Expose TimeSeriesProperty::valuesAsVector() to the Python interface

[Jose Borreguero]

Exposing std::vector<TYPE> valuesAsVector() const; to the python interface should result in a python method returning a read-only tuple of TYPE objects

[Martyn Gigg]

Expose TimeSeriesProperty valueAsVector as clone nparray Refs #6508

Changeset: b2d514108cafd358a8ef65de290c00f5a299776b

[Martyn Gigg]

Include required for some compilers. Refs #6508

Changeset: 96c6973615dd271227baefd10c13903f7986d285

[Martyn Gigg]

Fix include of CloneToNumpy. Refs #6508

Changeset: cfce7b15cda2728eb95e21a451e6f2ddaefa8658

[Martyn Gigg]

Tester:

• Load a file that contains a time series log, suggestion=LOQ49886.nxs;
• Right-click on the workspace when loaded and click show logs;
• Find a time-series log, suggestion TEMP1

Check in python that the .value property of the log contains the expected number of entries and is a numpy array:

run = ws.getRun()
log = run.getLogData("TEMP1")
values = log.value
print type(values)

[Russell Taylor]

Re #6508. Temporarily remove test that's failing on Windows.

Changeset: 6fbb4b1dec51f477b3e06037a880741a8cfb94ad

[Russell Taylor]

Test failure on Windows needs addressing, and the test re-enabling.

[Martyn Gigg]

Be less strict on element types Refs #6508

Windows dislikes the element type comparison even though it looks the same when you print them out...

Changeset: d2a724a94f812bdd281d905b2bbe02af0e0f9bf0

[Andrei Savici]

You can expose the times as numpy array as well

[Stuart Campbell]

This ticket has been transferred to github issue ​7354