Access C-Declared Nested Template from Python

Question

I statically declared a large structure in C, but I need to use the same data to do some analysis in Python. I would prefer not to replicate this data in Python to avoid errors, is there a way (read-only) to access the data directly in Python? I looked at "ctypes" and SWIG, and none of them seem to provide what I'm looking for ...

For example, I have:

/*.h file * /

typedef struct { double data[10]; } NestedStruct; typedef struct { NestedStruct array[10]; } MyStruct; 

/*.c file * /

 MyStruct the_data_i_want = { {0}, { {1,2,3,4} }, {0}, }; 

Ideally, I would like something that would allow me to get this in python and access it through the_data_i_want.array[1].data[2] or something like that. Any thoughts? I got swig to “work” in the sense that I was able to compile / import the .so created from my .c file, but I could not access it through cvars. Maybe there is another way? It seems to be so hard ....

---

Actually, I realized that. I add this because my reputation does not allow me to answer my question within 8 hours, and since I do not want to remember after 8 hours, I will add it now. I am sure there is a good reason for this that I do not understand.

Figured it out.

1st I compiled my .c file to a library:

Then I used types to define a python class that stores data:

 from ctypes import * class NestedStruct(Structure): _fields_ = [("data", c_double*10)] class MyStruct(Structure): _fields_ = [("array", NestedStruct*10)] 

Then I loaded the shared library in python:

 my_lib = cdll.LoadLibrary("my_lib.so") 

Then I used the in_dll method to get the data:

 the_data_i_want = MyStruct.in_dll(my_lib, "the_data_i_want") 

Then I could access it as if it were C. the_data_i_want.array[1].data[2]

Note. Maybe I messed up the syntax a bit because my actual data structure is nested 3 levels and I wanted to simplify it here for illustration.

Answer 1

You could also read the data in C and write to a JSON file that you could easily parse (usually there is a library that will even do this for you, python import json ) and access any other platform with almost every language setting, which you might think about. And at the same time, you could access you, the data is very similar compared to the way you accessed it in the original C code.

As a suggestion. This will make your data more portable and versatile, I think, but you will spend more time writing and parsing JSON, as if you were just reading the data stream directly from your C code in python.