Merge pull request #337 from s22s/fix/333

Fix for #333 and additional tests in that vein.

Author: metasim

core/src/main/scala/org/locationtech/rasterframes/expressions/aggregates/LocalMeanAggregate.scala

@@ -30,6 +30,7 @@ import org.apache.spark.sql.catalyst.expressions.{AttributeReference, Expression
 import org.apache.spark.sql.types.DataType
 import org.apache.spark.sql.{Column, TypedColumn}
 import org.locationtech.rasterframes.TileType
+import org.locationtech.rasterframes.expressions.accessors.RealizeTile
 
 @ExpressionDescription(
   usage = "_FUNC_(tile) - Computes a new tile contining the mean cell values across all tiles in column.",
@@ -58,11 +59,11 @@ case class LocalMeanAggregate(child: Expression) extends UnaryRasterAggregate {
   )
   override lazy val updateExpressions: Seq[Expression] = Seq(
     If(IsNull(count),
-      SetCellType(Defined(child), Literal("int32")),
-      If(IsNull(child), count, BiasedAdd(count, Defined(child)))
+      SetCellType(RealizeTile(Defined(child)), Literal("int32")),
+      If(IsNull(child), count, BiasedAdd(count, Defined(RealizeTile(child))))
     ),
     If(IsNull(sum),
-      SetCellType(child, Literal("float64")),
+      SetCellType(RealizeTile(child), Literal("float64")),
       If(IsNull(child), sum, BiasedAdd(sum, child))
     )
   )

core/src/main/scala/org/locationtech/rasterframes/expressions/transformers/RGBComposite.scala

@@ -33,7 +33,6 @@ import org.locationtech.rasterframes._
 import org.locationtech.rasterframes.encoders.CatalystSerializer._
 import org.locationtech.rasterframes.expressions.DynamicExtractors.tileExtractor
 import org.locationtech.rasterframes.expressions.row
-import org.locationtech.rasterframes.tiles.ProjectedRasterTile
 
 /**
   * Expression to combine the given tile columns into an 32-bit RGB composite.

core/src/main/scala/org/locationtech/rasterframes/extensions/DataFrameMethods.scala

@@ -22,7 +22,6 @@
 package org.locationtech.rasterframes.extensions
 
 import geotrellis.proj4.CRS
-import geotrellis.raster.{MultibandTile, ProjectedRaster}
 import geotrellis.spark.io._
 import geotrellis.spark.{SpaceTimeKey, SpatialComponent, SpatialKey, TemporalKey, TileLayerMetadata}
 import geotrellis.util.MethodExtensions
@@ -33,9 +32,7 @@ import org.apache.spark.sql.{Column, DataFrame, TypedColumn}
 import org.locationtech.rasterframes.StandardColumns._
 import org.locationtech.rasterframes.encoders.CatalystSerializer._
 import org.locationtech.rasterframes.encoders.StandardEncoders._
-import org.locationtech.rasterframes.expressions.{DynamicExtractors, aggregates}
-import org.locationtech.rasterframes.expressions.aggregates.TileRasterizerAggregate
-import org.locationtech.rasterframes.model.TileDimensions
+import org.locationtech.rasterframes.expressions.DynamicExtractors
 import org.locationtech.rasterframes.tiles.ProjectedRasterTile
 import org.locationtech.rasterframes.util._
 import org.locationtech.rasterframes.{MetadataKeys, RasterFrameLayer}

core/src/test/scala/org/locationtech/rasterframes/RasterFunctionsSpec.scala

@@ -525,11 +525,11 @@ class RasterFunctionsSpec extends TestEnvironment with RasterMatchers {
       checkDocs("rf_agg_local_max")
     }
 
-    ignore("should compute local mean") {
+    it("should compute local mean") {
       checkDocs("rf_agg_local_mean")
-      // https://github.com/locationtech/rasterframes/issues/333
       val df = Seq(two, three, one, six).toDF("tile")
           .withColumn("id", monotonically_increasing_id())
+
       df.select(rf_agg_local_mean($"tile")).first() should be(three.toArrayTile())
 
       df.selectExpr("rf_agg_local_mean(tile)").as[Tile].first() should be(three.toArrayTile())
@@ -539,7 +539,6 @@ class RasterFunctionsSpec extends TestEnvironment with RasterMatchers {
           .agg(rf_agg_local_mean($"tile"))
           .collect()
       }
-
     }
 
     it("should compute local data cell counts") {

core/src/test/scala/org/locationtech/rasterframes/TileStatsSpec.scala

@@ -24,6 +24,7 @@ package org.locationtech.rasterframes
 import geotrellis.raster._
 import geotrellis.raster.mapalgebra.local.{Max, Min}
 import geotrellis.spark._
+import org.apache.spark.sql.Column
 import org.apache.spark.sql.functions._
 import org.locationtech.rasterframes.TestData.randomTile
 import org.locationtech.rasterframes.stats.CellHistogram
@@ -317,4 +318,34 @@ class TileStatsSpec extends TestEnvironment with TestData {
       ndCount2 should be(count + 1)
     }
   }
+
+  describe("proj_raster handling") {
+    it("should handle proj_raster structures") {
+      val df = Seq(lazyPRT, lazyPRT).toDF("tile")
+
+      val targets = Seq[Column => Column](
+        rf_is_no_data_tile,
+        rf_data_cells,
+        rf_no_data_cells,
+        rf_agg_local_max,
+        rf_agg_local_min,
+        rf_agg_local_mean,
+        rf_agg_local_data_cells,
+        rf_agg_local_no_data_cells,
+        rf_agg_local_stats,
+        rf_agg_approx_histogram,
+        rf_tile_histogram,
+        rf_tile_stats,
+        rf_tile_mean,
+        rf_tile_max,
+        rf_tile_min
+      )
+
+      forEvery(targets) { f =>
+        noException shouldBe thrownBy {
+          df.select(f($"tile")).collect()
+        }
+      }
+    }
+  }
 }