[Numpy-discussion] [SciPy-user] median filter with clipping
Glen W. Mabey
Glen.Mabey at swri.org
Thu May 17 09:48:17 EDT 2007
On Thu, May 17, 2007 at 02:54:22AM -0600, Travis Oliphant wrote:
> Pierre GM wrote:
> >> How far away is maskedarray from being able to replace numpy.ma?
> >>
> >
> > So far, it does everything that numpy.core.ma does, with I believe more
> > flexibility and some additional features (hard/soft mask, easy
> > subclassing...). Personally, I stopped using numpy.core.ma completely (unless
> > for test purposes), and I even managed to convince another user to switch to
> > this module for his own work (cf the TimeSeries package).
> >
> > Of course, I expect that there are still some bugs here and there, but I'm
> > working on it (when I find them). It's a tad slower than numpy.core.ma, but
> > that's a side effect of the flexibility. In the long term, there are some
> > plans about porting the module to C, but we're talking in quarters more than
> > in weeks.
> >
> > About when it'll be promoted outside the sandbox: well, we need more feedback
> > from users, as usual. I guess that's the principal stopping block. I'd be
> > quite grateful if you could try it and let me know what you think. I grew
> > fond of this child born in pain (explaining to my bosses why I spent so much
> > time on something which is only remotely connected to what I paid to do...),
> > so I make sure that the baby behaves...
I never received this response from Pierre, even though I have a
white-filter on scipy-user ... strange.
> I'm inclined to move his masked array over to ma wholesale. The fact
> that Pierre sees it as his baby is very important to me. If it doesn't
> have significant compatibility issues than I'm all for it. I'm mainly
> interested in hearing how people actually using numpy.ma would respond.
>
> Perhaps we should move it over and start deprecating numpy.ma??
Meaning that there would exist both numpy.ma and numpy.maskedarray for a
period? That sounds great to me.
I'll happily switch over once I have more confidence in its stability.
Glen
More information about the NumPy-Discussion
mailing list