From heroxbd at gmail.com  Wed Dec  4 16:05:00 2013
From: heroxbd at gmail.com (heroxbd at gmail.com)
Date: Thu, 05 Dec 2013 00:05:00 +0900
Subject: [C++-sig] numpy integration with boost.python
Message-ID: <86iov4y7vn.fsf@moguhome00.in.awa.tohoku.ac.jp>

Dear all,

This is a cross post of stackoverflow[0].

I am interfacing a C++ data intense library with Python by
py++[1]/boost.python[2]. After profiling my program, I find 70% of run
time is spent on code like this:

    ni = range(v2o.getHits())
    tau = np.array([v2o.TofCorrectedTime[i] for i in ni])
    q = np.array[v2o.getCharge()[i] for i in ni]

v2o.TofCorrectedTime is typed __array_1_float_2368[3] from
py++. v2o.getCharge() is typed _impl_details_range_iterator_[4] from
py++, too. Size being about 2000, the convertion from these py++ array
wrappers to numpy is slow:

    In [42]: timeit np.array(v2o.TofCorrectedTime)
    100 loops, best of 3: 2.52 ms per loop

    In [43]: timeit np.array(v2o.getCharge())
    100 loops, best of 3: 4.94 ms per loop

    In [44]: timeit np.array([0]*2368)
    1000 loops, best of 3: 310 ?s per loop

    In [45]: timeit np.array(np.zeros(2368))
    100000 loops, best of 3: 4.41 ?s per loop

I searched the web for a solution. The candidates are:

 1. Cython[5] with memoryview[6]
 2. pyublas[7]
 3. Boost.NumPy[8]

Questions:
 
 - Is cython/memoryview easy to be integrated with boost.python and
   py++? I want to keep the rest of the library wrapper.

 - Which one best suites my problem in terms of convertion overhead?

Thanks,
Benda

  [0]: http://stackoverflow.com/questions/20376021/exchange-numpy-array-with-boost-python-pyublas-or-boost-numpy
  [1]: http://svn.code.sf.net/p/pygccxml/svn/pyplusplus_dev
  [2]: http://www.boost.org/libs/python
  [3]: http://svn.code.sf.net/p/pygccxml/svn/pyplusplus_dev/pyplusplus/code_repository/array_1.py
  [4]: http://svn.code.sf.net/p/pygccxml/svn/pyplusplus_dev/pyplusplus/code_repository/return_range.py
  [5]: http://cython.org
  [6]: http://docs.cython.org/src/userguide/memoryviews.html
  [7]: http://mathema.tician.de/software/pyublas/
  [8]: https://github.com/ndarray/Boost.NumPy

From ndbecker2 at gmail.com  Thu Dec  5 13:00:26 2013
From: ndbecker2 at gmail.com (Neal Becker)
Date: Thu, 05 Dec 2013 07:00:26 -0500
Subject: [C++-sig] numpy integration with boost.python
References: <86iov4y7vn.fsf@moguhome00.in.awa.tohoku.ac.jp>
Message-ID: <l7ppsg$qdl$1@ger.gmane.org>

heroxbd at gmail.com wrote:

> Dear all,
> 
> This is a cross post of stackoverflow[0].
> 
...
> I searched the web for a solution. The candidates are:
> 
>  1. Cython[5] with memoryview[6]
>  2. pyublas[7]
>  3. Boost.NumPy[8]
> 
...
I've been using ndarray lately:
https://github.com/ndarray/ndarray


From heroxbd at gmail.com  Fri Dec  6 14:55:09 2013
From: heroxbd at gmail.com (heroxbd at gmail.com)
Date: Fri, 06 Dec 2013 22:55:09 +0900
Subject: [C++-sig] numpy integration with boost.python
In-Reply-To: <86iov4y7vn.fsf@moguhome00.in.awa.tohoku.ac.jp>
 (heroxbd@gmail.com's message of "Thu, 05 Dec 2013 00:05:00 +0900")
References: <86iov4y7vn.fsf@moguhome00.in.awa.tohoku.ac.jp>
Message-ID: <86mwkeuls2.fsf@moguhome00.in.awa.tohoku.ac.jp>

Dear all,

heroxbd at gmail.com writes:

> This is a cross post of stackoverflow[0].
>
>   [0]: http://stackoverflow.com/questions/20376021/exchange-numpy-array-with-boost-python-pyublas-or-boost-numpy

Feedback as a happy user. Jim has answered on stackoverflow. I decided
to go with Boost.NumPy.

And thanks Neal for sharing with your experience.

The array data needs not to be copied anymore, giving a 2000-sized array
exposing to Python 1000x faster than iteration- over-elements scheme.

I had tens of such arrays to maintain. Too tedious to do it manually. So
I made a patch against Py++ SVN trunk[1]. I think Roman is in this
list. Could you please review it?

Jim, thanks again! I hope boost.python v3 is going on well.

Cheers,
Benda

1. https://sourceforge.net/p/pygccxml/patches/2/

From catchall at schlipf.it  Mon Dec 16 16:06:19 2013
From: catchall at schlipf.it (Mario Schlipf)
Date: Mon, 16 Dec 2013 16:06:19 +0100
Subject: [C++-sig] Boost.Python: How to expose std::unique_ptr
Message-ID: <008b01cefa70$6107b7e0$231727a0$@schlipf.it>

Hi all,

I am farily new to boost.python and trying to expose the return value of a
function to python.

The function signature looks like this:

 	std::unique_ptr<Message> someFunc(const std::string &str) const;


When calling the function in python, I get the following error:

	TypeError: No to_python (by-value) converter found for C++ type:
std::unique_ptr<Message, std::default_delete<Message> >

My function call in python looks like this:

	a = mymodule.MyClass()
	a.someFunc("some string here") # error here

I tried to expose the smart pointer with the following statements:

	class_<std::unique_ptr<Message, std::default_delete<Message>>,
bost::noncopyable ("Message", init<>()) ;

This example compiles, but I still get the error mentioned above.
Also, I tried to expose the class `Message` itself

	class_<Message>("Message", init<unsigned>())
		.def(init<unsigned, unsigned>()) 
		.def("f", &Message::f)
	;

Does someone know how to properly expose the pointer class?
Thanks!