[yt-users] SlicePlot y-axis problem

Matthew Turk matthewturk at gmail.com
Thu Apr 18 04:17:19 PDT 2013


Hi Caroline,

On Thu, Apr 18, 2013 at 6:29 AM, Caroline Van Borm <borm at astro.rug.nl> wrote:
> Hi Matthew,
>
> I tried both running it from a script and interactively (with iyt). I really
> only used a minimum of code:
>
> from yt.mods import *
> pf = EnzoStaticOutput("RD0007/RedshiftOutput0007")
> #I also tried: pf = load("RD0007/RedshiftOutput0007")
> SlicePlot(pf, 'y', 'Density').save()
>
> Sample data to run this on can be found on
>
> http://lca.ucsd.edu/software/enzo/data/cookbook/
>
> Just download and unpack any of the RD000* folders, for instance RD0007 for
> the example code above.
>
> On the screen, after the SlicePlot command it says
>
> yt : [INFO     ] 2013-04-18 12:16:59,830 Getting the binary hierarchy
> yt : [INFO     ] 2013-04-18 12:16:59,838 Finished with binary hierarchy
> reading
>
> and there it hangs. Pressing Ctrl-C doesn't do anything, I have to kill the
> process manually from another terminal.

Interesting.  I've never seen this behavior, and particularly not in a
way that biases one axis over another for slicing.  One thing we have
put into yt is the ability to respond to signals like SIGUSR1 ; so if
instead of "kill [pid]" you do "kill -SIGUSR1 [pid]" (where [pid] is
the process id) it should -- the next time the event loop hits --
output where in the python code it is.  But, if it's really hung, this
probably won't do anything.

Not responding to Ctrl-C indicates it's probably inside C code
somewhere; a likely candidate for this is in the IO.  Another thing to
try is this exact same script, but after you put these lines into
~/.yt/config

[yt]
loglevel: 1

this will increase the level of logging, which will tell us a lot
more.  Other scripts that use different patterns for data access may
also be illuminating:

pf = load( ... )
dd = pf.h.all_data()
print dd["Density"].min(), dd["Density"].max()

for instance.

>
> I installed yt using the installation script from the website, I don't think
> I changed any of the settings, or anything after the installation. The only
> thing I did before installing was "export CC=icc" to get HDF5 to install
> properly.

I'm hesitant to say this is a problem, but it may be.  This would mean
that the entire stack has been built with icc, but possibly only for
the C code, not the C++ code.  I've not seen a case in the past where
HDF5 wouldn't compile with GNU (i.e., gcc) and there may be a point
where matplotlib (C++ code internally) may not be interoperating well.
 My gut feeling is that rebuilding from scratch with the default
environment variables is the best bet, just for the purposes of
reducing the number of possible issues, but I'm not sure.

-Matt

>
> Thank you very much!
>
> Cheers,
> Caroline
>
>
>
> On 04/17/2013 01:42 PM, Matthew Turk wrote:
>
> Hi Caroline,
>
> Thanks for writing, and sorry that it's not working.  Don't worry too
> much about sending the data.  I'm sorry to hear that it's hanging and
> not responding, which isn't behavior we have often seen.  Here are a
> couple questions that will help us figure out how things might have
> gone wrong.
>
>  * Did you run this via a script that you wrote?  If you, could you
> run the command:
>
> yt pastebin that_script.py
>
> where "that_script.py" is the filename?  That will give a URL, which
> you can include in a reply so we can see it.
>
>  * What output do you see on screen up to the point where the execution
> hangs?
>  * How did you install yt?
>  * Is there one particular piece of sample data that we can run the
> script you paste'd (above) on to see the behavior?
>  * Have you made any changes to your installation -- for instance, to
> your matplotlib installation?
>  * When it's hung, if you press Ctrl-C so that execution terminates,
> what is output to the screen?  It should look like a traceback, where
> various functions and routines are listed in order and at the bottom
> it will say something like "KeyboardInterrupt"
>
> Thanks very much, and let's see if we can get this working for you.  :)
>
> -Matt
>
> On Wed, Apr 17, 2013 at 5:51 AM, Caroline Van Borm <borm at astro.rug.nl>
> wrote:
>
> Hi Nathan,
>
> I used the Enzo code (version 2.2) and was just running some of the included
> test problems. There is some example data available at
> http://lca.ucsd.edu/software/enzo/data/cookbook/ (I believe this is from an
> earlier version of Enzo), the plotting of which also results in the
> described problem. This is the first time I'm using yt.
>
> Thanks,
> Caroline
>
>
>
> On 04/16/2013 11:03 PM, Nathan Goldbaum wrote:
>
> Hi Caroline,
>
> Sorry to hear you're having trouble.
>
> Can you let us know what code generated the dataset you're trying to plot?
> Have you been able to visualize data produced by this code in the past?
>
> If you feel comfortable sharing your data and if it is small enough to be
> hosted on the web somewhere, it would be great if you could upload your
> dataset and either reply with a link to the list or send it to me privately
> so that one of us could try to take a look and see what exactly is going
> wrong.  I understand that datasets can sometimes be sensitive, however it
> can be slow and difficult to debug problems like this over e-mail.
>
> Cheers,
>
> Nathan
>
> On Apr 16, 2013, at 6:43 AM, Caroline Van Borm<borm at astro.rug.nl>  wrote:
>
> Dear yt users,
>
> I have a problem when trying to make a SlicePlot along the y-axis of any
> data, for any field. yt just hangs and no plot is created. For the x- and
> z-axes everything works fine, and a ProjectionPlot along the y-axis is also
> not a problem.
> I have tried updating yt and installing the current development version,
> but to no avail.
>
> Any help is much appreciated!
>
> Best regards,
> Caroline
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
>
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>



More information about the yt-users mailing list