podschwadt / cpp_test

Your task is to implement a simple matrix class using Python. The implementation of that class needs to be matrix.py. A matrix needs to be constructed from nested list objects. Additionally, it needs to support elementwise addition, elementwise multiplication, identity comparison, and matrix multiplication. Using the resulting code should look like this:

A = Matrix([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
B = Matrix([[9, 8, 7], [6, 5, 4], [3, 2, 1]])

C = A + B

However, when a Matrix object is created in Python, a corresponding object in C++ needs to be created as well. The C++ object needs access to the values that make up the matrix. The arithmetic operations (addition, multiplication, matrix multiplication) also need to be performed in C++. On C++ side, there are the classes Interface and C_Matrix (declared in the files interface.h and c_matrix.h). In your implementation, you need to implement the class C_Matrix. You can create additional classes and functions as you need, but you can not make changes to interface.h. The code that connects the Python and C++ objects needs to use ctypes. Besides the constructor, you should only call functions on C_Matrix instances that the functions declared in Interface. The implementation of the arithmetic functions should ideally be multi-threaded. A possible flow of data could look as follows:

In Python:
  A = Matrix([[1, 2, 3], [4, 5, 6], [7, 8, 9]]

  The Matrix constructor calls a ctypes function that creates a C_Matrix instance.

In C++:
  The functions called by cytpes read the matrix values and create a new 
  C_Matrix instance.

You need come up with a meachnism to connect the Python and C++ objects.


In Python: 
  B = A + A

  Calls the + operator of Matrix, which forwards using cytpes

In C++:
  Identify the two C_Matrix involved in the operation and return a new    C_Matrix instance

You need to submit your source code along with a build file that compiles the C++ code. To summarize, your tasks are:

• Create a Python class matrix that is backed by C++ objects.
• The connection between Python and C++ needs to be done using ctypes.
• Addition, multiplication, and matrix multiplication need to be done in C++; ideally using multithreading.
• Addition, multiplication, and matrix multiplication need to use the API specified by Interface

Important Notes:

• You are not allowed to use modules other than ctypes.
• You are not allowed to use libraries that are not in C++ standard library.
• You can view matrices as immutable.
• You only need to support element-wise operations for matrices of the same dimension