w, reader, dir);
   }
+
+  public void testIndexAndQuerySamePolygon() throws Exception {
+    Directory dir = newDirectory();
+    RandomIndexWriter w = new RandomIndexWriter(random(), dir);
+    Document doc = new Document();
+    Polygon polygon;
+    while(true) {
+      try {
+        polygon = GeoTestUtil.nextPolygon();
+        // quantize the polygon
+        double[] lats = new double[polygon.numPoints()];
+        double[] lons = new double[polygon.numPoints()];
+        for (int i = 0; i < polygon.numPoints(); i++) {
+          lats[i] = GeoEncodingUtils.decodeLatitude(GeoEncodingUtils.encodeLatitude(polygon.getPolyLat(i)));
+          lons[i] = GeoEncodingUtils.decodeLongitude(GeoEncodingUtils.encodeLongitude(polygon.getPolyLon(i)));
+        }
+        polygon = new Polygon(lats, lons);
+        Tessellator.tessellate(polygon);
+        break;
+      } catch (Exception e) {
+        // invalid polygon, try a new one
+      }
+    }
+    addPolygonsToDoc(FIELDNAME, doc, polygon);
+    w.addDocument(doc);
+    w.forceMerge(1);
+
+    ///// search //////
+    IndexReader reader = w.getReader();
+    w.close();
+    IndexSearcher searcher = newSearcher(reader);
+
+    Query q = LatLonShape.newPolygonQuery(FIELDNAME, QueryRelation.WITHIN, polygon);
+    assertEquals(1, searcher.count(q));
+    q = LatLonShape.newPolygonQuery(FIELDNAME, QueryRelation.INTERSECTS, polygon);
+    assertEquals(1, searcher.count(q));
+    q = LatLonShape.newPolygonQuery(FIELDNAME, QueryRelation.DISJOINT, polygon);
+    assertEquals(0, searcher.count(q));
+
+    IOUtils.close(w, reader, dir);
+  }
 }

lucene/core/src/test/org/apache/lucene/geo/TestPolygon2D.java

@@ -271,6 +271,23 @@ public class TestPolygon2D extends LuceneTestCase {
     assertTrue(poly.contains(-2, 0));  // left side: true
     assertTrue(poly.contains(-2, 1));  // left side: true
   }
+
+  /** Tests edge case behavior with respect to insideness */
+  public void testIntersectsSameEdge() {
+    Component2D poly = Polygon2D.create(new Polygon(new double[] { -2, -2, 2, 2, -2 }, new double[] { -2, 2, 2, -2, -2 }));
+    // line inside edge
+    assertEquals(Relation.CELL_INSIDE_QUERY, poly.relateTriangle(-1, -1, 1, 1, -1, -1));
+    assertEquals(Relation.CELL_INSIDE_QUERY, poly.relateTriangle(-2, -2, 2, 2, -2, -2));
+    // line over edge
+    assertEquals(Relation.CELL_CROSSES_QUERY, poly.relateTriangle(-4, -4, 4, 4, -4, -4));
+    assertEquals(Relation.CELL_CROSSES_QUERY, poly.relateTriangle(-2, -2, 4, 4, 4, 4));
+    // line inside edge
+    assertEquals(Relation.CELL_CROSSES_QUERY, poly.relateTriangle(-1, -1, 3, 3, 1, 1));
+    assertEquals(Relation.CELL_CROSSES_QUERY, poly.relateTriangle(-2, -2, 3, 3, 2, 2));
+    // line over edge
+    assertEquals(Relation.CELL_CROSSES_QUERY, poly.relateTriangle(-4, -4, 7, 7, 4, 4));
+    assertEquals(Relation.CELL_CROSSES_QUERY, poly.relateTriangle(-2, -2, 7, 7, 4, 4));
+  }
   
   /** Tests current impl against original algorithm */
   public void testContainsAgainstOriginal() {