Support exporting JEPG Compressed DICOMs

[Sharpz7]

def convert_images_to_dicom(
    images: Image.Image,
) -> pydicom.Dataset:

    empty_dicom = create_empty_dicom(dicom_type=DicomTypes.SERIES_ANONYMIZED)

    img_byte_list = []
    # re-write the above to use the images directly
    for img in images:
        with BytesIO() as output:
            img.save(output, format="jpeg")
            img_byte_list.append(output.getvalue())

    empty_dicom.PixelData = encapsulate([x for x in img_byte_list])
    empty_dicom["PixelData"].is_undefined_length = True
    empty_dicom.is_implicit_VR = False
    empty_dicom.LossyImageCompression = "01"
    empty_dicom.LossyImageCompressionRatio = 10  # default jpeg
    empty_dicom.LossyImageCompressionMethod = "ISO_10918_1"
    empty_dicom.file_meta.TransferSyntaxUID = "1.2.840.10008.1.2.4.51"

    empty_dicom.NumberOfFrames = len(images)
    empty_dicom.Rows, empty_dicom.Columns = (
        images[0].size[1],
        images[0].size[0],
    )

    empty_dicom.file_meta.TransferSyntaxUID = "1.2.840.10008.1.2.4.51"
    empty_dicom.PlanarConfiguration = 0

    # forces write_like_original=False
    with BytesIO() as output:
        empty_dicom.save_as(output, write_like_original=False)
        output.seek(0)
        empty_dicom = pydicom.dcmread(output)

    return empty_dicom

Make sure that we can open them fine from radstract.

[Sharpz7]

It is now clear we need to be more careful than the above, giving more options for the types of compression we support, and how faithful we are to non-RGB types (like monochrome2)

Support lossless, support lowest lossy dicom-native, Support lowest lossy custom encoder/decoder Support customisable compression lossy dicom-native