[IronPython] SciPy

Mon Dec 20 21:56:20 CET 2010

There's also MetaNumerics (http://metanumerics.codeplex.com/)

On Mon, Dec 20, 2010 at 10:32 AM, Mark Senko <msenko at completegenomics.com>wrote:

>  Thank you.
>
> My real goal is to find a decent math/numerical package that I can use
> without having to reinvent the wheel.
>
> My searches came up with numpy/SciPy.  Whether it actually uses the C
> interface or is a refactored version for .NET  really doesn’t matter to me.
>
>
>
> At least not much.
>
>
>
> I think I’ll take this opportunity to explain to this portion of the
> IronPython community (which seems to contain most of the real “movers” ) why
> I chose IronPython, and what I want to accomplish with it.
>
> I’m sure my use case may be of interest, at least academically, to some of
> you.
>
>
>
> I write software that controls hardware … it moves samples around with a
> robot and stage, controls shutters, light sources, moves optics, performs
> data acquisition and data analysis.
>
> It’s a complicated system. It requires scientists and engineers with high
> levels of knowledge in optics, physics, electronics to develop, study and
> understand performance, and to calibrate.
>
>
>
> At my old company, we wrote our own macro language (early ‘90s) which
> provided functions to move the pieces of hardware, call higher level
> routines in our code, call mathematical and higher level analysis routines,
> plot, macros could call other macros  … just about anything you want a
> scripting language to do.  The portion of my efforts I spent extending this
> macro language was time well spent.  Instead of being the bottleneck that
> had to write all new tests and experiments and new platform prototyping, I
> enabled the other scientists and engineers to do it themselves. I was an
> enabler, and productivity shot through the roof.
>
> Some of the macros written were simple, others complicated … but most were
> not written by me, they were written by the scientists and engineers.
>
>
>
> But, they were not programmers … at least not in the computer scientist
> sense.  They think procedurally, not object oriented.  They don’t want to
> argue about global vs. local variables, or whether a GOTO is good
> programming style.  They don’t develop web pages.
>
>
>
> At my new company, which builds a tool with similar demands, I am
> implementing similar scripting abilities.  I didn’t want to write my own
> scripting language again, that would be crazy with the number that are
> already available.
>
> The core needs are control over the hardware, numerical capabilities,
> plotting capabilities.  The language also needs to be straightforward
> without too much overhead, like a pile of import statements, that don’t
> really contribute to the functionality.  I also wanted a language with a
> some history, and lot’s of community written libraries that I could use
> without having to write my own.
>
>
>
> I did my internet search and looked at many different scripting languages,
> finally settling on Python as having the best set of language features for
> my needs.
>
> Granted, I HATE the indentation control, especially since an auto-indented
> line has a “different” indentation than the preceding line which was
> indented with spaces (unless you carefully set up your editor).
>
> I would much rather see braces or ENDIF,ENDFOR, ENDDEF … statements. That’s
> just an aside …
>
>
>
> Our company uses C#.  I quickly discovered that C# and Python don’t play
> well together …. Ah, but here is IronPython.
>
> So I learned how to embed it, wrote my own console, learned how to make
> static wrappers, how to make my functions global, and how to make python
> functions global.  I’m still learning the best and easiest ways to use it
> for our needs.
>
> But, I’m starting to find that the community developed libraries I was
> counting on are more often than not out of reach. That is what I found when
> I started looking for  a simple math package that would fit a polynomial,
> perform an FFT, maybe even do a non-linear least squares fit.  And I still
> need to find a plotting package …
>
>
>
> Anyway, I thought this might be interesting to some of you.
>
>
>
>
>
> *Mark Senko*
>
> Complete Genomics, Inc.
>
> 2071 Stierlin Court
>
> Mountain View, CA 94043
>
>
>
> *From:* users-bounces at lists.ironpython.com [mailto:
> users-bounces at lists.ironpython.com] *On Behalf Of *Jason McCampbell
> *Sent:* Monday, December 20, 2010 7:13 AM
> *To:* Discussion of IronPython
> *Subject:* Re: [IronPython] SciPy
>
>
>
> Hi Mark,
>
>
>
> As Dino mentioned we (Enthought) are working on refactoring Numpy into a
> pure "C" core with CPython and IronPython interface layers.  This is largely
> complete and available at github (https://github.com/numpy/numpy-refactor),
> though the core layer is largely undocumented thus far.  This is the
> multi-dimensional array.
>
>
>
> SciPy is in progress and we are updating it to work with the refactored
> numpy core and to add an IronPython interface.
>
>
>
> I assume you are looking for IronPython interfaces to SciPy as opposed to a
> C interface, correct?
>
>
>
> Regards,
>
> Jason
>
>
>
> On Thu, Dec 16, 2010 at 1:56 PM, Dino Viehland <dinov at microsoft.com>
> wrote:
>
> Enthought has been working on getting numpy/scipy ported over to work w/
> IronPython.  I believe numpy is working but I’m not sure of how far along
> SciPy is.  There’s a separate mailing list for this at:
>
>
>
> https://mail.enthought.com/mailman/listinfo/scipy4dotnet
>
>
>
> It’s very low traffic – it’s usually just working through issues Enthought
> has run into and either workarounds or suggested changes to IronPython.  I’d
> suggest sending a mail there – they might have something you can try.
>
>
>
>
>
> *From:* users-bounces at lists.ironpython.com [mailto:
> users-bounces at lists.ironpython.com] *On Behalf Of *Mark Senko
> *Sent:* Thursday, December 16, 2010 11:49 AM
> *To:* users at lists.ironpython.com
> *Subject:* [IronPython] SciPy
>
>
>
> I’ve been searching for the current state of support for “C” based
> libraries, specifically SciPy (I’m just looking for a decent numerical
> analysis package).  The responses I’ve seen on various websites are somewhat
> dated.
>
> What is the latest status, or is there no effort towards accommodating the
> C API? Is IronClad still the best option? Any info, suggestions and warnings
> would be appreciated before I start to invest a lot of time into installing
> and learning these packages.
>
>
>
> *Mark Senko*
>
> Complete Genomics, Inc.
>
> 2071 Stierlin Court
>
> Mountain View, CA 94043
>
>
>
>
>
>
>
> ____
>
>
>
> The contents of this e-mail and any attachments are confidential and only for
>
> use by the intended recipient. Any unauthorized use, distribution or copying
>
> of this message is strictly prohibited. If you are not the intended recipient
>
> please inform the sender immediately by reply e-mail and delete this message
>
> from your system. Thank you for your co-operation.
>
>
> _______________________________________________
> Users mailing list
> Users at lists.ironpython.com
> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
>
>
>
> ____
>
> The contents of this e-mail and any attachments are confidential and only for
> use by the intended recipient. Any unauthorized use, distribution or copying
> of this message is strictly prohibited. If you are not the intended recipient
> please inform the sender immediately by reply e-mail and delete this message
> from your system. Thank you for your co-operation.
>
>
> _______________________________________________
> Users mailing list
> Users at lists.ironpython.com
> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ironpython-users/attachments/20101220/ee8e565c/attachment.html>