Properties should return native types in python

[Peter Peterson]

It appears that all of the properties return a string for their value in python. This can be closed when

print type(RunProxy['gd_prtn_chrg'])
<type 'float'>

[Peter Peterson]

A note from Martyn on this

Since Python can only cast to types it knows about, TimeSeriesProperty
will need exporting to Python as well. You're correct in that this
should happen in kernel_exports.cpp. The export will be along the lines of

class_<TimeSeriesProperty<double>,
bases<Property>,boost::noncopyable>('TimeSeriesProperty_dbl', no_init)
   .def( .. and the rest )

The biggest annoyance with things like TimeSeriesProperty is that it is
templated and Python needs to export an actual type meaning that there
would need to be an export for each templated type. What I've done in
the past is define a macro for the export string so that I can easily
define each type without too much hassle.

If you have a look at the validator stuff below you'll see what I mean.
It's a bit annoying I know but I haven't found a good way of dealing
with templates/Python.


For the numpy stuff, the routines are in MantidVecHelper.cpp. They are
currently set up to build a numpy array from a MantidVec.

[Peter Peterson]

(In [7417]) Adding long to the list of vector types exported. Refs #1890.

[Peter Peterson]

(In [7418]) Changed the function references to properties. Also exposed TimeSeriesPropertyDouble and PropertyWithValueDouble. Refs #1890.

[Martyn Gigg]

(In [7433]) Update some reduction scripts in light of API changes. Fixes system tests. Re #1890

[Peter Peterson]

(In [7437]) Adding commented out version of vector properties. Refs #1890.

[Peter Peterson]

(In [7440]) Adding PropertyWithValue of vector. Refs #1890.

[Peter Peterson]

(In [7445]) Exposed ArrayProperty. Fixes #1890.

[Martyn Gigg]

(In [7447]) A minor tidy up after changes to properties for run objects so that the algorithm ones are consistent. This required access to getPointerToProperty on PropertyManager which Python couldn't get at through the TypedValue version. Now the properties from the run object and algorithm should behave in the same manner though. Re #1890

[Janik Zikovsky]

Doubles and strings work, but it doesn't look to me like the TimeSeriesProperty goes all the way:

a=LoadSNSEventNexus(Filename="/home/8oz/data/PG3_1879_event.nxs", OutputWorkspace="PG3_1879")
a=a.workspace()

In [42]: b = a.getRun()

In [43]: chrg = b["gd_prtn_chrg"]

In [44]: print type(chrg.value)
<type 'float'>

In [45]: c = b["proton_charge"]

In [46]: print type(c.value)
<class 'libMantidPythonAPI.cpp_list_dbl'>

c.value can be iterated, but it would be best if it returned a native Python list...

[Nick Draper]

Bulk move of tickets to iteration 27, if your ticket is essential for Iteration 26 then move it back.

[Peter Peterson]

This ticket was closed during Iteration 26 with a small amount of work left over in #2080.

[Martyn Gigg]

Verified this along with ticket #2080 as they both related to the same issue.

Strings, doubles, ints and time series properties now appear as native python types.

[Stuart Campbell]

This ticket has been transferred to github issue ​2737