Re: New version of PyNGL and PyNIO available

From: Jeff Whitaker <jswhit_at_nyahnyahspammersnyahnyah>
Date: Tue, 19 Aug 2008 09:49:06 -0600

Mary Haley wrote:
> On Mon, 18 Aug 2008, Jeff Whitaker wrote:
>> Mary Haley wrote:
>>> We are please to announce a new release of PyNIO and PyNGL, which are
>>> now being released as separate packages, and are available as open
>>> source.
>>> The new version of both packages is 1.3.0b1, and they have the
>>> following new major features:
>>> - Support for masked arrays added
>>> PyNIO will now return a numpy masked array if a missing value is
>>> detected on the NetCDF, HDF, or GRIB file. The PyNGL visualization
>>> routines were also updated to recognize masked arrays. You no
>>> longer need to set the *MissingValue* resources if you have a
>>> masked array that contains a valid missing value.
>>> - Extended subscripting added to PyNIO
>>> PyNIO now supports an extended form of subscripting to specify
>>> subsets using a description string that can reference associated
>>> coordinate variables if they exist in the file. Dimensions can be
>>> transposed by varying the order of the coordinate axes in the
>>> subset specification. Time dimension subsets may be specified using
>>> ISO-8061-conforming date/time strings if the time dimension has a
>>> CF-conforming 'units' attribute.
>>> - New analysis functions in PyNGL:
>>> o Ngl.betainc - Evaluates the incomplete beta function.
>>> o Ngl.chiinv - Evaluates the inverse chi-squared distribution
>>> function.
>>> o Ngl.linmsg - Linearly interpolates to fill in missing values.
>>> o Ngl.regline - Calculates the linear regression coefficient
>>> between two series.
>>> - New visualization routines in PyNGL:
>>> o Ngl.streamline_scalar_map - routine for generating streamlines
>>> colored by scalar fields over a map.
>>> o Ngl.wmstnm - draws station model data.
>>> For a full list and description of what's new, see:
>>> We are releasing this as a beta version so we can get some user
>>> testing of the new features, especially recognition of masked arrays
>>> and the extended subscripting capability in PyNIO.
>>> Please report to pyngl-talk or Dave or I if you run into problems, and
>>> we'll try to make sure it gets into the official 1.3.0 release.
>>> Note: building PyNGL and/or PyNIO from source code is not a trivial
>>> task. We have tried to document this as best as we could, but we
>>> would also appreciate your feedback in this area. We have provided
>>> some prebuilt binaries. Send email to me if you run into any
>>> installation problems.
>>> Cheers,
>>> --Mary
>> Mary: Thanks for releasing the source! I've just tried to build NIO
>> on my mac, and almost succeeded. I don't have HDFEOS, so I set
>> HAS_HDFEOS to 0. does
>> gcc -bundle -undefined dynamic_lookup -L/sw/lib -I/sw/include
>> build/temp.macosx-10.5-ppc-2.5/niomodule.o -L/sw/lib
>> -L/Users/jwhitaker/lib -lnio -lmfhdf -ldf -ljpeg -lpng -lz -lnetcdf
>> -lhdf5_hl -lhdf5 -lsz -lgrib2c -ljasper -lpng -o ./
>> but importing nio gives
>> ImportError: dlopen(./, 2): Symbol not found: _GDattach
>> Referenced from: /Volumes/User/jwhitaker/python/PyNIO-1.3.0b1/
>> Expected in: dynamic lookup
> Jeff,
> Are you linking against a libnio.a that we built? If so, it may have
> been built with HDFEOS included, and hence you will need the HDFEOS
> library to resolve these symbols. I realize this is a pain. I should
> update the documentation to reflect this. This is true for NetCDF4
> and GRIB2 as well, unfortunately.
> The one option is to build NCL from source and answer "no" to the
> HDFEOS support question.
> It looks like you are building on a Mac PPC running 10.5? I've
> attached the HDF-EOS software for 10.4 (I don't have the PPC/10.5
> combo). It's attached to this message.
> --Mary

Mary: Oops, I missed that I have to build NCL first before PyNIO. Is
libnio too entangled with the rest of NCL to split it out and ship it
with PyNIO? It would be really nice to be able to just get the I/O
capabilities of PyNIO without having to build all the ncar graphics stuff.

>> I think that symbol is from hdfeos - so it would appear that it's not
>> really optional. Funny thing is, I grepped the PyNIO source and I
>> don't see that symbol referenced anywhere. -Jeff
>> --
>> Jeffrey S. Whitaker Phone : (303)497-6313
>> Meteorologist FAX : (303)497-6449
>> NOAA/OAR/PSD R/PSD1 Email :
>> 325 Broadway Office : Skaggs Research Cntr 1D-113
>> Boulder, CO, USA 80303-3328 Web :

Jeffrey S. Whitaker         Phone  : (303)497-6313
Meteorologist               FAX    : (303)497-6449
NOAA/OAR/PSD  R/PSD1        Email  :
325 Broadway                Office : Skaggs Research Cntr 1D-113
Boulder, CO, USA 80303-3328 Web    :
Received on Tue Aug 19 2008 - 09:49:06 MDT

This archive was generated by hypermail 2.2.0 : Thu Aug 21 2008 - 22:13:03 MDT