v1.1.1

Changes:

• Fix grid type encoding. Grids were only encoded as analyses and now the user input for the type will be respected and properly read by GEMPAK.
• Updated and cleaned up gird projections.
• Update example documentation for grid encoding.
• Add testing for improved grid encoding/decoding.

v1.1.0

Changes:

• Add the ability to read GEMPAK VGF files.

v1.0.2

Changes:

• Add ability to read climate surface files.
• Fixes surface file type detection, Ship type could be assigned erroneously in certain cases.

v1.0.0

Changes:

• Truly putting the I/O in gempakIO. GEMPAK files can now be written and used by GEMPAK programs.

0.8.3

Changes:

• Minor bug fix release. Fixed issue that caused error when reading GEMPAK grids with two date/time fields defined.

0.8.2

Changes:

• nearest_time and sfjson now have an include_special parameter that will toggle whether SPECI observations are extracted from the GEMPAK surface file.
• Standard METAR observations that have text data no longer will have their timestamp changed to the actual observation time. The time will remain the same as what is stored in the GEMPAK surface file. This should be more intuitive as users expect hourly METAR observation to be associated with the top of the hour even when they were taken slightly before that hour.

0.8.1

Changes:

• Rearranged package structure. This should be transparent to users.
• Improved performance of gdxarray, snxarray, and sfjson methods.
• Fixed bug where sounding data could only be in pressure coordinates. Height coordinates should now also be processed correctly.
• RADAT attribute in sounding data is now wmo_codes as that is a more appropriate description.
• Added LAST and FIRST keywords for date/time filtering. This follows the GEMPAK standard.
• Surface data has improved processing and times are now taken from the actual text observation wherever possible. This should mean results from nearest_time will be as expected.

0.6.2

Bug fixes for sounding and surface data text.

0.6.0

Fixed sounding decorder bug

Part of the SIGW merging code erroneously added pressure to the wind
direction array. This was subsequently written over by actual wind
direction data which may explain why no errors were found in testing.
The pressure level is not placed in the correct array.