2.0.0b1

changelog

#32 #35 generator basics from @castelao
#33 #34 CI migration to ghactions and and pytests
#36 better use of netcdf metadata
#38 #39 #40 bugfixes around typecasting and wod and iquod flag parsing from @BecCowley

2.0.0b0

Version 2.0.0 of wodpy includes the long-awaited netCDF parser, intended to consume netCDF files and make a close-as-possible analogous profile class to the ASCII representation.

This release is marked b(eta)0 / pre-release as there's some improvements to be made before full release:

• numerous functions in ncProfile class just raise notImplemented
• need to make a clear statement about exactly what netCDF specification we're trying to parse here; priority will be WOD files with IQuOD information added, but need a clearer documentation reference / example profiles to build and test against. Currently built against the wodselect-returned netCDF file with the closest match to the ASCII file and profile the ASCII parser was built against.

1.6.2

• Tracks updates to CoTeDe from #25 (thanks as always @castelao!)
• A little metadata fiddling to play nice with twine and PyPI

1.6.1

This patch release fixes a regression we picked up in python 3, discussed in #22

1.6.0

• Adds support for IQuOD v0.1 ASCII files
• Adds level variable (temperature, salinity...) metadata dictionaries, including intelligent metadata keys

1.5.0

• Adds support for originator code extraction (secondary header ID 96)
• Exposes generic function to extract any secondary header ID

bugfixes

Fix bug in PI header pointed out by @CiaranEvans

Python 3.5 validation, new unpacks

• bugfixes for windows line breaks
• python 3.5 validation in CI (along with 2.7)
• new unpacks referenced in #14

Datetime handling, other checks

Adds smarter handling of dates and times, and other robustness checks.

CoTeDe support, bugfixes

Merge pull request #7 from IQuOD/end-of-file-error

Error when at end of file