                 Changes in version 0.7.4 (2026-05-29)                  

  - fix GCC16 issues on CRAN
  - fix compatibility with GDAL 3.13

                 Changes in version 0.7.3 (2026-01-18)                  

  - remove CXX11 dependency in configure
  - fix compilation error on CRAN r-devel-linux by updating datetime.cpp
    to fix date::parse() template matching

                 Changes in version 0.7.2 (2025-12-02)                  

  - fix CRAN issues due to missing error handling in
    add_collection_format()
  - add support for Sentinel data from EOPF
    (https://zarr.eopf.copernicus.eu) in stac_image_collection()

                 Changes in version 0.7.1 (2025-03-20)                  

  - fix CRAN build issues related to clang-20 and gcc_UBSAN

                 Changes in version 0.7.0 (2024-03-06)                  

  - add as.data.frame() to easily convert data cubes to data frames
  - add predict.cube() to predict pixel values based on models (lm, glm,
    caret, tidymodels, and similar)
  - add window_space() to apply (focal) moving window kernels or
    aggregation functions
  - extract() now combines extracted values with input geometries and
    attributes (if merge = TRUE)
  - add support for imagery with spatial reference from geolocation
    arrays (including curvilinear grids)
  - stac_image_collection() now accepts STACItemCollection objects
    directly and should be more robust
  - Windows build uses pkg-config if available
  - Improved error reporting for inaccessible imagery

                 Changes in version 0.6.4 (2023-04-14)                  

  - add native quartile reducers in reduce_time()
  - fix r-devel UCRT win build on CRAN
  - fix crashes on Windows UCRT due to unusable std::regex()
  - fix parallel data cube processing when nonstandard external package
    locations are used
  - stack_cube() now ignores files if not accessible / invalid instead
    of stopping all computations
  - The codebase has been reorganized R package is now maintained under
    https://github.com/appelmar/gdalcubes, whereas the C++ repo will be
    archived.

                 Changes in version 0.6.3 (2023-01-19)                  

  - fix gcc-13 compiler errors on CRAN
  - add datetime interval support in STAC collections
  - add support of new windows toolchain using Makevars.ucrt

                 Changes in version 0.6.2 (2022-10-10)                  

  - fix clang-15 compiler warnings on CRAN
  - new operation aggregate_space() to reduce spatial resolution of data
    cubes
  - improved / faster implementation of plot()
  - handle WKT strings as spatial reference systems in STAC responses
  - handle special characters in variable / band names

                 Changes in version 0.6.1 (2022-03-23)                  

  - fix gcc-12 builds on CRAN
  - fix automatic reprojection in extract_geom()
  - update GDAL on Windows

                 Changes in version 0.6.0 (2022-03-08)                  

  - major stability improvements:
      - fix unexpected stack overflows due to to GDAL error handler from
        sf calling Rf_warning()
      - if GDALOpen() fails to read an image, it will now be simply
        ignored but not stop processing the current chunk
      - improved handling and checks for empty chunks in data cube
        operations
      - parallel processing now uses worker processes instead of threads
  - new extract_geom() function to extract data cube values from spatial
    or spatiotemporal features and to compute summary statistics
  - remove functions query_points(), query_timeseries(), and
    zonal_statistics() in favor of extract_geom()
  - fix filter_geom() issues with larger polygons
  - fix filter_geom() error while checking if polygon is within data
    cube
  - use WKT strings or authority codes in image collections instead of
    proj4 strings
  - chunk sizes can now be set as a global package option either as
    constant sizes or as a function of data cube size
  - default chunk sizes consider the number of parallel worker processes
  - animate() now can produce mp4 and GIF animations
  - animate() works for larger image sequences using the av or gifski
    packages
  - remove dependency on RcppProgress

                 Changes in version 0.5.1 (2021-12-03)                  

  - fix CRAN vignette issue on Mac due to data download failures
  - fix image_mask() function for minimum and maximum values

                 Changes in version 0.5.0 (2021-10-28)                  

  - new operation aggregate_time() to reduce temporal resolution of data
    cubes
  - new stack_cube() function to build data cubes from aligned images
    without image collection creation
  - new operations slice_time() and slice_space() to extract single time
    series or slices
  - new crop() function can be used to crop a data cube by space and/or
    time
  - single bands of data cubes can be selected using the $ operator
  - fix datetime parser to support strings with fractional seconds
  - fix CRAN issues due to obsolete autoconf warnings
  - the[] operator can now be used for flexible cropping, slicing, and
    band selection on data cubes

                 Changes in version 0.4.1 (2021-07-29)                  

  - fix build issues on MacOS
  - fixes for Windows build including ucrt support

                 Changes in version 0.4.0 (2021-07-13)                  

  - new operator ncdf_cube() to read data cubes from (intermediate)
    results
  - new operator rename_bands() to change band names
  - image collection creation from STAC API queries with
    stac_image_collection()
  - progress bar can now be disabled with gdalcubes_options()
  - removed gdalcubes_set_threads() in favor of gdalcubes_options()
  - removed gdalcubes_debug_output() in favor of gdalcubes_options()
  - removed gdalcubes_set_ncdf_compression() in favor of
    gdalcubes_options()
  - removed gdalcubes_use_cache() in favor of gdalcubes_options()
  - removed image collection operations translate_COG() and
    translate_gtiff()
  - fix installation issues on MacOS and GCC11 warnings

                 Changes in version 0.3.1 (2020-08-27)                  

  - make GEOS dependency optional
  - remove CURL dependency from configure

                 Changes in version 0.3.0 (2020-08-05)                  

New Features

  - Compute summary statistics of data cubes over polygons with
    zonal_statistics()
  - Extracts time series at irregular spatial points with
    query_timeseries()
  - Time dimension may ow be irregular / labeled after selecting slices
    with the new select_time() function
  - Filter pixels of a data cube by a spatial polygon with filter_geom()
  - Apply an R function on time series without reduction using
    apply_time()
  - Batch format conversion of images in a collection with
    translate_cog() and translate_gtiff()

Minor improvements

  - conversion to stars objects with st_as_stars()
  - add support for image collections without collection format in
    create_image_collection()
  - optional global SRS definition in collection formats
  - default chunk size is now (t,y,x) = (1,256,256)
  - remove reduce() function
  - remove cube argument in cube_view function
  - new collection format for daily 0.25° AVHRR Optimum Interpolation
    Sea Surface Temperature
  - new collection formats for ESA CCI soil moisture products
  - new collection format for daily precipitation observations from GPM
    / IMERG
  - new collection format for MODIS MOD09GA (aqua and terra)
  - add na.color argument in plot.cube()

Bug fixes

  - fix CRS metadata in produced netCDF files
  - fix multithreading locking issues with GDAL 3

                 Changes in version 0.2.5 (2020-05-17)                  

  - fixed compiler warnings on CRAN
  - temporarily removed as_stars(), will be added again in 0.3

                 Changes in version 0.2.4 (2020-02-02)                  

  - fixed axis order issues with GDAL3 and PROJ6
  - fixed compiler warnings with GDAL3

                 Changes in version 0.2.3 (2019-10-21)                  

  - fixed clang compiler warnings
  - fixed MODIS collection formats
  - new collection formats MxD14A2 and MxD13A2

                 Changes in version 0.2.2 (2019-10-15)                  

  - support for GDAL subdatasets in collection format
  - MODIS collection formats now use subdatasets automatically
  - fixed configure.ac for R-devel
  - add query_points() to query data cube values at irregular
    spatiotemporal points

                 Changes in version 0.2.1 (2019-08-21)                  

  - new collection format for PlanetScope data
  - fixed R CMD check warnings on CRAN (caused by compiler warning
    -Wdeprecated-declarations)
  - fixed mean aggregation

                 Changes in version 0.2.0 (2019-08-08)                  

New Features

  - add animate() function to create data cube time series animations
  - apply mask bands on pixel values during the construction of the data
    cube, see ?image_mask
  - add write_tif() to export data cubes as (possibly cloud-optimized)
    GeoTIFF files (one per time slice)
  - export of data cubes with write_tif() and write_ncdf() supports
    packing data values to smaller integer types
  - processing cubes is interruptible, though it can still take time to
    let all threads finish their current chunk
  - add as_array() function to convert a data cube to a native in-memory
    R array
  - new operator fill_time() fills NA pixels of data cubes based on time
    series interpolation
  - changed image collection database schema, existing collections must
    be recreated
  - new global configuration function gdalcubes_options() as a
    replacement to gdalcubes_set_threads() etc.
  - new function add_images() adds images to an existing image
    collection

Minor improvements

  - rename filter_predicate() -> filter_pixel()
  - collection format Sentinel2_L2A now includes WVP, AOT, and SCL bands
  - consistent output for printing data cube views and data cubes
  - new collection format for Sentinel-2 data on Theia (credits to
    Xavier Laviron)
  - new collection format for MODIS MxD13Q1 vegetation index data
  - add write_json_descrargument to write_ncdf()
  - new argument with_VRT in write_ncdf() to write GDAL VRT datasets for
    data cube time slices
  - collection formats can now overwrite scale, offset, and unit for
    bands
  - write_ncdf() can produce netCDF files without bounds variables if
    desired
  - write_ncdf() and write_tif() return created files as character
    vectors.

Bug fixes

  - fix windows source compilation on CRAN
  - bands of multiband files are now read in correct order
  - fix package build with PROJ 6.1 (credits to Roger Bivand)

                 Changes in version 0.1.0 (2019-05-24)                  

  - First release
