Release Notes

0.9.x

0.9.1

  • Upgraded to Spark 2.4.7
  • Added pyspark.sql.DataFrame.display(num_rows:int, truncate:bool) extension method when rf_ipython is imported.
  • Added users’ manual section on IPython display enhancements.
  • Added method_name parameter to the rf_resample method.
  • BREAKING: In SQL, the function rf_resample now takes 3 arguments. You can use rf_resample_nearest with two arguments or refactor to rf_resample(t, v, "nearest").
  • Added resample method parameter to SQL and Python APIs. See updated docs.
  • Upgraded many of the pyrasterframes dependencies, including: descartes, fiona, folium, geopandas, matplotlib, numpy, pandas, rasterio, shapely
  • Changed rasterframes.prefer-gdal configuration parameter to default to False, as JVM GeoTIFF performs just as well for COGs as the GDAL one.
  • Fixed #545.

0.9.0

  • Upgraded to GeoTrellis 3.3.0. This includes a number of breaking changes enumerated as a part of the PR’s change log. These include:
  • Add Int type parameter to Grid
  • Add Int type parameter to CellGrid
  • Add Int type parameter to GridBounds… or TileBounds
  • Use GridBounds.toGridType to coerce from Int to Long type parameter
  • Update imports for layers, particularly geotrellis.spark.tiling to geotrellis.layer
  • Update imports for geotrellis.spark.io to geotrellis.spark.store...
  • Removed FixedRasterExtent
  • Removed FixedDelegatingTile
  • Removed org.locationtech.rasterframes.util.Shims
  • Change Extent.jtsGeom to Extent.toPolygon
  • Change TileLayerMetadata.gridBounds to TileLayerMetadata.tileBounds
  • Add geotrellis-gdal dependency
  • Remove any conversions between JTS geometry and old geotrellis.vector geometry
  • Changed org.locationtech.rasterframes.encoders.StandardEncoders.crsEncoder to crsSparkEncoder
  • Change (cols, rows) dimension destructuring to Dimensions(cols, rows)
  • Revisit use of Tile equality since it’s more strict
  • Update reference.conf to use geotrellis.raster.gdal namespace.
  • Replace all uses of TileDimensions with geotrellis.raster.Dimensions[Int].
  • Upgraded to gdal-warp-bindings 1.0.0.
  • Upgraded to Spark 2.4.5
  • Formally abandoned support for Python 2. Python 2 is dead. Long live Python 2.
  • Introduction of type hints in Python API.
  • Add functions for changing cell values based on either conditions or to achieve a distribution of values. (#449)
    • Add rf_local_min, rf_local_max, and rf_local_clip functions.
    • Add cell value scaling functions rf_rescale and rf_standardize.
    • Add rf_where function, similar in spirit to numpy’s where, or a cell-wise version of Spark SQL’s when and otherwise.
  • Add rf_sqrt function to compute cell-wise square root.

0.8.x

0.8.5

  • Added rf_z2_index for constructing a Z2 index on types with bounds.
  • Breaking: rf_spatial_index renamed rf_xz2_index to differentiate between XZ2 and Z2 variants.
  • Added withSpatialIndex to RasterSourceDataSource to pre-partition tiles based on tile extents mapped to a Z2 space-filling curve
  • Add rf_mask_by_bit, rf_mask_by_bits and rf_local_extract_bits to deal with bit packed quality masks. Updated the masking documentation to demonstrate the use of these functions.
  • Added toDF extension method to MultibandGeoTiff
  • Added rf_agg_extent and rf_agg_reprojected_extent to compute the aggregate extent of a column
  • Added rf_proj_raster for constructing a proj_raster structure from individual CRS, Extent, and Tile columns.
  • Added rf_render_color_ramp_png to compute PNG byte array for a single tile column, with specified color ramp.
  • In rf_ipython, improved rendering of dataframe binary contents with PNG preamble.
  • Throw an IllegalArgumentException when attempting to apply a mask to a Tile whose CellType has no NoData defined. (#409)
  • Add rf_agg_approx_quantiles function to compute cell quantiles across an entire column.

0.8.4

  • Upgraded to Spark 2.4.4
  • Add rf_mask_by_values and rf_local_is_in raster functions; added optional inverse argument to rf_mask functions. (#403, #384)
  • Added forced truncation of WKT types in Markdown/HTML rendering. (#408)
  • Add rf_local_is_in raster function. (#400)
  • Added partitioning to catalogs before processing in RasterSourceDataSource (#397)
  • Fixed bug where rf_tile_dimensions would cause unnecessary reading of tiles. (#394)
  • Breaking (potentially): removed GeoTiffCollectionRelation due to usage limitation and overlap with RasterSourceDataSource functionality.

0.8.3

  • Updated to GeoTrellis 2.3.3 and Proj4j 1.1.0.
  • Fixed issues with LazyLogger and shading assemblies (#293)
  • Updated rf_crs to accept string columns containing CRS specifications. (#366)
  • Added rf_spatial_index function. (#368)
  • Breaking (potentially): removed pyrasterframes.create_spark_session in lieu of pyrasterframes.utils.create_rf_spark_session

0.8.2

  • Added ability to pass config options to convenience PySpark session constructor. (#361)
  • Bumped Spark dependency to version 2.3.4. (#350)
  • Fixed handling of aggregate extent and image size on GeoTIFF writing. (#362)
  • Fixed issue with RasterSourceDataSource swallowing exceptions. (#267)
  • Fixed SparkML memory pressure issue caused by unnecessary reevaluation, overallocation, and primitive boxing. (#343)
  • Fixed Parquet serialization issue with RasterRefs (#338)
  • Fixed TileExploder, rf_agg_local_mean and TileColumnSupport to support proj_raster struct (#287, #163, #333).
  • Various documentation improvements.
  • Breaking (potentially): Synchronized parameter naming in Python and Scala for spark.read.raster (#329).

0.8.1

  • Added rf_local_no_data, rf_local_data and rf_interpret_cell_type_as raster functions.
  • Added: rf_rgb_composite and rf_render_png.
  • Added toMarkdown and toHTML extension methods for DataFrame, and registered them with the IPython formatter system when rf_ipython is imported.
  • New documentation theme (thanks @jonas!).
  • Fixed: Removed false return type guarantee in cases where an Expression accepts either Tile or ProjectedRasterTile (#295)

0.8.0

  • Super-duper new Python-centric RasterFrames Users’ Manual!
  • Upgraded to the following core dependencies: Spark 2.3.3, GeoTrellis 2.3.0, GeoMesa 2.2.1, JTS 1.16.0.
  • Build pyrasterframes binary distribution for pip installation.
  • Added support for rendering RasterFrame types in IPython/Jupyter.
  • Added new tile functions rf_round, rf_abs, rf_log, rf_log10, rf_log2, rf_log1p, rf_exp, rf_exp10, rf_exp2, rf_expm1, rf_resample.
  • Support Python-side Tile User-Defined Type backed by numpy ndarray or ma.MaskedArray.
  • Support Python-side Shapely geometry User-Defined Type.
  • SQL API support for rf_assemble_tile and rf_array_to_tile.
  • Introduced at the source level the concept of a RasterSource and RasterRef, enabling lazy/delayed read of sub-scene tiles.
  • Added withKryoSerialization extension methods on SparkSession.Builder and SparkConf.
  • Added rf_render_matrix debugging function.
  • Added RasterFrameLayer.withExtent extension method.
  • Added SinglebandGeoTiff.toDF extension method.
  • Added DataFrame.rasterJoin extension method for merging two dataframes with tiles in disparate CRSs.
  • Added rf_crs for ProjectedRasterTile columns.
  • Added st_extent (for Geometry types) and rf_extent (for ProjectedRasterTile and RasterSource columns).
  • Added st_geometry (for Extent types) and rf_geometry (for ProjectedRasterTile and RasterSource columns).
  • Reworked build scripts for RasterFrames Jupyter Notebook.
  • Breaking: The type RasterFrame renamed RasterFrameLayer to be reflect its intended purpose.
  • Breaking: All asRF methods renamed to asLayer.
  • Breaking: Root package changed from org.locationtech.rasterframes to org.locationtech.rasterframes.
  • Breaking: Removed envelope, in lieu of st_extent, rf_extent or st_envelope
  • Breaking: Renamed rf_extent_geometry to st_geometry
  • Breaking: Renamed rf_tile_dimensions to rf_dimensions
  • Breaking: Renamed rf_reproject_geometry to st_reproject
  • Breaking: With the upgrade to JTS 1.16.0, all imports of com.vividsolutions.jts need to be changed to org.locationtech.jts.
  • Deprecation: Tile column functions (in RasterFunctions) and SQL registered names have all been renamed to follow snake_case conventions, with an rf_ prefix, matching SQL and Python. A temporary compatibility shim is included so that code built against 0.7.1 and earlier still work. These will be marked as deprecated.
  • Breaking: In Scala and SQL, ..._scalar functions (e.g. local_add_scalar) have been removed. Non-scalar forms now dynamically detect type of right hand side.
  • Breaking: tileToArray has been replaced with _tile_to_array_double and _tile_to_array_int.
  • Breaking: Renamed bounds_geometry to rf_extent_geometry.
  • Breaking: renamed agg_histogram to rf_agg_approx_histogram, local_agg_stats to rf_agg_local_stats, local_agg_max to rf_agg_local_max, local_agg_min to rf_agg_local_min, local_agg_mean to rf_agg_local_mean, local_agg_data_cells to rf_agg_local_data_cells, local_agg_no_data_cells to rf_agg_local_no_data_cells.
  • Breaking: CellHistogram no longer carries along approximate statistics, due to confusing behavior. Use rf_agg_stats instead.
  • Introduced LocalCellStatistics class to wrap together results from LocalStatsAggregate.
  • Breaking: TileDimensions moved from astraea.spark.rasterframes to org.locationtech.rasterframes.model.
  • Breaking: Renamed RasterFrame.withBounds to RasterFrameLayer.withGeometry for consistency with DataSource schemas.

Known issues

  • #188: Error on deserialization of a Tile with a bool cell type to the Python side; see issue description for work around.

0.7.x

0.7.1

  • Fixed ColorRamp pipeline in MultibandRender
  • Fixed Python wrapper for explodeTiles

0.7.0

  • Now an incubating project under Eclipse Foundation LocationTech! GitHub repo moved to locationtech/rasterframes.
  • PySpark support! See pyrasterframes/python/README.rst to get started.
  • Exposed Spark JTS spatial operations in Python.
  • Added RasterFrames-enabled Jupyter Notebook Docker Container package. See deployment/README.md for details.
  • Updated to GeoMesa version 2.0.1.
  • Added convertCellType, normalizedDifference mask and inverseMask operations on tile columns.
  • Added tile column + scalar operations: localAddScalar, localSubtractScalar, localMultiplyScalar, localDivideScalar
  • Added rasterize and reprojectGeometry operations on geometry columns.
  • Added for for writing GeoTIFFs from RasterFrames via DataFrameWriter.
  • Added spark.read.geotrellis.withNumPartitions(Int) for setting the initial number of partitions to use when reading a layer.
  • Added spark.read.geotrellis.withTileSubdivisions(Int) for evenly subdividing tiles before they become rows in a RasterFrame.
  • Added experimental package for sandboxing new feature ideas.
  • Added experimental GeoJSON DataSource with schema inferfence on feature properties.
  • Added Scala, SQL, and Python tile-scalar arithmetic operations: localAddScalar, localSubtractScalar, localMultipyScalar, localDivideScalar.
  • Added Scala, SQL, and Python tile functions for logical comparisons both tile-tile and tile-scalar variants: localLess, localLessEqual, localGreater, localGreaterEqual, localEqual, and localUnequal.
  • Added SlippyExport experimental feature for exporting the contents of a RasterFrame as a SlippyMap tile image directory structure and Leaflet/OpenMaps-enabled HTML file.
  • Added experimental DataSource implementations for MODIS and Landsat 8 catalogs on AWS PDS.
  • Change: Default interpoation for toRaster and toMultibandRaster has been changed from Bilinear to NearestNeighbor.
  • Breaking: Renamed/moved astraea.spark.rasterframes.functions.CellStatsAggregateFunction.Statistics to astraea.spark.rasterframes.stats.CellStatistics.
  • Breaking: HistogramAggregateFunction now generates the new type astraea.spark.rasterframes.stats.CellHistogram.
  • Breaking: box2D renamed envelope.

0.6.x

0.6.1

  • Added support for reading striped GeoTiffs (#64).
  • Moved extension methods associated with querying tagged columns to DataFrameMethods for supporting temporal and spatial columns on non-RasterFrame DataFrames.
  • GeoTIFF and GeoTrellis DataSources automatically initialize RasterFrames.
  • Added RasterFrame.toMultibandRaster.
  • Added utility for rendering multiband tile as RGB composite PNG.
  • Added RasterFrame.withRFColumnRenamed to lessen boilerplate in maintaining RasterFrame type tag.

0.6.0

  • Upgraded to Spark 2.2.1. Added VersionShims to allow for Spark 2.1.x backwards compatibility.
  • Introduced separate rasterframes-datasource library for hosting sources from which to read RasterFrames.
  • Implemented basic (but sufficient) temporal and spatial filter predicate push-down feature for the GeoTrellis layer datasource.
  • Added Catalyst expressions specifically for spatial relations, allowing for some polymorphism over JTS types.
  • Added a GeoTrellis Catalog DataSource for inspecting available layers and associated metadata at a URI
  • Added GeoTrellis Layer DataSource for reading GeoTrellis layers from any SPI-registered GeoTrellis backend (which includes HDFS, S3, Accumulo, HBase, Cassandra, etc.).
  • Ability to save a RasterFrame as a GeoTrellis layer to any SPI-registered GeoTrellis backends. Multi-column RasterFrames are written as Multiband tiles.
  • Addd a GeoTiff DataSource for directly loading a (preferably Cloud Optimized) GeoTiff as a RasterFrame, each row containing tiles as they are internally organized.
  • Fleshed out support for MultibandTile and TileFeature support in datasource.
  • Added typeclass for specifying merge operations on TileFeature data payload.
  • Added withTemporalComponent convenince method for creating appending a temporal key column with constant value.
  • Breaking: Renamed withExtent to withBounds, and now returns a JTS Polygon.
  • Added EnvelopeEncoder for encoding JTS Envelope type.
  • Refactored build into separate core and docs, paving way for pyrasterframes polyglot module.
  • Added utility extension method withPrefixedColumnNames to DataFrame.

Known Issues

  • Writing multi-column RasterFrames to GeoTrellis layers requires all tiles to be of the same cell type.

0.5.x

0.5.12

  • Added withSpatialIndex to introduce a column assigning a z-curve index value based on the tile’s centroid in EPSG:4326.
  • Added column-appending convenience methods: withExtent, withCenter, withCenterLatLng
  • Documented example of creating a GeoTrellis layer from a RasterFrame.
  • Added Spark 2.2.0 forward-compatibility.
  • Upgraded to GeoTrellis 1.2.0-RC2.

0.5.11

  • Significant performance improvement in explodeTiles (1-2 orders of magnitude). See #38
  • Fixed bugs in NoData handling when converting to Double tiles.

0.5.10

  • Upgraded to shapeless 2.3.2
  • Fixed #36, #37

0.5.9

  • Ported to sbt 1.0.3
  • Added sbt-generated astraea.spark.rasterframes.RFBuildInfo
  • Fixed bug in computing aggMean when one or more tiles are null
  • Deprecated rfIinit in favor of SparkSession.withRasterFrames or SQLContext.withRasterFrames extension methods

0.5.8

  • Upgraded to GeoTrellis 1.2.0-RC1
  • Added REPLsent-based tour of RasterFrames
  • Moved Giter8 template to separate repository s22s/raster-frames.g8 due to sbt limitations
  • Updated Getting Started to reference new Giter8 repo
  • Changed SQL function name rf_stats and rf_histogram to rf_aggStats and rf_aggHistogram for consistency with DataFrames API

0.5.7

  • Created faster implementation of aggregate statistics.
  • Fixed bug in deserialization of TileUDTs originating from ConstantTiles
  • Fixed bug in serialization of NoDataFilter within SparkML pipeline
  • Refactoring of UDF organization
  • Various documentation tweaks and updates
  • Added Giter8 template

0.5.6

  • TileUDFs are encoded using directly into Catalyst–without Kryo–resulting in an insane decrease in serialization time for small tiles (int8, <= 128²), and pretty awesome speedup for all other cell types other than float32 (marginal slowing). While not measured, memory footprint is expected to have gone down.

0.5.5

  • aggStats and tileMean functions rewritten to compute simple statistics directly rather than using StreamingHistogram
  • tileHistogramDouble and tileStatsDouble were replaced by tileHistogram and tileStats
  • Added tileSum, tileMin and tileMax functions
  • Added aggMean, aggDataCells and aggNoDataCells aggregate functions.
  • Added localAggDataCells and localAggNoDataCells cell-local (tile generating) fuctions
  • Added tileToArray and arrayToTile
  • Overflow fix in LocalStatsAggregateFunction
0.9.1