glx2dicom - Convert Sirona GALAXIS (GALILEOS Viewer) CBCT format to DICOM

README.rst

glx2dicom - Convert Sirona GALAXIS (GALILEOS Viewer) CBCT format to DICOM

Usage:

python3 glx2dicom.py --tag PatientName='Surname^Name' \
       --tag PatientBirthDate='19700101' \
       [--tag TAG=VALUE]... SRC_DIR OUT_DIR

This script is not intended for usage in clinics since nobody may assure correctness of conversion result. Perhaps the proper way to get CT from a Sirona device in DICOM format is to request to add an attachment with DICOM files in addition to the proprietary format. I would not be surprised if it is impossible without purchasing an extra option for the device.

I got a CD with cone beam computed tomography (CBCT) results and a proprietary Sirona GALILEOS Viewer (GALAXIS) application for MS Windows. I have not managed to get it running under wine, so I decided to convert it to a more wide spread file format.

The result may be explored using e.g. SightViewer (or dicomXplorer, another application create with SIGHT framework developed by IRCAD) or amide. Some applications do not have 3D features such as aeskulap, but still able to display individual slices.

Requirements: python3 with pydicom, GDCM, and lxml packages.

My guess concerning CT format is the following. It has an ID that is used in file names.

<ID>.gwg

file likely contain metadata such as patient name and birth date, but I have no idea which way it is encoded or encrypted. The file header (signature) is GWG16. It does not look like (possibly compressed) QDataStream. I am unsure what serializers are available in .NET. This script does not use the file, so it is necessary to specify metadata using command line options.

<ID>_vol_0

directory contains data and should be used as the argument of this script.

<ID>_vol_0/<ID>_proj_0

is an XML file with parameters that might be used to create panoramic projection. It seems the image is not saved and the vendor viewer computes it in runtime. Unused by this script.

<ID>_vol_0/<ID>_vol_0

is a gzipped XML file describing device settings and parameters of the saved files. There is a nested element having entity-encoded XML subtree as its content. This script takes some values, however more parameters may be obtained such as Xray tube current and spectrum peak.

<ID>_vol_0/<ID>_vol_0_<NNN>

files having different trailing number. This is CT slices written as gzip compressed file. They are 2 bytes per pixel (int16 or uint16) little endian raw greyscale images. Actually 12 bits are used despite the <ID>_vol_0/<ID>_vol_0 file declares 14 bits. Instead of using this script you may uncompress these files using e.g. zcat(1) and load them as raw images to e.g. ImageJ.

Some additions to the usual disclaimer:

• I never faced DICOM format before.
• This script is tested on a single CT, unsure ether it was produced by Galileos or Orthophos
• Not all mandatory DICOM attributes are set.
• Images are compressed using RLE method. Unsure if it is possible to easily use some lossless JPEG compression from, python however the gdcmconv(1) CLI tool can do it.
• I am completely unsure concerning transfer function (DICOM Look up Table), so "Window" and "Rescale" attributes may be incorrect, see DICOM PS3.3 C.11 Look Up Tables and Presentation States
• Likely it is possible to add the original files with parameters as custom DICOMDIR entries. It will make conversion lossless since the result may be updated later without access to source files. It is not done however.

glx2dicom 0.1

Copyright (C) 2022 Max Nikulin

License GPLv3+: GNU GPL version 3 or later https://gnu.org/licenses/gpl.html.

This is free software: you are free to change and redistribute it.

There is NO WARRANTY, to the extent permitted by law.

glx2dicom.py

#!/usr/bin/env python3
# glx2dicom - Convert Sirona GALILEOS Viewer CBCT format to DICOM
#
# Copyright (C) 2022 Max Nikulin
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.

"""glx2dicom - Convert Sirona GALAXIS (GALILEOS Viewer) CBCT format to DICOM

Usage::

    python3 glx2dicom.py --tag PatientName='Surname^Name' \\
           --tag PatientBirthDate='19700101' \\
           [--tag TAG=VALUE]... SRC_DIR OUT_DIR

This script is not intended for usage in clinics
since nobody may assure correctness of conversion result.
Perhaps the proper way to get CT from a Sirona device in DICOM_ format
is to request to add an attachment with DICOM files
in addition to the proprietary format.
I would not be surprised if it is impossible without purchasing
an extra option for the device.

I got a CD with cone beam computed tomography (CBCT) results
and a proprietary Sirona GALILEOS Viewer (GALAXIS) application
for MS Windows. I have not managed to get it running under wine_,
so I decided to convert it to a more wide spread file format.

The result may be explored using e.g. SightViewer_ (or dicomXplorer_,
another application create with `SIGHT framework`_ developed by IRCAD)
or amide_. Some applications do not have 3D features such as
aeskulap_, but still able to display individual slices.

Requirements: python3_ with pydicom_, GDCM_, and lxml_ packages.

My guess concerning CT format is the following. It has an ID
that is used in file names.

``<ID>.gwg``
    file likely contain metadata such as patient name
    and birth date, but I have no idea which way it is encoded
    or encrypted. The file header (signature) is ``GWG16``.
    It does not look like (possibly compressed) ``QDataStream``.
    I am unsure what serializers are available in .NET.
    This script does not use the file, so it is necessary
    to specify metadata using command line options.
``<ID>_vol_0``
    directory contains data and should be used
    as the argument of this script.
``<ID>_vol_0/<ID>_proj_0``
    is an XML file with parameters that
    might be used to create panoramic projection.
    It seems the image is not saved and the vendor viewer
    computes it in runtime. Unused by this script.
``<ID>_vol_0/<ID>_vol_0``
    is a gzipped XML file describing
    device settings and parameters of the saved files.
    There is a nested element having entity-encoded XML
    subtree as its content. This script takes some values,
    however more parameters may be obtained such as Xray
    tube current and spectrum peak.
``<ID>_vol_0/<ID>_vol_0_<NNN>``
    files having different trailing number. This is CT slices written
    as gzip compressed file. They are 2 bytes per pixel (int16 or uint16)
    little endian raw greyscale images. Actually 12 bits are used
    despite the ``<ID>_vol_0/<ID>_vol_0`` file declares 14 bits.
    Instead of using this script you may uncompress these files
    using e.g. ``zcat(1)`` and load them as raw images to e.g.
    ImageJ_.

Some additions to the usual disclaimer:

- I never faced DICOM format before.
- This script is tested on a single CT,
  unsure ether it was produced by Galileos or Orthophos
- Not all mandatory DICOM attributes are set.
- Images are compressed using RLE method. Unsure if it is possible
  to easily use some lossless JPEG compression from, python however
  the |gdcmconv|_ CLI tool can do it.
- I am completely unsure concerning transfer function (DICOM Look up Table),
  so "Window" and "Rescale" attributes may be incorrect,
  see |dicom ps3.3 C.11|_
- Likely it is possible to add the original files with parameters
  as custom DICOMDIR entries. It will make conversion lossless
  since the result may be updated later without access to source files.
  It is not done however.

.. _aeskulap: https://sight.pages.ircad.fr/sight/
.. _amide: https://github.com/ferdymercury/amide
.. _DICOM: https://www.dicomstandard.org/current
.. |dicom ps3.3 C.11| replace:: DICOM PS3.3 C.11
   Look Up Tables and Presentation States
.. _`dicom ps3.3 C.11`: https://dicom.nema.org/medical/dicom/current/
   output/chtml/part03/sect_C.11.html
.. _dicomXplorer: https://sight.pages.ircad.fr/sight-doc/
   Introduction/src/applications.html#dicomxplorer
.. _GDCM: https://sourceforge.net/projects/gdcm/
.. |gdcmconv| replace:: ``gdcmconv(1)``
.. _gdcmconv: https://gdcm.sourceforge.net/html/gdcmconv.html
.. _ImageJ: https://github.com/imagej/ImageJ
.. _`lxml - XML and HTML with Python`:
.. _lxml: https://lxml.de/
.. _pydicom: https://pydicom.github.io/pydicom/
.. _python3: https://www.python.org/
.. _`SIGHT framework`: https://github.com/IRCAD/sight/
.. _SightViewer: https://www.ircad.fr/research/software/
.. _wine: https://www.winehq.org/
"""

__title__ = "glx2dicom"
__version__ = "0.1"
__author__ = "Max Nikulin"
__email__ = "[email protected]"
__description__ = \
    """Convert Sirona GALAXIS (GALILEOS Viewer) CBCT format to DICOM"""
__license__ = "GPLv3+"
__copyright__ = "Copyright (C) 2022 Max Nikulin"

import datetime
import gzip
from lxml import etree
# import numpy
from pathlib import Path
from pydicom import uid, datadict
from pydicom.dataset import FileDataset
from pydicom.dataset import FileMetaDataset
from pydicom.fileset import FileSet
import re
from typing import Any, Dict, List


usage = """%(prog)s --tag PatientName='Surname^Name' \\
           --tag PatientBirthDate='19700101' \\
           [--tag TAG=VALUE]... SRC_DIR OUT_DIR
   or: %(prog)s [--help | --version]
"""
version = f"""%(prog)s {__version__}
{__copyright__}
License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>.
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
"""

default_dicom_attrs = {
    # Random UUID obtained using ``uid.generate_uid(None)``
    'ImplementationClassUID': '2.25.88075347621178512596802224299896711910',
    'SpecificCharacterSet': 'ISO_IR 100',
    'TransferSyntaxUID': uid.ExplicitVRLittleEndian,
    # PS3.3 C.7.1 patient
    # No UID
    # Likely encoded in ``.gwg` files
    'PatientName': 'Test^Firstname',
    'PatientID': '1234456789',
    # TODO (0018, 5100) PatientPosition
    # PS3.3 C.7.2 study
    'StudyDescription': 'Dental CBCT Study',
    # Physisian
    # PS3.3 C.7.3 series
    'Modality': 'CT',
    'SeriesNumber': "1",
    'SeriesDescription': "CT series",
    # Unsure, perhaps "ORIGINAL" would be for images from which
    # CT is reconstructed.
    'ImageType': ['DERIVED', 'PRIMARY'],
    'NumberOfFrames': 1,
    # PS3.3 C.7.6 Image
    # TODO 'PatientOrientation'
    # PS3.3 C.7.6.2, 10.7
    # Unsure, but do not see obviously wrong orientation.
    'ImageOrientationPatient': [1.0, 0.0, 0.0, 0.0, 1.0, 0.0],
    # PS3.3 C7.6.3, C7.6.3.3
    'SamplesPerPixel': 1,
    'PhotometricInterpretation': 'MONOCHROME2',  # 2 bytes per pixel greyscale
    # SightViewer-21.1.1 can not handle 0 (unsigned) 16 bit images
    'PixelRepresentation': 1,  # 2's component (unsigned)
    'BitsAllocated': 16  # multiple of 8
}

uid_prefix = None
"""UID prefix used to generate e.g. image identifiers

``None``
    UUID (default since I have not registered prefix),
``""``
    prefix registered for ``pydicom``.
"""


def read_fbp_params(ct_dir: Path) -> Dict[str, Any]:
    with gzip.open(ct_dir / ct_dir.name, "rb") as f:
        tree = etree.fromstring(f.read())

    # Other <LibParams> children:
    # - TimeStamp: no seconds
    # - SequenceID
    # - CrossHairX, Y, Z.
    assert tree.tag == "FBPParams"
    params = tree.find("LibParams")
    assert params is not None
    assert params.find("ZoomFactor").text == "1"
    # <DEVICEPLUGIN> (parsed from <ScanParamsXml> text)
    # attributes that might be useful:
    # - deviceid
    # - dx89serialsno
    # - p2kserialno
    # - kv
    # - ma
    # - xraypulsewidth
    scan = etree.fromstring(params.find("ScanParamsXml").text)
    assert scan is not None
    assert scan.tag == "DEVICEPLUGIN"
    # print(etree.tostring(scan, pretty_print=True, encoding="unicode"))
    image = scan.find('IMAGEDESCRIPTION/IMAGE')
    assert image is not None

    img_date = image.attrib['takedate']
    img_time = image.attrib['taketime']
    assert re.match("[0-9]{4}-[0-9]{2}-[0-9]{2}", img_date)
    assert re.match("[0-9]{2}:[0-9]{2}:[0-9]{2}", img_time)
    img_date = re.sub("-", "", img_date)
    img_time = re.sub(":", "", img_time)

    retval = {}

    # PS3.3 C.7.6 Image
    retval['ContentDate'] = img_date
    retval['ContentTime'] = img_time

    # PS3.3 C.7.6.3 Image Pixel Module
    # PS3.3 C.7.6.3.3 Image Pixel Description Macro
    rows = int(params.find('VolSizeX').text)
    assert rows > 0
    columns = int(params.find('VolSizeY').text)
    assert columns > 0
    retval['Rows'] = rows
    retval['Columns'] = columns
    # 14, but really 12
    bits = int(params.find('ImageInBitDepth').text)
    assert bits > 8 and bits <= 16
    value_max = int(params.find('VoxelValueMax').text)
    assert value_max < (2 << bits)
    retval['BitsStored'] = bits
    # PS3.3 C.7.6.3.3
    # and PS3.5 8.1.1 Pixel Data Encoding of Related Data Elements
    retval['HighBit'] = bits - 1

    # PS3.3 C.7.6.2, 10.7
    retval['PixelSpacing'] = [
        float(params.find('VoxelSizeX').text),
        float(params.find('VoxelSizeY').text)]
    thickness = float(params.find('VoxelSizeZ').text)
    assert thickness > 0.0
    retval['SliceThickness'] = thickness
    retval['SpacingBetweenSlices'] = thickness
    z = float(params.find('CenterZ').text)
    retval['ImagePositionPatient'] = [
        float(params.find('CenterX').text)
        - 0.5*(columns - 1)*retval['PixelSpacing'][0],
        float(params.find('CenterY').text)
        - 0.5*(rows - 1)*retval['PixelSpacing'][1],
        z]
    retval['SliceLocation'] = z  # likely unused by applications

    retval['SeriesDate'] = img_date
    retval['SeriesTime'] = img_time

    scan_id = params.find("ScanID").text
    assert scan_id is not None
    # PS3.3 C.7.2 study
    # Unsure, perhaps a scan may contain more than single volume
    retval['StudyID'] = scan_id
    retval['StudyDate'] = img_date
    retval['StudyTime'] = img_time

    retval['FileSetID'] = scan_id

    # PS3.3 C.11.2.1.2.1 Note 4
    # Look Up Tables and Presentation States
    # > VOI LUT Module
    # > Window Center and Window Width
    # > Default LINEAR Function
    #
    # Added with hope that sight viewer will use it to construct
    # default transfer function instead of displaying white brick at startup
    x1 = int(params.find('VoxelValueMin').text)
    assert x1 >= 0
    x2 = int(params.find('VoxelValueMax').text)
    assert x2 > x1
    retval['WindowCenter'] = (x1 + x2 + 1)//2
    retval['WindowWidth'] = x2 - x1 + 1
    # FIXME I am completely unsure in the following values,
    # they are taken from a DICOM file sample,
    # but they makes SightViewer behavior better.
    # Otherwise it displays a white brick when a CT is loaded
    # and it is necessary to adjust transfer function.
    retval['RescaleIntercept'] = -1000
    retval['RescaleType'] = 'HU'
    retval['RescaleSlope'] = 1.0

    return retval


def glx2dicom(srcdir: Path, dstdir: Path, dicom_attrs) -> None:
    # It seems trailing "/" does not matter.
    attrs = dicom_attrs.copy()

    media_storage_sop_class_uid = attrs.pop('MediaStorageSOPClassUID', None)
    sop_class_uid = attrs.pop('SOPClassUID', media_storage_sop_class_uid)
    if not sop_class_uid:
        sop_class_uid = uid.CTImageStorage
    elif (
            media_storage_sop_class_uid and
            sop_class_uid != media_storage_sop_class_uid):
        raise ValueError(
            'SOPClassUID != MediaStorageSOPClassUID',
            (sop_class_uid, media_storage_sop_class_uid))

    file_meta = FileMetaDataset()
    file_meta.MediaStorageSOPClassUID = sop_class_uid
    ds = FileDataset(
        None, {},
        file_meta=file_meta,
        is_implicit_VR=False,
        is_little_endian=True)
    # Not necessary in the case of
    #     ds.save_as(ds.filename, write_like_original=False)
    # It seems ``FileSet.write()`` ensures proper format as well.
    # ds.preamble = b'\0'*128

    # PS3.3 C.12
    # C.12.1.1.1.1 should be equal to File Meta Information header
    ds.SOPClassUID = file_meta.MediaStorageSOPClassUID

    # ``FileSet`` attribute, image values are generated using UID prefix
    media_storage_sop_instace_uid = attrs.pop(
        'MediaStorageSOPInstanceUID', None)
    sop_instance_uid = attrs.pop(
        'SOPInstanceUID', media_storage_sop_instace_uid)
    if not sop_instance_uid:
        sop_instance_uid = uid.generate_uid(uid_prefix)
    elif (
            media_storage_sop_instace_uid and
            sop_instance_uid != media_storage_sop_instace_uid):
        raise ValueError(
            'SOPInstanceUID != MediaStorageSOPInstanceUID',
            (sop_instance_uid, media_storage_sop_instace_uid))

    # :dcm:`part03/sect_C.7.4.html` PS3.3 C.7.4 Frame of Reference Module
    # Identifies frame of reference withing a series.
    for tag in [
            'StudyInstanceUID', 'SeriesInstanceUID', 'FrameOfReferenceUID']:
        value = attrs.pop(tag, None) or uid.generate_uid(uid_prefix)
        setattr(ds, tag, value)

    fs = FileSet()
    fs.MediaStorageSOPInstanceUID = sop_instance_uid

    # PS3.3 C.12
    instance_creation_date = attrs.pop('InstanceCreationDate', '')
    instance_creation_time = attrs.pop('InstanceCreationTime', '')
    if (not instance_creation_date) and (not instance_creation_date):
        dt = datetime.datetime.now()
        instance_creation_date = dt.strftime("%Y%m%d")
        instance_creation_time = dt.strftime("%H%M%S.%f")
    ds.InstanceCreationDate = instance_creation_date
    ds.InstanceCreationTime = instance_creation_time

    for k, v in attrs.items():
        tag1 = (datadict.tag_for_keyword(k) & 0xffff0000) >> 16
        obj = ds
        if tag1 == 0x0002:
            obj = file_meta
        elif tag1 == 0x0004:
            obj = fs
        setattr(obj, k, v)

    transfer_syntax_uid = attrs['TransferSyntaxUID']
    photometric_interpretation = attrs['PhotometricInterpretation']
    slice_spacing = attrs['SpacingBetweenSlices']
    z_center = attrs['ImagePositionPatient'][2]
    pixel_bytes, bits_allocated_remnant = divmod(attrs['BitsAllocated'], 8)
    assert bits_allocated_remnant == 0
    data_length = attrs['Rows']*attrs['Columns']*pixel_bytes

    src = sorted(src_dir.glob(src_dir.name + '_[0-9]*[0-9]'))
    if not len(src) > 0:
        raise RuntimeError(f'No files found in {src_dir:r}', src_dir)
    z_base = z_center - 0.5*slice_spacing*(len(src) + 1)

    i = 1
    for f in src:
        print(f'{i:3d} {f}')

        # It seems compressing image or adding dataset to fileset
        # resets some attributes
        ds.PhotometricInterpretation = photometric_interpretation
        file_meta.TransferSyntaxUID = transfer_syntax_uid
        file_meta.MediaStorageSOPInstanceUID = uid.generate_uid(uid_prefix)
        # PS3.3 C.12
        # C12.1.1.1.1 should be equal to File Meta Information header
        ds.SOPInstanceUID = file_meta.MediaStorageSOPInstanceUID
        # PS3.3 C.7.6 Image
        ds.InstanceNumber = str(i)  # Earlier "Image Number"
        z = z_base + i*slice_spacing
        ds.ImagePositionPatient[2] = z
        ds.SliceLocation = z  # likely unused by applications
        with gzip.open(f, "rb") as img:
            data = img.read()
        assert len(data) == data_length
        ds.PixelData = data
        # Attempt to set ``SmallestImagePixelValue``
        # and ``LargestImagePixelValue`` causes
        #     TypeError: object of type 'numpy.int16' has no len()
        # on writing DICOM file. It does not matter if ``pixel_array``
        # is obtained from ``ds`` or created explicitly.
        # pixel_type = numpy.dtype(numpy.int16)
        # pixel_type.newbyteorder('<')
        # pixel_array = numpy.frombuffer(data, pixel_type)
        # ds.SmallestImagePixelValue = pixel_array.min()
        # ds.LargestImagePixelValue = pixel_array.max()
        # ds.PixelData = pixel_array.tobytes()
        # Docs mention 'pylibjpeg' plugin as well.
        ds.compress(uid.RLELossless, encoding_plugin='gdcm')
        # Adds to a temporary directory at first.
        # It seems there is no option to control it.
        fs.add(ds)
        # TODO Should 'Rows' and 'Columns' attributes be added
        # to each image directory record?
        i += 1
    fs.write(dstdir)


def create_argument_parser():
    from argparse import ArgumentParser, RawDescriptionHelpFormatter
    parser = ArgumentParser(
        usage=usage,
        description=__description__,
        # Do not wrap --version
        formatter_class=RawDescriptionHelpFormatter)
    parser.add_argument("-V", "--version", action="version", version=version)
    parser.add_argument(
        "--tag", metavar="DICOM_TAG=VALUE",
        default=[], action="append",
        help="set or override DICOM tag")
    parser.add_argument(
        "src_dir", metavar="SRC_DIR",
        help='Directory with source CT images'
        ', "ID_vol_0" adjucent to the "ID.gwg" file')
    parser.add_argument(
        "dst_dir", metavar="DST_DIR",
        help='Directory to write the result DICOMDIR and images')
    return parser


def cli_tags2dict(tags: List[str]):
    retval = {}
    errors = []
    for kv in tags:
        k, v = kv.split("=", 2)
        try:
            datadict.dictionary_VR(k)
            # TODO convert integer values
            retval[k] = v
        except (KeyError, ValueError):
            errors.append(f'Unknown DICOM tag: {kv!r}')
    return retval, errors


def readme():
    """Print description intended for ``README.rst`` file.

    - Merge the module docstring with disclaimer and license statement.
    - Check metadata consistency.

    Usage::
        python3 -c 'from glx2dicom import readme; print(readme())'
    """
    title, usage_par, usage_cmd, body = __doc__.split(sep='\n\n', maxsplit=3)
    assert title == f'{__title__} - {__description__}'
    use = re.sub(r'\n.*--version.*\n', '', usage)
    assert re.sub(r'^ *', '', usage_cmd, flags=re.MULTILINE) == re.sub(
        r'^ *', '',
        use % dict(prog=f'python3 {__title__}.py'),
        flags=re.MULTILINE)
    title_bar = re.sub('.', '=', title)
    print('\n'.join([
        title_bar, title, title_bar, '', usage_par,
        '', usage_cmd, '', body, '']),
        end='')
    print(
        re.sub('<([^>]+)>', '\\1', version).strip()
        .replace('\n', '\n\n') % dict(prog=__title__))


if __name__ == '__main__':
    parser = create_argument_parser()
    args = parser.parse_args()
    src_dir = Path(args.src_dir)
    # ``collections.ChainMap(dicom_attrs, default_dicom_attrs)``
    # would not work because ``pop(key)`` removes element
    # from first dict only and next time the ``key`` may be obtained
    # from second dict.
    dicom_attrs = default_dicom_attrs.copy()
    glx_attrs = read_fbp_params(src_dir)
    dicom_attrs.update(glx_attrs)

    cli_attrs, errors = cli_tags2dict(args.tag)
    if errors:
        # "%(prog)s" is not supported here
        parser.error("\n".join(errors))
    dicom_attrs.update(cli_attrs)

    # TODO --verbose option
    # from pprint import pprint
    # pprint(dicom_attrs)

    glx2dicom(src_dir, Path(args.dst_dir), dicom_attrs)

If you get errors with gdcm first 'pip uninstall gdcm'…

You may comment out the line ds.compress(uid.RLELossless, encoding_plugin='gdcm'), if you can afford larger size of the conversion result. It should be possible to use another tool to compress DICOM image files.

Sorry, for me, to pull all dependencies, it was enough to type (Debian or Ubuntu Linux)

sudo apt install python3-pydicom python3-lxml python3-gdcm

ensar@Ensars-MacBook-Pro ~ % python3 /Users/ensar/Downloads/bd9f630466e1cac6aea162279ce5e15e-728495d5292bd21b3ecfc84be536f6f26dfa9fd2/glx2dicom.py --tag PatientName=‘Tahic^Ensar’
--tag PatientBirthDate=‘20021112’
/Users/ensar/Downloads/ZUBI/f790c03f73af4765aaf000348776925af790c03f73af_vol_0 /Users/ensar/Desktop/Test
Traceback (most recent call last):
File "/Users/ensar/Downloads/bd9f630466e1cac6aea162279ce5e15e-728495d5292bd21b3ecfc84be536f6f26dfa9fd2/glx2dicom.py", line 515, in
glx_attrs = read_fbp_params(src_dir)
^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/ensar/Downloads/bd9f630466e1cac6aea162279ce5e15e-728495d5292bd21b3ecfc84be536f6f26dfa9fd2/glx2dicom.py", line 220, in read_fbp_params
scan = etree.fromstring(params.find("ScanParamsXml").text)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "src/lxml/etree.pyx", line 3264, in lxml.etree.fromstring
File "src/lxml/parser.pxi", line 1988, in lxml.etree._parseMemoryDocument
ValueError: can only parse strings

Currently experiencing this error, any way to resolve this?

File "glx2dicom.py", line 220, in read_fbp_params
scan = etree.fromstring(params.find("ScanParamsXml").text)

It seems structure of gzipped XML file (with the same name as directory *_vol0) has changed or developers have fixed a bug when a part of XML is inserted as text with quoting of <>&.

The following dirty hack may be a workaround in such case. The script tries to obtain study date and time from this XML element, so for one time conversion it should be enough to replace the following code (line 220)

    scan = etree.fromstring(params.find("ScanParamsXml").text)
    assert scan is not None
    assert scan.tag == "DEVICEPLUGIN"
    # print(etree.tostring(scan, pretty_print=True, encoding="unicode"))
    image = scan.find('IMAGEDESCRIPTION/IMAGE')
    assert image is not None

    img_date = image.attrib['takedate']
    img_time = image.attrib['taketime']
    assert re.match("[0-9]{4}-[0-9]{2}-[0-9]{2}", img_date)
    assert re.match("[0-9]{2}:[0-9]{2}:[0-9]{2}", img_time)
    img_date = re.sub("-", "", img_date)
    img_time = re.sub(":", "", img_time)

by something like

    img_date = "2024-01-31"
    img_time = "15:45:51" 

Thanks for the work. I got some files from my dentist which is a folder named "ID_vol_0" and a single file "ID.gwg", got error message "FileNotFoundError: [Errno 2] No such file or directory: 'src/src'". So I manually tar & gzip them, then I got following error:

Traceback (most recent call last):
  File "/Users/dm/Downloads/CT/glx2dicom.py", line 515, in <module>
    glx_attrs = read_fbp_params(src_dir)
                ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/dm/Downloads/CT/glx2dicom.py", line 202, in read_fbp_params
    tree = etree.fromstring(f.read())
           ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "src/lxml/etree.pyx", line 3264, in lxml.etree.fromstring
  File "src/lxml/parser.pxi", line 1989, in lxml.etree._parseMemoryDocument
  File "src/lxml/parser.pxi", line 1876, in lxml.etree._parseDoc
  File "src/lxml/parser.pxi", line 1164, in lxml.etree._BaseParser._parseDoc
  File "src/lxml/parser.pxi", line 633, in lxml.etree._ParserContext._handleParseResultDoc
  File "src/lxml/parser.pxi", line 743, in lxml.etree._handleParseResult
  File "src/lxml/parser.pxi", line 672, in lxml.etree._raiseParseError
  File "<string>", line 
lxml.etree.XMLSyntaxError: Start tag expected, '<' not found, line 1, column 1

It's not the expected structure of the SRC_DIR? Any way to solve? Thanks in adv.

Thanks for the work. I got some files from my dentist which is a folder named "ID_vol_0" and a single file "ID.gwg", got error message "FileNotFoundError: [Errno 2] No such file or directory: 'src/src'". So I manually tar & gzip them

If you put ID_vol_0 (with some long ID) into the src folder then specify src/ID_vol_0 (not just src) as the SRC_DIR argument. Do not tar or gzip ID_vol_0 directory. Check that the ID_vol_0 file exists in that folder.

Ask your clinic, perhaps they can provide the same scan in DICOM format.

Thanks for the work. I got some files from my dentist which is a folder named "ID_vol_0" and a single file "ID.gwg", got error message "FileNotFoundError: [Errno 2] No such file or directory: 'src/src'". So I manually tar & gzip them

If you put ID_vol_0 (with some long ID) into the src folder then specify src/ID_vol_0 (not just src) as the SRC_DIR argument. Do not tar or gzip ID_vol_0 directory. Check that the ID_vol_0 file exists in that folder.

Ask your clinic, perhaps they can provide the same scan in DICOM format.

i have '[Errno 2] No such file or directory' error when i enter src/id_vol_0 as SRC_DIR argument. problem is, it just appends last directories name to the argument ('no such directory src/id_vol_0/id_vol_0' when i specified just src/id_vol_0).
edit: ok i got it, it expects file of the same name in the directory, but still that file is present.

it expects file of the same name in the directory, but still that file is present.

Yes, it is layout of files that I have. So use the directory where you put your CT as SRC_DIR.

it expects file of the same name in the directory, but still that file is present.

Yes, it is layout of files that I have. So use the directory where you put your CT as SRC_DIR.

it appends the directories name eg 'CTdirectory' and says no such file or directory "CTdirectory/CTdirectory'.

it appends the directories name eg 'CTdirectory' and says no such file or directory "CTdirectory/CTdirectory'.

I have no idea what is the file layout in your case. I have a CD with a directory having a long ID and _vol_0 in its name. Path to this folder should be specified as the SRC_DIR argument. The directory contains a file having the same name is the directory and the script reads metadata from this file. Check that files and directories in your case are consistent to the structure described above.

it appends the directories name eg 'CTdirectory' and says no such file or directory "CTdirectory/CTdirectory'.

I have no idea what is the file layout in your case. I have a CD with a directory having a long ID and _vol_0 in its name. Path to this folder should be specified as the SRC_DIR argument. The directory contains a file having the same name is the directory and the script reads metadata from this file. Check that files and directories in your case are consistent to the structure described above.

mine's the same: CTdir/id_vol_0/id_vol_0. no matter if i specify CTdir or id_vol_0 as SRC_DIR it wont work

mine's the same: CTdir/id_vol_0/id_vol_0. no matter if i specify CTdir or id_vol_0 as SRC_DIR it wont work

If the *_vol_0 file exists in the directory with the same name then ask a friend or a colleague how to properly specify path to this directory. ls (or dir in the case of Windows) with the directory as its argument should not report an error.

mine's the same: CTdir/id_vol_0/id_vol_0. no matter if i specify CTdir or id_vol_0 as SRC_DIR it wont work

If the *_vol_0 file exists in the directory with the same name then ask a friend or a colleague how to properly specify path to this directory. ls (or dir in the case of Windows) with the directory as its argument should not report an error.

apologies, ran it again and it worked with CTdir/id_vol_0 as SRC_DIR. i applied forementioned "image date 'dirty hack' " because it didnt work before and apparently i didn't pay attention that i was entering wrong src_dir after.

Hello! I have problems with code execution. Help me)

Traceback (most recent call last):
File "/home/rav/Development/mrt/glx2dicom.py", line 528, in
glx2dicom(src_dir, Path(args.dst_dir), dicom_attrs)
File "/home/rav/Development/mrt/glx2dicom.py", line 321, in glx2dicom
sop_class_uid = uid.CTImageStorage
File "/usr/lib/python3/dist-packages/pydicom/uid.py", line 44, in getattr
raise AttributeError(f"module {name} has no attribute {name}")
AttributeError: module pydicom.uid has no attribute CTImageStorage

sop_class_uid = uid.CTImageStorage
File "/usr/lib/python3/dist-packages/pydicom/uid.py", line 44, in getattr
raise AttributeError(f"module {name} has no attribute {name}")
AttributeError: module pydicom.uid has no attribute CTImageStorage

What version of pydicom do you use? It seems, CTImageStorage is still in its place:
https://github.com/pydicom/pydicom/blob/main/src/pydicom/uid.py#L680
What is the result of the following command in your case?

python3 -c "from pydicom import uid; print(uid.CTImageStorage)"

When inserting verbatim text, please, add triple backticks around to suppress treating some characters as markup, see https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/creating-and-highlighting-code-blocks