Extracting Data from a DICOMDIR file using Pydicom

Question

I'm unable to read in a DICOM file as I usually would, citing the error:

AttributeError: 'DicomDir' object has no attribute 'DirectoryRecordSequence'

I've tried:

• pydicom.fileset.FileSet
• using specific tags with dcmread
• pydicom.filereader.read_dicomdir
• pydicom.filereader.read_partial
• using force=True in dcmread

pydicom.filereader.read_file_meta_info is about the only thing that's not returned an error and yields;

(0002, 0000) File Meta Information Group Length  UL: 172
(0002, 0001) File Meta Information Version       OB: b'\x00\x01'
(0002, 0002) Media Storage SOP Class UID         UI: Media Storage Directory Storage
(0002, 0003) Media Storage SOP Instance UID      UI: 2.25.330614241706723499239981063503184149269
(0002, 0010) Transfer Syntax UID                 UI: Explicit VR Little Endian
(0002, 0012) Implementation Class UID            UI: 1.3.6.1.4.1.30071.8
(0002, 0013) Implementation Version Name         SH: 'fo-dicom 4.0.7'

Moreover, the image is supposed to be a regular DICOM file, not a DICOMDIR. I can open the file in ImageJ and view header information there so I know the data is recoverable. Is there a way for me to read in this file in Python or alternatively force it to ignore looking for DirectoryRecordSequence?

Edit: Code and stacktrace from using FileSet:

from pydicom.fileset import FileSet
fs = FileSet("unprocessed.dcm")

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-4-2b6ba2e435fe> in <module>
      1 from pydicom.fileset import FileSet
----> 2 fs = FileSet("unprocessed.dcm")

c:\****\appdata\local\programs\python\python38-32\lib\site-packages\pydicom\fileset.py in __init__(self, ds)
    998         # Check the DICOMDIR dataset and create the record tree
    999         if ds:
-> 1000             self.load(ds)
   1001         else:
   1002             # New File-set

c:\****\appdata\local\programs\python\python38-32\lib\site-packages\pydicom\fileset.py in load(self, ds_or_path, include_orphans, raise_orphans)
   1641             ds = ds_or_path
   1642         else:
-> 1643             ds = dcmread(ds_or_path)
   1644 
   1645         sop_class = ds.file_meta.get("MediaStorageSOPClassUID", None)

c:\****\appdata\local\programs\python\python38-32\lib\site-packages\pydicom\filereader.py in dcmread(fp, defer_size, stop_before_pixels, force, specific_tags)
   1027         stop_when = _at_pixel_data
   1028     try:
-> 1029         dataset = read_partial(
   1030             fp,
   1031             stop_when,

c:\****\appdata\local\programs\python\python38-32\lib\site-packages\pydicom\filereader.py in read_partial(fileobj, stop_when, defer_size, force, specific_tags)
    879             DeprecationWarning
    880         )
--> 881         ds = DicomDir(
    882             fileobj,
    883             dataset,

c:\****\appdata\local\programs\python\python38-32\lib\site-packages\pydicom\dicomdir.py in __init__(self, filename_or_obj, dataset, preamble, file_meta, is_implicit_VR, is_little_endian)
     94 
     95         self.patient_records: List[Dataset] = []
---> 96         self.parse_records()
     97 
     98     def parse_records(self) -> None:

c:\****\appdata\local\programs\python\python38-32\lib\site-packages\pydicom\dicomdir.py in parse_records(self)
    125 
    126         # Build the mapping from file offsets to records
--> 127         records = self.DirectoryRecordSequence
    128         if not records:
    129             return

c:\****\appdata\local\programs\python\python38-32\lib\site-packages\pydicom\dataset.py in __getattr__(self, name)
    834             return {}
    835         # Try the base class attribute getter (fix for issue 332)
--> 836         return object.__getattribute__(self, name)
    837 
    838     @property

AttributeError: 'DicomDir' object has no attribute 'DirectoryRecordSequence'

​

Answer 1

pydicom reads the dataset correctly, but because it identifies as Media Storage Directory it gets processed by the deprecated DicomDir class, even when passed directly to the FileSet class. Because the dataset isn't a valid Media Storage Directory instance this fails, producing the exception seen.

You should be able to fix this by changing the file meta information's (0002,0002) Media Storage SOP Class UID during read:

from pydicom import dcmread
from pydicom import config

def fix_sop_class(elem, **kwargs):
    if elem.tag == 0x00020002:
        # DigitalXRayImageStorageForProcessing
        elem = elem._replace(value=b"1.2.840.10008.5.1.4.1.1.1.1.1")

    return elem

config.data_element_callback = fix_sop_class

ds = dcmread('path/to/file')

By changing the SOP Class UID, that processing is skipped and the dataset returned.