Hi guys,
I've added a new means of plotting, currently only in trunk/ and
currently undocumented, that will pop up windows in the 'pylab'
environment with the displayed plots in them.
Currently, this works best within IPython, run in pylab mode ('ipython -pylab').
I've installed IPython on datastar, along with the TkAgg backend for
pylab. If you've never played with pylab, then you shouldn't have any
problems -- but if you have, you may need to modify your matplotlibrc
file to get it to use the TkAgg backend. Additionally, for some
reason IPython crashes on tab completion for me. I'll let you know if
I find a solution to that. (IPython has a ton of magic in it; I am by
no means an expert, and I don't even usually use it very often, but
it's pretty cool.)
So if you want a PlotCollection object that will pop up windows, you
need to ask for PlotCollectionInteractive. I've included below an
example script for how to do basic stuff. These commands should be
typed in or %run (see below) at the prompt given by running ipython
-pylab . All of these require that you have my python libraries in
your setup (http://yt.enzotools.org/doc/getting_started.html#datastar-sdsc)
but you additionally need to have my bin path in your PATH variable
(/users/stanford/mturk/local/bin/). The documentation will be updated
accordingly, but it's still in testing and I wanted to give you a
heads up.
>>>
from yt.mods import *
pf = EnzoStaticOutput("/users/stanford/mturk/local/src/yt/trunk/tests/DD0010/moving7_0010")
pc = PlotCollectionInteractive(pf, center=[0.75,0.75,0.75])
pc.add_phase_sphere(0.1,'1', ["Density","Temperature","x-velocity"])
pc.add_slice("Density", 0)
<<<
Note that there's *nothing* special here -- except that instead of
making a PlotCollection, we make a PlotCollectionInteractive.
IPython has a "%run" command that will execute a script that exists in
the file system. So if you have these commands (or any commands,
really) in a file, you can type "%run my_file.py" at the IPython
prompt and it will execute them. For more information about IPython,
which I've been told will "change your life!" you can see:
http://ipython.scipy.org/
or the videos at:
http://showmedo.com/videos/series?name=CnluURUTV
Additionally, Jeff (our man in Orion-ville) is an expert at IPython,
and I be he can answer any questions you might have.
Please let me know what you think, and if you have any suggestions.
Also, if any of these instructions don't work, please reply to the
list!
-Matt
Hi there,
I've created a couple binaries installations of yt along with an
installation script. If anybody has time to try it out and get back
to me, that'd be pretty awesome. Right now they're experimental, but
I've tested them a couple places and found them to be functional.
The installation script is here:
http://yt.enzotools.org/browser/trunk/doc/install_script.sh?format=raw
You should be able to download it, modify the parametet MY_DIR at the
top, and then execute it. It will download the appropriate software,
assuming a base level of installed system-level packages, install it
and then echo to you the locations of various items. Once it's
finished and your path is set according to the output from the script,
you should be able to run 'reason' or 'yt' to get the GUI or the
command line, respectively.
I've also packaged up binaries for ia64 and x86_64.
http://yt.enzotools.org/files/yt-0.3-ia64.ziphttp://yt.enzotools.org/files/yt-0.3-x86_64.zip
If you're on either of these types of machines, you should be able to
download the appropriate file and unzip it. It will create a
directory named yt-0.3 containing three executables: 'yt', 'reason'
and 'py'. Executing yt-0.3/yt will give you a command prompt with
appropriately pre-loaded modules, executing yt-0.3/py will give you a
vanilla command prompt, and executing yt-0.3/reason will give you the
GUI.
Again, I've tested these on a couple systems, but I'd *really*
appreciate if you could download and give them a shot. Please let me
know if you have success or failure!
-Matt
Hi guys,
I discovered a bug in the cutting plane for yt. The buffer, received
from my C-code for pixelization, was not transposed, as it should have
been. I have fixed it, and what this ends up meaning is that the
resultant image is now flipped about a line going from lower-left to
upper-right. This is not a huge deal; what is a bigger deal is that
the cutting-plane vectors were also not transposed, and because that
created a two-component plot, they were previously not correct. Now,
they are correct, as of r655.
I've placed example before/after pictures here:
http://yt.enzotools.org/files/cp_before.pnghttp://yt.enzotools.org/files/cp_after.png
I hope this has not been an inconvenience for anyone.
-Matt
