hspatial 4.1.0

extract_point_from_raster() now also accepts a GDALRaster object.

Updated requirements to be compatible with Django 4 and 5.

Now works with htimeseries 6 and 7.

PointTimeseries() has changed in that default_time must be aware.

Now requires htimeseries 6. This recent htimeseries version works with aware time zones, and hspatial, which works with naive, internally handles them as UTC.

Reverted previous change, which had not been properly tested and doesn’t work; in effect 2.1.6 is the same as 2.1.4.

Added compatibility with htimeseries 6.

Added compatibility with gdal 3.

Updated dependencies (essentially htimeseries).

Updated requirements to be compatible with Django 3.

Fixed an error in extract_point_from_raster(). It was sometimes picking up a neigbouring pixel instead of the correct one.

h_integrate() now adds a “UNIT” metadata item to the produced raster with the unit of measurement as specified in the input files. Conversely, PointTimeseries reads the “UNIT” metadata item from the rasters and sets it in the returned HTimeseries object.

Now uses version 5 of hts file (i.e. different time step notation).

Added option default_time to PointTimeseries.

Fixed a bug where extract_point_from_raster() was modifying the point passed to it.

Fixed a bug where hts files were opened in the wrong mode, with inconsistent results (this is another occurence of the bug that had been fixed in 0.1.2).

Added test utility setup_test_raster() to make unit testing easier.

Added option “version” to PointTimeseries.get_cached() so that it can save in different file format versions.

Fixed a bug in extract_point_from_timeseries() where it was sometimes raising the wrong exception type (depending on GDAL version).

The API for extract_point_timeseries_from_rasters() has changed. Instead of a function, there’s now a PointTimeseries class.

When extracting a point timeseries from rasters, start_date and end_date can now be specified.

When extracting point time series from a set of rasters, it is now possible to provide a prefix as well as a list of files, and it is also possible to save the extracted time series to a file, optionally only if the file is outdated.

extract_point_from_raster (and extract_point_timeseries_from_rasters) now supports GeoDjango (GEOS) Point objects besides GDAL Point objects.

Function coordinates2point has been added.

Fixed a bug where hts files were opened in the wrong mode, with inconsistent results.

Fixed an ugly timezone bug that caused the data to refer to a different time than what the timestamp actually said.

When the timezone was missing from the input files, there was an incomprehensible AttributeError. This was fixed and now it provides an understandable error message.

Initial release