Allow DiagnosticAnalysis with missing opt config metrics (#5159)

Summary:
Pull Request resolved: #5159

Filter DiagnosticAnalysis metrics to only those present in the adapter's optimization config, gracefully handling cases where the adapter cannot be extracted. This prevents errors when the experiment has metrics not included in the optimization configuration.

The commit adds a try/except block in DiagnosticAnalysis that uses extract_relevant_adapter to identify which metrics are actually in the optimization config, filtering out any that aren't. If the adapter can't be extracted, it silently continues with the original metric list.

Reviewed By: mgarrard

Differential Revision: D100103284

fbshipit-source-id: 60071ea87c708cc1b7725e3ccf97af990121e19a

Author: bletham

ax/analysis/diagnostics.py

@@ -11,10 +11,11 @@
 from ax.analysis.analysis import Analysis
 from ax.analysis.graphviz.generation_strategy_graph import GenerationStrategyGraph
 from ax.analysis.plotly.cross_validation import CrossValidationPlot
-from ax.analysis.utils import validate_experiment
+from ax.analysis.utils import extract_relevant_adapter, validate_experiment
 from ax.core.analysis_card import AnalysisCardGroup
 from ax.core.experiment import Experiment
 from ax.core.utils import is_bandit_experiment
+from ax.exceptions.core import UserInputError
 from ax.generation_strategy.generation_strategy import GenerationStrategy
 from pyre_extensions import none_throws, override
 
@@ -84,6 +85,19 @@ def compute(
         is_bandit = generation_strategy and is_bandit_experiment(
             generation_strategy_name=generation_strategy.name
         )
+        try:
+            relevant_adapter = extract_relevant_adapter(
+                experiment=experiment,
+                generation_strategy=generation_strategy,
+                adapter=adapter,
+            )
+            adapter_metric_names = [
+                relevant_adapter._experiment.signature_to_metric[signature].name
+                for signature in relevant_adapter._metric_signatures
+            ]
+            metric_names = [m for m in metric_names if m in adapter_metric_names]
+        except UserInputError:
+            pass
 
         cross_validation_plots = (
             [