iyt, IPython
by Matthew Turk
Hi guys,
Earlier today Jeff Oishi committed a script called 'iyt' to
trunk/scripts/ . If you've got IPython (http://ipython.scipy.org/)
installed and you run iyt, it will load up an IPython session with the
appropriate threading necessary to access the 'pylab' interface to
Matplotlib, as well as all the yt modules pre-imported and in the
current namespace. This, combined with the PlotCollectionInteractive
object, simplify interactive exploration of data significantly.
(IPython is pretty sweet -- I bet Jeff has some great tips; I'm not an
expert on it.)
If you're on branches/yt-1.0 , it'll be committed over there after it
gets tested a bit.
-Matt
10 years, 3 months
New means of simple interaction
by Matthew Turk
Hi guys,
Are you tired of having to include lots of boilerplate code in your
python scripts?
I've committed a change (r737) in trunk/ that should help cut down on
this -- but might also make things significantly more terse!
The idea has long been that there were standard workflows in yt --
generating static outputs, adding these to plot collections, and
making plots. These are now accessible from a pipe-like interface.
For example:
>>> from yt.mods import *
>>> pf = "galaxy1200" | static
>>> pf | slicer
>>> pf | slicer(field="Temperature", width=(1.0,'kpc'))
>>> pf | slicer_y(field="H2I_Fraction", width=(1.0,'au'))
>>> pf | projector(field="Density", width=(1.0,'kpc'), name='hiya')
or if you want to be ultra terse, forget the pf part:
>>> from yt.mods import *
>>> "galaxy1200" | slicer
>>> pf | projector(field="Temperature", weight_field="Density", width=(1.0,'kpc'), name='hiya')
I'm thinking about adding some more of these -- for contour finding
among other things -- and using iterators to speed this whole workflow
up.
Let me know if you like this, or if there are problems, or whatever!
There are likely some bugs, but I've run a bunch of tests and it seems
to work.
-Matt
10 years, 3 months
Reason improvement
by Matthew Turk
Greetings,
Just a note -- in revision 719 -- I've committed a fix for the GUI
component (reason) that was causing huge and inappropriate delays to
projecting a dataset, if that dataset had more than about 10,000
grids. The GUI was updating every time a new grid was masked and
examined, which was oftentimes many times more than once per second.
It should be working now. Fixes have been committed to both trunk/
and branches/yt-1.0 .
-Matt
10 years, 4 months