Return numpy array from python to Lazarus

[Chemistza74]

Hi,

I've recently started playing around with P4D in Lazarus/FPC. I'm trying to read a numpy array created in python back in to FPC. The python code which I have in SynEdit1:

import numpy as np

py_array = np.zeros(10)
for i in range (10):
    py_array[i]=i

My attempt at the FPC code below to extract py_array is based on FPC Demo35:

const
  N = 10;
type
  PIntArray = ^TIntArray;
  TIntArray = array[0..N - 1] of NativeInt;  
procedure TForm1.Button1Click(Sender: TObject);
var
  np_arr: PPyObject;
  PyBuffer: Py_buffer;
  I:integer;
  res:int64;
begin
  self.PythonEngine1.ExecStrings(self.SynEdit1.lines);
  np_arr := ExtractPythonObjectFrom(MainModule.py_array);
  PythonEngine1.PyObject_GetBuffer(np_arr, @PyBuffer, PyBUF_CONTIG);
  PythonEngine1.CheckError;
  try
    for I := 0 to N - 1 do
      res:= PIntArray(PyBuffer.buf)^[I];
      self.ListBox1.Items.Add(inttostr(res));
  finally
    PythonEngine1.PyBuffer_Release(@PyBuffer);
  end;
end;

However, when I run this, I only end up with a single, random number in ListBox1. Any suggestions on how to resolve this would be greatly appreciated!

[pyscripter]

If you run this code in PyScripter

import numpy as np

py_array = np.zeros(10)
for i in range (10):
    py_array[i]=i

print(py_array)

The output is:

[0. 1. 2. 3. 4. 5. 6. 7. 8. 9.]

Notice that the values are floating point and you try to read them as Integers.

I think if you change your code to:

import numpy as np

py_array = np.zeros(10, dtype= np.int32)
for i in range (10):
    py_array[i]=i

or simpler:

import numpy as np
py_array = np.array(range(10))

it should work.