We Keep Coding

How to use cppyy to embed Python instead of boost-python

I am currently using boost-python to embed a Python interpreter into my C++ application and facilitate passing data from the executed Python process to the running C++ application through the boost-python bindings, as per https://www.boost.org/doc/libs/1_75_0/libs/python/doc/html/tutorial/tutorial/embedding.html
I have some troubles with respect to performance, especially when calling wrapped functions with a large number of arguments, the overhead of parsing and boxing all those arguments for "passing" to the C++ "side" is considerable.
I checked alternatives to boost-python, for example pybind11, which can also be embedded, but the performance is unlikely to improve. I also found out about cppyy, but from the documentation I am at loss as how to facilitate embedding an interpreter into my program, or rather, how I should convert my current approach of an embedded interpreter to be able to use cppyy. My aim in trying out cppyy is to check whether using cppyy and/or PyPy as an interpreter can increase the performance of my code, since neither boost-python nor pybind11 support embedding PyPy.
Can anyone give any pointers in how to replace an embedded Python interpreter using boost-python with cppyy?

The cppyy embedding interface isn't documented yet because it's not functional on PyPy/cppyy (which seems to be what you're asking for most specifically), only for CPython. And for the latter, I don't necessarily see whether/how it would be faster than boost.python or pybind11, as it still relies on boxing variables and the C-API to call into Python. Potentially, C++ type lookups are faster, but that would be all.
You can play with it rather easily to get some performance numbers first by calling into Python from C++ (Cling) from cppyy and see how it looks. Here's a trivial example:
import cppyy
import time
N = 10000000
def pycall(a):
return a
cppyy.cppdef("""\
int (*ptr)(int) = 0;
void func(uint64_t N) {
for (uint64_t i = 0; i < N; ++i)
ptr(1);
}""")
cppyy.gbl.ptr = pycall
ts = time.perf_counter()
cppyy.gbl.func(N)
print('time per call:', (time.perf_counter()-ts)/N)
To use your own code and types, rather than int, just include headers with cppyy.include and load libraries with cppyy.load_library(). Both Cling on the C++ side and cppyy on the Python side will then have full access, so you can use the types in the callback.
If the numbers look better, the main pieces you need are in CPyCppyy/API.h, see here: https://github.com/wlav/CPyCppyy/blob/master/include/CPyCppyy/API.h
My best recommendation, at this point in time, would however have to be CFFI, which is C only, but will give you callbacks that you can use directly and that are JIT friendly on PyPy:
https://cffi.readthedocs.io/en/latest/embedding.html

Can one author a library API – one accessible outside Python – using Cython?

I know Cython's purpose is to create Python extensions modules, but can the compiled libraries made with Cython be loaded by non-python programs? If not, why?

I doubt that you can load them directly on a non-python program; looking at the C code generated by the simplest Cython script it's apparent that you need all the Python scaffolding to make it work. That said, you can do it indirectly for example from C++. In C++ I use boost.python to embed the Python interpreter and load some of my modules and scripts. This may seem convoluted, but allows you to quickly use whatever extensions you already have written in Python from C++, provided that you build the appropriate gluing code (see the boost.python wiki).
The disadvantage of this approach is that you are really loading a full Python interpreter just to be able to use some extension. This wasn't an issue for me since I already had the python extension and was embedding Python in my application to provide basic scripting ability, but I would not use this approach to write new libraries.

There are two mechanisms by which you can make cdef’ed Cython structures available externally:
Use a cdef public declaration – and/or also
Use the api keyword.
Here’s one of the examples of both mechanisms, from the aforelinked pages:
cdef public struct Vehicle:
int speed
float power
cdef api void activate(Vehicle *v):
if v.speed >= 88 and v.power >= 1.21:
print "Time travel achieved"
Each of these methods will instruct the Cython compiler to generate a header (“.h”) file that you can then integrate with your orthogonal C/C++ project.
A cdef public declaration yields a file named modulename.h; using the structures in this file will require you to link with the compiled Cython extension module.
An api declaration (which may be used simultaneously with cdef public, if you so desire) yields a modulename_api.h file; code consuming an api-based header will not need to link to the extension module – but it will need to call the cdef’d function import_modulename() before any of the API code can be used (a tactic the Cython-inclined users of NumPy will find familiar).
In my personal experience, it takes little effort to expose and subsequently employ encythoned structures as an external API in this fashion, provided that the struct layouts dovetail well with the consuming code, and that you’re willing to contend with manually managing the GIL in your C/C++ code in order to make things work.

LIBSVM Python ctypes string function pointer segmentation faults

I've been porting a Python package that uses libsvm onto some production servers and ran into a strange segmentation fault which I traced to a ctypes function pointer. I'm trying to determine where the ctypes wrapper failed and if this is a distro specific problem or not.
The system I am running this on is a very clean virtual machine with almost nothing installed:
Solaris 5.11
amd64 pentium_pro+mmx pentium_pro pentium+mmx pentium i486 i386 i86
Python 2.7.2
Now for the problem description and how I narrowed to ctypes. In libsvm you can specify the print function by passing a void (*print_func)(const char *) pointer into the svm_set_print_string_function function. The default with a NULL pointer is to print to stdout. Now the interesting part is that the Python wrapper for libsvm (which works fine on a variety of other systems) makes such a function pointer when asking for quiet mode (no printing) via the following:
PRINT_STRING_FUN = CFUNCTYPE(None, c_char_p)
def print_null(s):
return
if argv[i] == "-q":
self.print_func = PRINT_STRING_FUN(print_null)
libsvm.svm_set_print_string_function(self.print_func)
When I set quiet mode libsvm accepts the function pointer but hangs after a few seconds when calling svm_train then seg faults. I tried making a void * argument function pointer and then casting it to a const char * function pointer with the same results, which means it wasn't the conversion from const char * to a PyStringObject.
Then I finally just wrote a C++ function to set the function pointer to a no-op in the library itself by:
void print_null(const char *) {}
void svm_set_print_null() {
svm_set_print_string_function(&print_null);
}
which worked as expected with no segmentation faults. This leads me to think that the ctypes is failing at some internal point of function pointer conversion. Looking through the ctypes source files hasn't revealed anything obvious to me though I haven't worked a lot with ctypes explicitly so it's difficult to narrow down where the bug might be.
I can use my library addition solution for now, but if I want to silently process the returns I would need to actually be able to pass a function pointer into libsvm. Plus it doesn't give me peace of mind about stability if I need to implement such workarounds without knowing what's the true root cause of the problem.
Has anyone else had problems with libsvm print functions on Solaris or specifically ctypes function pointers in Python on Solaris? I couldn't find anything searching online about either such problems with Solaris. I planning on playing around with library calls and making some function processing libs to find the exact boundaries of failure, but someone else's input might save me a day or two of debug testing.
UPDATE
The problem is reproducible on the 32bit version of Solaris 5.11 as well.

Is wrapping C++ library with ctypes a bad idea?

I read through the following two threads on wrapping C library and C++ library, I am not sure I get it yet. The C++ library I am working with does use class and template, but not in any over-sophisticated way. What are issues or caveats of wrapping it with ctypes (besides the point that you can do so in pure python etc)?
PyCXX , Cython and boost::python are three other choices people mentioned, is there any consensus which one is more suitable for C++?
Thanks
Oliver

In defence of boost::python, given Alexander's answer on ctypes:
Boost python provides a very "c++" interface between c++ and python code - even doing things like allowed python subclasses of c++ classes to override virtual methods is relatively straightforward. Here's a potted list of good features:
Allow virtual methods of C++ classes to be overridden by python subclasses.
Bridge between std::vector<>, std::map<> instances and python lists and dictionaries (using vector_indexing_suite and map_indexing_suite)
Automatic sharing of reference counts in smart pointers (boost::shared_ptr, etc) with python reference counts (and you can extend this to any smart pointer).
Fine grained control of ownership when passing arguments and returning values from functions.
Basically, if you have a design where you want to expose a c++ interface in a way faithful to the language, then boost::python is probably the best way to do it.
The only downsides are increased compile time (boost::python makes extensive use of templates), and sometimes opaque error messages if you don't get things quite right.

For C++ a library to be accessible from Python it must use C export names, which basically means that a function named foo will be accessible from ctypes as foo.
This can be achieved only by enclosing the public interface with export C {}, which in turn disallows function overloading and templates therein (only the public interface of the library to be wrapped is relevant, the inner workings are not and may use any C++ features they like).
Reason for this is that C++ compilers use a mechanism called name mangling to generate unique names for overloaded or templated symbols. While ctypes would still find a function provided you knew its mangled name, the mangling scheme depends on the compiler/linker being used and is nothing you can rely on. In short: do not use ctypes to wrap libraries that use C++ features in their public interface.
Cython takes a different approach. It aids you at building a C extension module that does the interfacing with the original library. Therefore, linking to the C++ library is done by the regular C++ linkage mechanism, thus avoiding the aforementioned problem. The trouble with Cython is that C extension libraries need to to be recompiled for every platform, but anyway, this applies to the C++ library to be wrapped as well.
Personally, I'd say that in most cases the time to fire up Cython is a time that is well-spent and will eventually pay off in comparison to ctypes (with an exception for really simple Cish interfaces).
I don't have any experience with boost.python, so I can't comment on it (however, I don't have the impression that it is very popular either).

Prototyping with Python code before compiling

I have been mulling over writing a peak-fitting library for a while. I know Python fairly well and plan on implementing everything in Python to begin with but envisage that I may have to re-implement some core routines in a compiled language eventually.
IIRC, one of Python's original remits was as a prototyping language, however Python is pretty liberal in allowing functions, functors, objects to be passed to functions and methods, whereas I suspect the same is not true of say C or Fortran.
What should I know about designing functions/classes which I envisage will have to interface into the compiled language? And how much of these potential problems are dealt with by libraries such as cTypes, bgen, SWIG, Boost.Python, Cython or Python SIP?
For this particular use case (a fitting library), I imagine allowing users to define mathematical functions (Guassian, Lorentzian etc.) as Python functions which can then to be passed an interpreted by the compiled code fitting library. Passing and returning arrays is also essential.

Finally a question that I can really put a value answer to :).
I have investigated f2py, boost.python, swig, cython and pyrex for my work (PhD in optical measurement techniques). I used swig extensively, boost.python some and pyrex and cython a lot. I also used ctypes. This is my breakdown:
Disclaimer: This is my personal experience. I am not involved with any of these projects.
swig:
does not play well with c++. It should, but name mangling problems in the linking step was a major headache for me on linux & Mac OS X. If you have C code and want it interfaced to python, it is a good solution. I wrapped the GTS for my needs and needed to write basically a C shared library which I could connect to. I would not recommend it.
Ctypes:
I wrote a libdc1394 (IEEE Camera library) wrapper using ctypes and it was a very straigtforward experience. You can find the code on https://launchpad.net/pydc1394. It is a lot of work to convert headers to python code, but then everything works reliably. This is a good way if you want to interface an external library. Ctypes is also in the stdlib of python, so everyone can use your code right away. This is also a good way to play around with a new lib in python quickly. I can recommend it to interface to external libs.
Boost.Python: Very enjoyable. If you already have C++ code of your own that you want to use in python, go for this. It is very easy to translate c++ class structures into python class structures this way. I recommend it if you have c++ code that you need in python.
Pyrex/Cython: Use Cython, not Pyrex. Period. Cython is more advanced and more enjoyable to use. Nowadays, I do everything with cython that i used to do with SWIG or Ctypes. It is also the best way if you have python code that runs too slow. The process is absolutely fantastic: you convert your python modules into cython modules, build them and keep profiling and optimizing like it still was python (no change of tools needed). You can then apply as much (or as little) C code mixed with your python code. This is by far faster then having to rewrite whole parts of your application in C; you only rewrite the inner loop.
Timings: ctypes has the highest call overhead (~700ns), followed by boost.python (322ns), then directly by swig (290ns). Cython has the lowest call overhead (124ns) and the best feedback where it spends time on (cProfile support!). The numbers are from my box calling a trivial function that returns an integer from an interactive shell; module import overhead is therefore not timed, only function call overhead is. It is therefore easiest and most productive to get python code fast by profiling and using cython.
Summary: For your problem, use Cython ;). I hope this rundown will be useful for some people. I'll gladly answer any remaining question.
Edit: I forget to mention: for numerical purposes (that is, connection to NumPy) use Cython; they have support for it (because they basically develop cython for this purpose). So this should be another +1 for your decision.

I haven't used SWIG or SIP, but I find writing Python wrappers with boost.python to be very powerful and relatively easy to use.
I'm not clear on what your requirements are for passing types between C/C++ and python, but you can do that easily by either exposing a C++ type to python, or by using a generic boost::python::object argument to your C++ API. You can also register converters to automatically convert python types to C++ types and vice versa.
If you plan use boost.python, the tutorial is a good place to start.
I have implemented something somewhat similar to what you need. I have a C++ function that
accepts a python function and an image as arguments, and applies the python function to each pixel in the image.
Image* unary(boost::python::object op, Image& im)
{
Image* out = new Image(im.width(), im.height(), im.channels());
for(unsigned int i=0; i<im.size(); i++)
{
(*out)[i] == extract<float>(op(im[i]));
}
return out;
}
In this case, Image is a C++ object exposed to python (an image with float pixels), and op is a python defined function (or really any python object with a __call__ attribute). You can then use this function as follows (assuming unary is located in the called image that also contains Image and a load function):
import image
im = image.load('somefile.tiff')
double_im = image.unary(lambda x: 2.0*x, im)
As for using arrays with boost, I personally haven't done this, but I know the functionality to expose arrays to python using boost is available - this might be helpful.

The best way to plan for an eventual transition to compiled code is to write the performance sensitive portions as a module of simple functions in a functional style (stateless and without side effects), which accept and return basic data types.
This will provide a one-to-one mapping from your Python prototype code to the eventual compiled code, and will let you use ctypes easily and avoid a whole bunch of headaches.
For peak fitting, you'll almost certainly need to use arrays, which will complicate things a little, but is still very doable with ctypes.
If you really want to use more complicated data structures, or modify the passed arguments, SWIG or Python's standard C-extension interface will let you do what you want, but with some amount of hassle.
For what you're doing, you may also want to check out NumPy, which might do some of the work you would want to push to C, as well as offering some additional help in moving data back and forth between Python and C.

f2py (part of numpy) is a simpler alternative to SWIG and boost.python for wrapping C/Fortran number-crunching code.

In my experience, there are two easy ways to call into C code from Python code. There are other approaches, all of which are more annoying and/or verbose.
The first and easiest is to compile a bunch of C code as a separate shared library and then call functions in that library using ctypes. Unfortunately, passing anything other than basic data types is non-trivial.
The second easiest way is to write a Python module in C and then call functions in that module. You can pass anything you want to these C functions without having to jump through any hoops. And it's easy to call Python functions or methods from these C functions, as described here: https://docs.python.org/extending/extending.html#calling-python-functions-from-c
I don't have enough experience with SWIG to offer intelligent commentary. And while it is possible to do things like pass custom Python objects to C functions through ctypes, or to define new Python classes in C, these things are annoying and verbose and I recommend taking one of the two approaches described above.

Python is pretty liberal in allowing functions, functors, objects to be passed to functions and methods, whereas I suspect the same is not true of say C or Fortran.
In C you cannot pass a function as an argument to a function but you can pass a function pointer which is just as good a function.
I don't know how much that would help when you are trying to integrate C and Python code but I just wanted to clear up one misconception.

In addition to the tools above, I can recommend using Pyrex
(for creating Python extension modules) or Psyco (as JIT compiler for Python).