Re: mpFillColors and nglSpreadColors and LINUX

From: Mary Haley <haley_at_nyahnyahspammersnyahnyah>
Date: Thu, 21 Oct 2004 17:28:33 -0600 (MDT)

>
> Hi Mary,
>
> Thanks for the response, but I'm afraid I'm no further along. I understand
> your comments and I was able to set the land fill color directly using a
> named colormap, and I was able to get the length of the assigned colormap as
> you suggested. But, no matter what I do, nglSpreadColors does nothing. I
> still have a gray bar at the top of the BlueRedGray colorbar. I can use a
> different colormap but it doesn't change the underlying problem that
> nglSpreadColors doesn't seem to work.
>
> I am supposed to pass the resources which have nglSpreadColors=True to the
> contour function right?

Derrick,

Yes, that's correct, and I think I see a big OOPS on my part here.

The "nglSpreadColorEnd" resource has no "s" after the word "Color".

I didn't see this before in your email, I apologize!

I'm looking into why PyNGL didn't complain about your misspelled
resource. I think I know what the problem is...

> map = ngl.contour_map(wks,annual,resources) # Draw contours over a map.
>
> Is there another possibility? I've tried a few simpler tests to and as far as
> I can tell, nglSpreadColors, nglSpreadColorsEnd/Start don't do anything.
>
> Could this be another gcc3.3.3/Red Hat FC2 problem?
>
> Thanks again,
> Derrick
>
> PS: At the risk of starting a firestorm, can anyone comment on other linux
> distributions as far as reliability and smooth operations? Also, I'm
> interested to hear from any Mac OS/X users regarding ncl/PyNGL performance on
> a Mac.

I do know there are problems that somebody had with a Mandrake
installation, and I'm baffled by that one.

My experience with building NCL for different systems seems to
indicate that it's a real pain to keep all these Linux binaries
straight. Now that 64-bit is entering the field, this adds a whole
level of complexity. On top of that, we have compilers galore (Intel,
Portland Group, PathScale, GNU) and it's fun trying to get them all
to play well together.

>From the PyNGL perspective, I'm hoping the Python distutils stuff
handles all this. I haven't had a chance to try it on these systems
yet. I would like to hear from others on this issue as well.

--Mary

> On Thu October 21 2004 11:02 am, Mary Haley wrote:
> > > I'm having difficulty deciphering how the colors in a colormap are
> > > assigned to various elements in the plot. I'm following the chkbay.py
> > > example from pynglex somewhat.
> > >
> > > I'm trying to figure out the colors which are assigned to the background
> > > (I assume element 0), foreground(1), land, ocean, etc.
> > >
> > > I am able to set the foreground and background correctly.
> > >
> > > resources = ngl.Resources()
> > > rlist = ngl.Resources()
> > > rlist.wkColorMap = 'BlueRedGray'
> > > ngl.set_values(wks,rlist)
> > > resources.mpFillOn = True # Turn on map fill.
> > > resources.mpLandFillColor = 254
> > > resources.mpOceanFillColor = -1
> > >
> > > This gives me a white background with black letters and a gray fill color
> > > for land, which is what I want. But, I also get a gray box in the
> > > "warmest" value for my colorbar. I have tried the nglSpreadColors
> > > resource like so ...
> > >
> > > resources.nglSpreadColors = True
> > > resources.nglSpreadColorsEnd=-2
> > >
> > > which I then pass to the map function ...
> > > map = ngl.contour_map(wks,sstma,resources) # Draw contours over a map.
> > >
> > > This has no effect that I can see. I thought it would tell the contour
> > > function to stop using the colormap at the ncol - 2 entry. Is this the
> > > wrong interpretation or am I doing something else?
> >
> > Hi Derrick,
> >
> > It looks like you are doing everything correctly. Are you using
> > the same colormap, "rainbow+gray"? If not, then you should check
> > your colormap to make sure that you don't have more than one gray
> > entry at the end of your colormap.
> >
> > > On that note, is there a way, short of defining your own colormap as a
> > > Numeric array, that you can find the length of a colormap in order to
> > > know beforehand the indices to assign to mpOceanFillColor?
> >
> > You can get the length as follows: any time after the "open_wks" call,
> > insert:
> >
> > cmap_len = Ngl.get_integer(wks,"wkColorMapLen")
> >
> > You can also draw the colormap, for debugging purposes with the
> > following call:
> >
> > Ngl.draw_colormap(wks)
> >
> > With regard to your question, however, you don't need to use index
> > values to specify what color you want for mpOceanFillColor. Instead,
> > you can use one of 650 "named" colors, if the named color is
> > already in your colormap. For example:
> >
> > resources.mpOceanFillColor = "DeepSkyBlue"
> >
> > To see all the named colors, visit:
> >
> > http://www.pyngl.ucar.edu/selectedtopics.html#CreatingColorMapNamed
> >
> > You can click on any of the graphics for a larger view, and also
> > to be able to read the name of the color. The color names
> > are case insensitive, so "PapayaWhip" is the same as "papAYAWHiP".
> >
> > Each named color has an associated RGB value with it. If this RGB
> > value is not found in your colormap, PyNGL will find the closest color
> > to using a special algorithm. (I can't think of the name of it at the
> > moment.)
> >
> > --Mary
> >
> > > Any help is appreciated,
> > > Derrick
> > >
> > > --
> > > Derrick Snowden
> > > PhOD/AOML/NOAA
> > > 4301 Rickenbacker Causeway
> > > Miami, FL 33149
> > > Office: 305.361.4322 Fax: 305.361.4392
> > > _______________________________________________
> > > pyngl-talk mailing list
> > > pyngl-talk_at_ucar.edu
> > > http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
>
> --
> Derrick Snowden
> PhOD/AOML/NOAA
> 4301 Rickenbacker Causeway
> Miami, FL 33149
> Office: 305.361.4322 Fax: 305.361.4392
> _______________________________________________
> pyngl-talk mailing list
> pyngl-talk_at_ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
>

-- 
-------------------------------------------------
Mary Haley                   haley_at_ucar.edu
NCAR/SCD/VETS                303-497-1254 (voice)
1850 Table Mesa Dr           303-497-1804 (fax)
Boulder, CO  80305
-------------------------------------------------
_______________________________________________
pyngl-talk mailing list
pyngl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
Received on Thu Oct 21 2004 - 17:28:58 MDT

This archive was generated by hypermail 2.2.0 : Thu Jan 19 2006 - 21:30:16 MST