- Update readme.
- Remove unnecessary dimensions while reading point data.
- Pin squinx version to fix rtd
- Update netcdf4 requirement to v1.4.2
- Update installation to pyscaffold 2.5.x to fix pyscaffold/pyscaffold#148
- Restrict netcdf4 package to versions <=1.2.8 because of Unidata/netcdf4-python#784
- Allow writing and reading of PointData in append mode.
- Set default filenames for GriddedPointData to include .nc ending.
- Translate RuntimeError of older versions of the netCDF4 library to IOError.
- Avoid race conditions when creating directories for new files.
- Fix bug that lost the datatype during writing of timeseries with pandas > 0.17.1
- Fix Python 3 compability.
- fix bug that read the wrong timeseries if a non existing location id was given.
- fix bug when reading not completely filled files in read_bulk mode in contigous ragged.
- Catch RuntimeError and IOError to be compatible with older netCDF4 versions.
- Fix compression of variables in point_data.
- IndexedRaggedTs are now compatible with numpy record arrays and dictionaries.
- IndexedRaggedTs.write_ts can now write data for multiple grid points at once.
- Add interface to write a complete cell to GriddedNcIndexedRaggedTs
- No changes in functionality
- Fix setup.py for correct installation
- Fix n_loc bug
- Add recarray for point data
- Excluding pandas==0.19.0
- Deprecate pynetcf.time_series.GriddedTs please use pynetcf.time_series.GriddedNcTs in the future. Be aware that the init arguments have changed slightly to path, grid, ioclass.
- Add support of read/write netCDF point data following CF conventions
- Add support for disabling automatic masking during reading. Useful if the data has fill values but needs to be scaled to a datatype that does not support NaN values.
- Add support for disabling automatic scaling in base netCDF4 library.
- Add support for dtype conversion before scaling and offset.
- Add classes for gridded datasets based on pygeobase
- improve test coverage
- make compatible with newest netCDF4 releases
- support read_bulk keyword for all dataset types
- fix open/closing of netCDF file
- fixed issue #9
- fixed issue #10
- fixed issue #4
- moved netcdf classes out of rs data readers