fix: stop passing dataset=None to fit_cls when sensitivity Job is complete

Job.perform() in autofit/non_linear/grid/sensitivity/job.py used to
short-circuit `dataset = None` when `is_complete` was True (commit
41095a0, Oct 2024 "skip simulation if job is complete") but still called
base_fit_cls / perturb_fit_cls with that None. Consumer Analysis classes
that unpack `dataset.data` / `dataset.noise_map` in __init__ — including
the user-facing autofit_workspace/scripts/features/sensitivity_mapping.py
tutorial — crash on re-runs with `AttributeError: 'NoneType' object has
no attribute 'data'`.

Always call simulate_cls. The expensive non-linear search is still skipped
on re-runs via Search.fit's load-from-zip path (paths.restore() unzips,
paths.is_complete returns True, result_via_completed_fit loads samples
from disk). Only the typically-cheap simulator cost is no longer
optimized away.

The library's own `test_perform_twice` test in
`test_autofit/non_linear/grid/test_sensitivity/test_functionality.py`
masked the contract issue because the test conftest's `Analysis.__init__`
simply stores `dataset` without unpacking — real workspace code unpacks
it eagerly.

Author: Jammy2211

autofit/non_linear/grid/sensitivity/job.py

@@ -156,13 +156,10 @@ def perform(self) -> JobResult:
         An object comprising the results of the two fits
         """
 
-        if self.is_complete:
-            dataset = None
-        else:
-            dataset = self.simulate_cls(
-                instance=self.simulate_instance,
-                simulate_path=self.paths.image_path.with_name("simulate"),
-            )
+        dataset = self.simulate_cls(
+            instance=self.simulate_instance,
+            simulate_path=self.paths.image_path.with_name("simulate"),
+        )
 
         result = self.base_fit_cls(
             model=self.model,