MultiValue kV causes skin map server error

[Ed McDonagh]

Study with multiple kV values per exposure causes server error when detail view opened due to skin map calculations.

Refs #584

Internal Server Error: /openrem/rf/261383/skin_map/
Traceback (most recent call last):
  File "/home/user/veOpenREM/local/lib/python2.7/site-packages/django/core/handlers/base.py", line 132, in get_response
    response = wrapped_callback(request, *callback_args, **callback_kwargs)
  File "/home/user/veOpenREM/local/lib/python2.7/site-packages/django/contrib/auth/decorators.py", line 22, in _wrapped_view
    return view_func(request, *args, **kwargs)
  File "/home/user/bbOpenREM/openrem/remapp/views.py", line 858, in rf_detail_view_skin_map
    make_skin_map(pk)
  File "/home/user/veOpenREM/local/lib/python2.7/site-packages/celery/local.py", line 188, in __call__
    return self._get_current_object()(*a, **kw)
  File "/home/user/veOpenREM/local/lib/python2.7/site-packages/celery/app/task.py", line 420, in __call__
    return self.run(*args, **kwargs)
  File "/home/user/bbOpenREM/openrem/remapp/tools/make_skin_map.py", line 171, in make_skin_map
    kvp = float(irrad.irradeventxraysourcedata_set.get().kvp_set.get().kvp)
  File "/home/user/veOpenREM/local/lib/python2.7/site-packages/django/db/models/manager.py", line 127, in manager_method
    return getattr(self.get_queryset(), name)(*args, **kwargs)
  File "/home/user/veOpenREM/local/lib/python2.7/site-packages/django/db/models/query.py", line 338, in get
    (self.model._meta.object_name, num)
MultipleObjectsReturned: get() returned more than one Kvp -- it returned 22!

[David Platten]

@edmcdonagh, can you send me an RDSR of a study that has this problem, then I'll take a look at it?

[Ed McDonagh]

Thanks David.

[David Platten]

Calculate the median kVp per exposure. Fixes error that otherwise occurs if a single exposure contains multiple kVp values stored for it. However, is calculating the median the best thing to do? References issue #602

→ <<cset 43270ace05b5>>

[David Platten]

I've just realised that calculating the median is a bad idea because only users with PostgreSQL have access to the median function.

I could use the mean or max aggregate function instead.

I think the mean is the way to go.

@edmcdonagh, what do you think?

[David Platten]

Calculate the median kVp per exposure using the numpy median function. Will work for any OpenREM install. Fixes error that otherwise occurs if a single exposure contains multiple kVp values stored for it. However, is calculating the median the best thing to do? References issue #602

→ <<cset 7df793cc1ef0>>

[David Platten]

Updating skin dose map docs to say how we deal with an irradiation event that has multiple kVp values. References issue #602 and #540

→ <<cset 2bb34a054be6>>

[Ed McDonagh]

I think it probably is. This RDSR is the only I have ever seen with per pulse data, and generally of the two that I have been sent the value is mostly the same, except one of the series has a 0 value at the start. If you used mean then that wouldn't be representative. The only other thing you could do is exclude any zeros before taking a mean, but I think median is probably the way to go. We just need to put it in the docs somewhere.

[David Platten]

@edmcdonagh, your comment must have crossed paths with my commit just now. I'm now using numpy's median function, which I think is the best way forward. I've also updated the docs to reflect this.

[Ed McDonagh]

Thanks David. Will you create a PR?

[David Platten]

Updating changes files. References issue #602

→ <<cset 4648f128d9c3>>

[David Platten]

Added a check for np.isnan(kvp), and if true set kvp to None. I think this addresses the comment that Ed made in the pull request. References issue #602

→ <<cset 83379155dcf4>>

[Ed McDonagh]

Merged in issue602multiplekVpError (pull request #162)

Refs #584, #602

Approved-by: Ed McDonagh [e-mail]

→ <<cset bba534d5588a>>

[David Platten]

Now calculating the mean kVp, excluding any None or zero values, for events with multiple kVp values. References issue #602

→ <<cset 03af6df6c564>>

[Ed McDonagh]

Merged in issue602multiplekVpErrorMean (pull request #164)

Now calculating the mean kVp, excluding any None or zero values, for events with multiple kVp values. References issue #602

Approved-by: Ed McDonagh [e-mail]

→ <<cset 86994c6983d4>>