docs: fix dead iterations_per_update in wiki (use quick/full_update)

wiki/core/api/analysis_objects.md

@@ -170,7 +170,8 @@ concise API reference. This is the ground the
 ## Visualisations during fit
 
 Analyses also produce diagnostic plots that PyAutoFit writes to `output/.../image/`
-at each update step (controlled by `iterations_per_update` on the search). The
+at each update step (controlled by `iterations_per_quick_update` /
+`iterations_per_full_update` on the search). The
 plot routines live in the same packages as the analyses:
 
 - `PyAutoLens:autolens/imaging/model/visualizer.py`

wiki/core/api/searches.md

@@ -31,7 +31,7 @@ af.Nautilus(
     path_prefix=..., name=..., unique_tag=...,
     n_live=200,                  # 100 (exploration) / 200 (production) / 400+ (complex)
     number_of_cores=4,
-    iterations_per_update=2500,
+    iterations_per_quick_update=2500,  # cadence for the on-the-fly max-likelihood fit
 )
 ```
 

wiki/core/concepts/non_linear_search.md

@@ -92,8 +92,9 @@ and of the SLaM pipeline.
   to produce a unique output directory.
 - **`unique_tag`** — extra identifier to distinguish otherwise-identical runs.
 - **`number_of_cores`** — parallel likelihood evaluations.
-- **`iterations_per_update`** — how often the search writes intermediate samples to
-  disk.
+- **`iterations_per_quick_update`** / **`iterations_per_full_update`** — how often the
+  search writes the on-the-fly max-likelihood fit (quick) vs the full output — samples,
+  visuals, `model.result` (full). There is no `iterations_per_update`.
 
 Per-search knobs (`n_live` for nested, `nwalkers` / `nsteps` for MCMC) are in the
 search-specific entries of [`../api/searches`](../api/searches.md).