Elekta Iview Files

[Jeremy Donaghue]

I can not seem to open files that are sent from Elekta's IView, and was wondering if anyone can help.

 

just running:

ds=dicom.read_file('Iview.dcm',force=True)

 

gives me:

 

Traceback (most recent call last):

  File "C:\Users\jdonaghue\Documents\Python\DICOM Stuff\src\QuickViewer.py", line 12, in <module>

    ds=dicom.read_file('Iview.dcm',force=True)

  File "C:\Python26\lib\site-packages\dicom\filereader.py", line 472, in read_file

    dataset = read_partial(fp, stop_when, defer_size=defer_size, force=force)

  File "C:\Python26\lib\site-packages\dicom\filereader.py", line 406, in read_partial

    file_meta_dataset = _read_file_meta_info(fileobj)

  File "C:\Python26\lib\site-packages\dicom\filereader.py", line 350, in _read_file_meta_info

    group_length = unpack("<L", fp.read(length))[0] # XXX should prob check (gp, el), VR before read

struct.error: unpack requires a string argument of length 4

 

I understand what that it's finding information that does not match the length, but have no idea how to get around it.  I have attached one the file.

 

Thank you.

 

j

[Darcy Mason]

On Dec 1, 2:13 pm, Jeremy Donaghue <canod...@gmail.com> wrote:
> I can not seem to open files that are sent from Elekta's IView, and was
> wondering if anyone can help.

>...

>     group_length = unpack("<L", fp.read(length))[0] # XXX should prob check
> (gp, el), VR before read
>
> struct.error: unpack requires a string argument of length 4
>
> I understand what that it's finding information that does not match the
> length, but have no idea how to get around it.  I have attached one the
> file.

Thanks for the example file. It didn't jump out at me at first, but
then I noticed that the file meta information does not have the
mandatory group length data element (0002,0000); it goes straight to
(0002,0001). The "XXX" note about checking them was prophetic, though,
as pydicom could have checked the tag and generated a sensible error
message.

Work-around? I don't know off-hand. Pydicom relies on the group length
to read the file meta info. The only other way I can think of at the
moment would be to create a whole new section of the
_read_file_meta_info method for just this case, using read_dataset
with a "stop_when" callback to finish on the first tag that is not
group 2. Then rewind the file pointer back to the start of the group 8
tag to get ready to read the file dataset. But that rewind amount will
depend on the transfer syntax. It's do-able, but a bit messy.

I've added an issue to the issue list.

Darcy

[Darcy Mason]

On Dec 1, 10:28 pm, Darcy Mason <darcyma...@gmail.com> wrote:
...

> then I noticed that the file meta information does not have the
> mandatory group length data element (0002,0000);
>

> Work-around? ... The only other way I can think of at the

> moment would be to create a whole new section of the
> _read_file_meta_info method for just this case, using read_dataset
> with a "stop_when" callback to finish on the first tag that is not
> group 2.

I've implemented that solution in the latest revision [1]. Pydicom can
now read the example iview file.

-Darcy

[1] http://code.google.com/p/pydicom/source/detail?r=349753235209