Function not found

[Naveenagv]

I'm working on extracting the rtstruct co-ordinates of my file using the example in rtstruct.rst but :
from dicom_csv.rtstruct.csv import collect_rtstructs is not working. I have tried looking through the documentation and there is no function defining RTstruct. Please let me know of any ambiguities.

[kurmukovai]

Hi, we temporarily remove this functionality, I am currently working on the updated (more general) version. Till then, you could use something like this to get the contours:

def read_contour_sequence(dataset: pydicom.dataset.Dataset) -> dict:
   
    contour_sequence_dict = dict()
    
    if not hasattr(dataset, 'ContourSequence'):
        raise AttributeError('Dataset does not have Contour Sequence.')
    
    for _slice in dataset.ContourSequence:
        key = _slice.ContourImageSequence[0].ReferencedSOPInstanceUID
        coords = np.array(_slice.ContourData)
        n = len(coords) // 3
        coords = coords.reshape((n, 3))
        if key not in contour_sequence_dict:
            contour_sequence_dict[key] = [coords]
        else:
            contour_sequence_dict[key].append(coords)
            
    return contour_sequence_dict

Usage example:

rtstruct = dcmread('path/to/rtstruct.dcm')
contours = list(rtstruct.ROIContourSequence)
contour_coords = read_contour_sequence(contours[0]) # if you have more than one contour in rtstruct you need to iterate over `contours`

contour_coords is a dictionary with keys being SOPInstanceUID of a reference image and values - coordinates of a contour. You might still need to postprocess them in case your image has non-unit ImageOrientationPatient and non-zero ImagePositionPatient fields.