Python ctypes: calling a function with custom types in c

Question

I'm trying to wrap pre-existing c code for use in Python in Linux. I have little experience with c, and I'm currently approaching this problem using ctypes. My C function requires a 2d array with custom type entries and I don't know how to recreate this in python to pass it to the c function.

Here is the function I'm attempting to call:

void usbBuildGainTableAI_USB1808(libusb_device_handle *udev, Calibration_AIN table[NCHAN_1808][NGAINS_1808])
{

  int i, j;
  uint16_t address = 0x7000;  // base address of ADC calibration coefficients.

  for (i = 0; i < NCHAN_1808; i++) {
    for (j = 0; j < NGAINS_1808; j++) {
      usbMemAddressW_USB1808(udev, address);
      usbMemoryR_USB1808(udev, (uint8_t *) &table[i][j].slope, sizeof(float));
      address += 4;
      usbMemAddressW_USB1808(udev, address);
      usbMemoryR_USB1808(udev, (uint8_t *) &table[i][j].offset, sizeof(float));
      address += 4;
    }
  }
  return;
}

The header file has defined

typedef struct Calibration_AIN_t {
  float slope;
  float offset;
} Calibration_AIN;

where NCHAN_18081 and NGAINS_1808 are constants and udev is an integer. I've followed an older question regarding multidimensional arrays and attempted to make a structure like the one in the c code.

_1808 = CDLL(os.path.abspath("lib1808.so"))

NCHAN_1808 = 8  # max number of A/D channels in the device
NGAINS_1808 = 4  # max number of gain levels

class Calibration_AIN(Structure):
    _fields_ = [("slope", c_float), ("offset", c_float)]

class AINarray(Structure):
    _fields_ = [("array", (Calibration_AIN() * NCHAN_1808) * NGAINS_1808)]

table_AIN = AINarray()

_1808.usbBuildGainTableAI_USB1808(udev, table_AIN)

But there's a couple problems with this: the custom type Calibration_AIN can't be populated in an array with the operator * like an int or float can, and I can't pass a custom type through to c. I've also tried making the array with a list of lists in Python, but I can't convert that to anything useful to c passable through ctypes.

How can I call this function from python without modifying the c code? Any help would be much appreciated, and also let me know if I should just learn c and try to write my program in c or Cython. Ctypes may not be the best way of doing this.

CristiFati · Accepted Answer · 2018-11-29 23:01:28Z

0

The line:

_fields_ = [("array", (Calibration_AIN() * NCHAN_1808) * NGAINS_1808)]

should raise a TypeError, since you instantiate Calibration_AIN. I believe that what you meant was:

_fields_ = [("array", (Calibration_AIN * NCHAN_1808) * NGAINS_1808)]

But even so, you don't need the AINarray wrapper. According to [Python]: Arrays, you could do something like:

Calibration_AIN_Table = (Calibration_AIN * NCHAN_1808) * NGAINS_1808

and then, in order to initialize an instance do smth like:

>>> cat = Calibration_AIN_Table()
>>> for i in range(NGAINS_1808):
...     for j in range(NCHAN_1808):
...         cat[i][j].slope = i * j
...         cat[i][j].offset = i * i * j
...
>>> cat[2][3].slope
6.0

edited Nov 29, 2018 at 23:01

answered Apr 25, 2018 at 10:46

CristiFati

41.5k9 gold badges68 silver badges115 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

mgpanas Over a year ago

Thanks for your reply. When I do this I get an error on the line cat[i][j].slope = i * j TypeError: _cytpes.PyCArrayType object does not support indexing

CristiFati Over a year ago

cat = Calibration_AIN_Table(), cat should be an instance note the () after Calibration_AIN_Table.

Collectives™ on Stack Overflow

Python ctypes: calling a function with custom types in c

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related