Welcome to CodePy’s documentation!¶
CodePy is a C metaprogramming toolkit for Python. It handles two aspects of metaprogramming:
- Generating C source code.
- Compiling this source code and dynamically loading it into the Python interpreter.
Both capabilities are meant to be used together, but also work on their own. In particular, the code generation facilities work well in conjunction with PyCuda. Dynamic compilation and linking are so far only supported in Linux with the GNU toolchain.
A taste of CodePy¶
This sample CodePy program builds a Boost.Python C++ module that returns the string “hello world”:
from cgen import *
from codepy.bpl import BoostPythonModule
mod = BoostPythonModule()
mod.add_function(
FunctionBody(
FunctionDeclaration(Const(Pointer(Value("char", "greet"))), []),
Block([Statement('return "hello world"')])
))
from codepy.toolchain import guess_toolchain
cmod = mod.compile(guess_toolchain())
print cmod.greet()
Boost.Python and CodePy¶
You may notice that the above code snippet refers to Boost.Python. Boost.Python is a Python wrapper generator library for C++. CodePy does not depend on it being available, but its use is strongly encouraged–otherwise the generated sourcecode would need to use another wrapper generator or talk directly to the Python C API to make its functionality accessible from Python. Boost.Python is only used at run time. CodePy has no install-time dependencies
Instructions
on how to install Boost.Python are available. As described in the
instructions, CodePy looks for the installation location of Boost.Python
in $HOME/.aksetup-defaults.py and /etc/aksetup-defaults.py.
Note
Boost.Python is not needed at all if CodePy is used to generate code for PyCuda.
Contents¶
If you are looking for the codepy.cgen module, it has been moved to a
separate package called cgen, see also that package’s package index
page.