ct_philips.py extractor fails for Philips Brilliance 16P CT scanner

[David Platten]

For the only example I have from this type of scanner the extractor fails because the Acquisition DateTime is blank for each acquisition in the dose summary object.

[David Platten]

I have fixed this on my local system by altering get_date_time in the dcmdatetime.py file. This routine currently fails if the contents of the tag are blank. A tweak is also required in _ctradiationdose within the ct_philips.py file.

Altering get_date_time to the following fixes the first problem:

def get_date_time(tag,dataset):
    """Get DICOM date time string and return Python date time.

    :param tag:         DICOM keyword, no spaces or plural as per dictionary.
    :type tag:          str.
    :param dataset:     The DICOM dataset containing the tag.
    :type dataset:      dataset
    :returns:           Python date time value
    """
    import datetime
    if (tag in dataset):
        dicomdatetime = getattr(dataset,tag)
        if '+' in dicomdatetime or '-' in dicomdatetime:
            import re
            dicomdatetime = re.split('\+|-',dicomdatetime)[0]
        if '.' in dicomdatetime:
            return datetime.datetime.strptime(dicomdatetime, "%Y%m%d%H%M%S.%f")
        try:
            return datetime.datetime.strptime(dicomdatetime, "%Y%m%d%H%M%S")
        except ValueError:
            return

_ctradiationdose needs to have a try put around the microsecond calculation code:

    else:
        # Come back and set start and end of irradiation after creating the x-ray events
        proj.start_of_xray_irradiation = events.aggregate(Min('date_time_started'))['date_time_started__min']
        try:
            latestlength = int(events.latest('date_time_started').exposure_time * 1000) # in microseconds
            lastevent = events.aggregate(Max('date_time_started'))['date_time_started__max']
            if lastevent and latestlength:
                last = lastevent + timedelta(microseconds=latestlength)
                proj.end_of_xray_irradiation = last
        except TypeError:
            pass
        proj.save()

[David Platten]

Is it a good idea to configure get_date_time to return nothing if it encounters a ValueError? Would it be better to try and catch the error in the calling routine? Any thoughts @edmcdonagh?

[Ed McDonagh]

I think returning nothing probably is the best way to do it. It is consistent with the other get functions.

We'd need to make the same change for the other date_time functions.

[David Platten]

Modifed datetime routines to make them return None if there is a ValueError encountered. References issue #500

→ <<cset 963b0a36c5cf>>

[David Platten]

Added a catch for TypeError to prevent a missing datetime value from preventing import. References issue #500

→ <<cset 43f4fbdd2378>>

[David Platten]

The two commits solve this problem - the Philips Brilliance 16P study with empty Acquisition DateTime imports without a problem. I can also confirm that studies from the Philips Brilliance 64 continue to import correctly.

[David Platten]

Added anonymised Philips Brilliance 16P CT scanner enhanced structured report file. References issue #500

→ <<cset c88e8ea48bd7>>

[Ed McDonagh]

Adding a very basic test to ensure sample Philips SC imports - fails with current develop. Refs #500

→ <<cset b31b72c3ce9e>>

[Ed McDonagh]

Adding ref #500 to changes

→ <<cset 7d1e5e9f136f>>

[Ed McDonagh]

Issue #478 was marked as a duplicate of this issue.