write unit test for sft with no geometry

adeet1 · Mar 22, 2024 · c385def · c385def
1 parent e8e044d
commit c385def
Showing 1 changed file with 45 additions and 16 deletions.
diff --git a/...torage-parquet/src/test/scala/org/locationtech/geomesa/parquet/ParquetReadWriteTest.scala b/...torage-parquet/src/test/scala/org/locationtech/geomesa/parquet/ParquetReadWriteTest.scala
@@ -40,9 +40,10 @@ import java.io.{File, RandomAccessFile}
 import java.nio.file.Files
 import java.nio.file.Paths
 import scala.collection.mutable.ArrayBuffer
-
 import com.fasterxml.jackson.databind.ObjectMapper
-import com.networknt.schema.{JsonSchemaFactory, SpecVersion}
+import com.networknt.schema.{JsonSchemaFactory, SpecVersion, ValidationMessage}
+
+import scala.collection.mutable
 
 @RunWith(classOf[JUnitRunner])
 class ParquetReadWriteTest extends Specification with AllExpectations with LazyLogging {
@@ -116,6 +117,20 @@ class ParquetReadWriteTest extends Specification with AllExpectations with LazyL
     }
   }
 
+  // Helper method that validates the file metadata against the GeoParquet metadata json schema
+  def validateMetadata(metadataString: String): mutable.Set[ValidationMessage] = {
+    val schema = {
+      val schemaFile = new File(getClass.getClassLoader.getResource("geoparquet-metadata-schema.json").toURI)
+      val schemaReader = scala.io.Source.fromFile(schemaFile)
+      val schemaString = schemaReader.mkString
+      schemaReader.close()
+      JsonSchemaFactory.getInstance(SpecVersion.VersionFlag.V4).getSchema(schemaString)
+    }
+    val metadata = new ObjectMapper().readTree(metadataString)
+
+    schema.validate(metadata).asScala
+  }
+
   "SimpleFeatureParquetWriter" should {
 
     "fail if a corrupt parquet file is written" >> {
@@ -138,31 +153,45 @@ class ParquetReadWriteTest extends Specification with AllExpectations with LazyL
       }
     }
 
-    "write parquet files" >> {
+    "write geoparquet files" >> {
       val writer = SimpleFeatureParquetWriter.builder(new Path(f.toUri), sftConf).build()
       WithClose(writer) { writer =>
         features.foreach(writer.write)
       }
-
       Files.size(f) must beGreaterThan(0L)
 
-      // Check that the metadata in the file validates against the GeoParquet json schema
-      val schema = {
-        val schemaFile = new File(getClass.getClassLoader.getResource("geoparquet-metadata-schema.json").toURI)
-        val schemaReader = scala.io.Source.fromFile(schemaFile)
-        val schemaString = schemaReader.mkString
-        schemaReader.close()
-        JsonSchemaFactory.getInstance(SpecVersion.VersionFlag.V4).getSchema(schemaString)
+      val metadata = writer.getFooter.getFileMetaData.getKeyValueMetaData.get(GeoParquetSchemaKey)
+      validateMetadata(metadata) must beEmpty
+    }
+
+    "write a geoparquet file for an sft with no geometry" >> {
+      val f = Files.createTempFile("geomesa", ".parquet")
+      val sft = SimpleFeatureTypes.createType("test", "name:String,age:Int,dtg:Date")
+      val sftConf = {
+        val c = new Configuration()
+        StorageConfiguration.setSft(c, sft)
+        // Use GZIP in tests but snappy in prod due to license issues
+        c.set(ParquetCompressionOpt, CompressionCodecName.GZIP.toString)
+        c
       }
-      val metadata = {
-        val metadataString = writer.getFooter.getFileMetaData.getKeyValueMetaData.get(GeoParquetSchemaKey)
-        new ObjectMapper().readTree(metadataString)
+
+      val features = Seq(
+        ScalaSimpleFeature.create(sft, "1", "first", 100, "2017-01-01T00:00:00Z"),
+        ScalaSimpleFeature.create(sft, "2", null, 200, "2017-01-02T00:00:00Z"),
+        ScalaSimpleFeature.create(sft, "3", "third", 300, "2017-01-03T00:00:00Z")
+      )
+
+      val writer = SimpleFeatureParquetWriter.builder(new Path(f.toUri), sftConf).build()
+      WithClose(writer) { writer =>
+        features.foreach(writer.write)
       }
+      Files.size(f) must beGreaterThan(0L)
 
-      schema.validate(metadata).asScala must beEmpty
+      val metadata = writer.getFooter.getFileMetaData.getKeyValueMetaData.get(GeoParquetSchemaKey)
+      validateMetadata(metadata) must beEmpty
     }
 
-    "read parquet files" >> {
+    "read geoparquet files" >> {
       val result = readFile(FilterCompat.NOOP, sftConf)
       result mustEqual features
     }