Question

I have a standalone Python module used to perform analysis on some raw data. The module is working great.

Now I need the output generated by the Python module in C source that will do further processing on output.

Here is the rough idea of flow:

1. C source will call the Python module.
2. Python module will fill the results generated into a C structure and
   then somehow pass it to C source.

So what is the best way to get the results from Python module in C source code?

Should I use something like the following to fill the results?

from ctypes import *
class result(Structure):       
    _fields_ = [("status", c_int), ...]

Answer 1

I don’t think you’ll want to use the ctypes module — that module is for accessing DLLs and shared objects from Python code, not the other way around.

You don’t specify how the C program is going to invoke the Python module. Will it

• invoke the Python script using something like popen from C? If so, then you can write data to stdout in Python and interpret the data appropriately in C. You’ll have to design your own little protocol for this. This can be an ad-hoc protocol, or you can use existing formats like JSON, for which you’ll find libraries for both C and Python.

• embed the Python VM in your C program? Then you’ll need to convert the C values into appropriate Py* values and put them into a PyTuple that you can pass as an argument to the Python function, from C. The output of the Python function will be a PyObject, which you’ll have to unpack (in C) to get the data.

In both cases, you’ll need to do the packing and unpacking yourself.