Merge branch 'dev'

Author: codingmiao

README.md

@@ -5,14 +5,14 @@ a spatial index for neo4j 4.x.
 
 ## What can it do
 
-create a spatial index
+### create a spatial index
 ~~~java
 //5个参数依次是:  neo4j的GraphDatabaseService实例    索引名(唯一) 空间属性名   rtree最大子节点数 最大缓存geometry对象数
 RTreeIndex rTreeIndex = RTreeIndexManager.createIndex(db, "index1", "geometry", 64, 1024);
 ~~~
 
 
-add node(s) to spatial index
+### add node(s) to spatial index
 ~~~java
 Transaction tx = db.beginTx();
 Node node = tx.createNode(testLabel);//create node
@@ -23,7 +23,7 @@ rTreeIndex.add(node,tx);//add to spatial index(if you have many vertices, add as
 
 ~~~
 
-spatial query
+### spatial query
 ~~~java
 //Enter a rectangle and query the node covered by the rectangle
 double[] bbox = new double[]{3, 1, 8, 9};
@@ -45,7 +45,7 @@ try (Transaction tx = db.beginTx()) {
 }
 ~~~
 
-nearest neighbor search
+### nearest neighbor search
 ~~~java
 //Query the 5 nodes closest to point (10.2, 13.2)
 try (Transaction tx = db.beginTx()) {
@@ -61,11 +61,22 @@ try (Transaction tx = db.beginTx()) {
     System.out.println(res);//DistanceResult里包含了node、距离以及geometry，详见测试用例
 }
 ~~~
-
+### Faster spatial analysis of big geometry
+As for very fine geometry, because it contains too many points, the performance of spatial relationship calculation by using JTS's ``intersects`` function directly will be poor. At this point, it can be considered to use the Tool BigShapeManager to cut it open and establish index before analysis:
+~~~java
+String wkt = "...";//一个很多点构成的多边形
+Geometry geometry = wktReader.read(wkt);
+String indexId = "123";//声明一个id来构建索引，注意id唯一性
+BigShapeManager.build(graphDb, indexId, geometry, 10, 10);//切割多边形为10行10列并构建索引
+BigShape bigShape = BigShapeManager.get(graphDb, indexId);//获取到BigShape对象
+Geometry s1 = new WKTReader().read("POINT (866.8184900283813 132.99309968948364)");
+System.out.println(bigShape.intersects(tx, s1));//利用BigShape来做空间关系计算
+~~~
+The actual measurement performance of BigShape is up to 100 times faster than that of direct JTS for fine and complex geometry (simple geometry does not need BigShape cutting, otherwise the performance may decline)
 
 
 ## install
-The latest version is `1.3.0`
+The latest version is `1.4.0`
 
 maven import in your project
 ```

README_zh.md

@@ -5,14 +5,14 @@
 
 ## 这个项目能干什么
 
-新建空间索引
+### 新建空间索引
 ~~~java
 //5个参数依次是:  neo4j的GraphDatabaseService实例    索引名(唯一) 空间属性名   rtree最大子节点数 最大缓存geometry对象数
 RTreeIndex rTreeIndex = RTreeIndexManager.createIndex(db, "index1", "geometry", 64, 1024);
 ~~~
 
 
-简洁地为node加入空间索引:
+### 简洁地为node加入空间索引:
 ~~~java
 Transaction tx = db.beginTx();
 Node node = tx.createNode(testLabel);//新建节点
@@ -23,7 +23,7 @@ rTreeIndex.add(node,tx);//加入索引(效率起见，多个node的话用list ad
 
 ~~~
 
-空间查询
+### 空间查询
 ~~~java
 //输入一个矩形范围，查询矩形覆盖的节点
 double[] bbox = new double[]{3, 1, 8, 9};
@@ -45,7 +45,7 @@ try (Transaction tx = db.beginTx()) {
 }
 ~~~
 
-最邻近搜索
+### 最邻近搜索
 ~~~java
 //查询距离点(10.2, 13.2)最近的5个node
 try (Transaction tx = db.beginTx()) {
@@ -61,11 +61,22 @@ try (Transaction tx = db.beginTx()) {
     System.out.println(res);//DistanceResult里包含了node、距离以及geometry，详见测试用例
 }
 ~~~
-
+### 精细对象的快速空间分析
+对于非常精细的geometry，由于它包含了太多的点，直接用jts的intersects等函数去进行空间关系计算性能会很差，此时，可以考虑用BigShapeManager工具将其切割开并建立索引后再分析：
+~~~java
+String wkt = "...";//一个很多点构成的多边形
+Geometry geometry = wktReader.read(wkt);
+String indexId = "123";//声明一个id来构建索引，注意id唯一性
+BigShapeManager.build(graphDb, indexId, geometry, 10, 10);//切割多边形为10行10列并构建索引
+BigShape bigShape = BigShapeManager.get(graphDb, indexId);//获取到BigShape对象
+Geometry s1 = new WKTReader().read("POINT (866.8184900283813 132.99309968948364)");
+System.out.println(bigShape.intersects(tx, s1));//利用BigShape来做空间关系计算
+~~~
+实测对于精细复杂的geometry，BigShape的性能比直接jts快了上百倍（简单的geometry就不要BigShape切割了，否则可能反而性能下降）
 
 
 ## install
-The latest version is `1.3.0`
+The latest version is `1.4.0`
 
 maven import in your project
 ```

pom.xml

@@ -12,7 +12,7 @@
     </properties>
     <groupId>org.wowtools</groupId>
     <artifactId>neo4j-rtree</artifactId>
-    <version>1.4-SNAPSHOT</version>
+    <version>1.4.0</version>
     <name>neo4j-rtree</name>
 
     <description>a spatial index for neo4j 4.x.</description>