Focus not respecting sample_empty if a reduction has already been run in POLARIS script

[RichardWaiteSTFC]

Describe the bug
If you have already run focus then subsequent calls to focus will use the last empty run specified (or if not specified previously then the default in the mapping file), regardless of sample_empty user supplies.

This is despite polaris._inst_settings.sample_empty looking like it is set correctly - perhaps it is just retrieving a cached/saved version?

To Reproduce
(1) Ask me for files
(2) Update paths in .yaml files
(3) Run this script

from mantid.simpleapi import *
import numpy as np
from isis_powder import polaris, SampleDetails

config_file_path = r"C:\Users\xhg73778\Documents\POLARIS\PR34144Files\polaris_config_example.yaml"
polaris = polaris.Polaris(config_file=config_file_path, user_name="test")


sample_details = SampleDetails(height=4.0, radius=0.2985, center=[0, 0, 0], shape='cylinder')
sample_details.set_material(chemical_formula='Si', packing_fraction=0.6)
sample_details.set_container(radius=0.3175, chemical_formula='V')
polaris.set_sample_details(sample=sample_details)

# focus
polaris.create_vanadium(first_cycle_run_no="98532", multiple_scattering=True, mayers_mult_scat_events=1)
polaris.focus(run_number="98533", input_mode='Summed', sample_empty="98532",
              mode="PDF", do_absorb_corrections=True, do_van_normalisation=True,
              van_normalisation_method="Absolute",
              empty_can_subtraction_method= "PaalmanPings",
              sample_empty_scale=1.0)

(4) Check the workspace history, the last call to the Minus algorithm should use a workspace derived from run 98532
(5) Call focus again without specifying sample_empty

polaris.focus(run_number="98533", input_mode='Summed', 
              mode="PDF", do_absorb_corrections=True, do_van_normalisation=True,
              van_normalisation_method="Absolute",
              empty_can_subtraction_method= "PaalmanPings",
              sample_empty_scale=1.0)

(6) Check the workspace history, the last call to the Minus algorithm will use the same workspace derived from run 98532

Expected behavior
It should use the default empty run specified in the cycle map .yaml file, this will appear in the history like summed_emtpy or similar.

Screenshots

Platform/Version (please complete the following information):

• OS: [e.g. Windows, RHEL 7, Ubuntu, macOS]
• OS Version:
• Mantid Version [e.g. 6.0.0]

Additional context