Record bounded Qwen3.5-2B pilot and stratified diagnostics

tests/test_pilot.py

@@ -107,6 +107,25 @@ class PilotTests(unittest.TestCase):
                 plot_main()
             self.assertTrue(validate(root)['passed'])
             self.assertEqual(len(list((root / 'figures').glob('*.pdf'))), 6)
+            import csv
+            from collections import defaultdict
+            with (root / 'analysis/null_histograms.csv').open() as f:
+                bins = list(csv.DictReader(f))
+            areas = defaultdict(float)
+            counts = defaultdict(int)
+            sizes = {}
+            for row in bins:
+                key = (row['mode'], row['ndim'], row['method'])
+                areas[key] += float(row['density']) * (float(row['bin_right']) - float(row['bin_left']))
+                counts[key] += int(row['count'])
+                sizes[key] = int(row['samples'])
+            self.assertTrue(areas)
+            for key in areas:
+                self.assertAlmostEqual(areas[key], 1.)
+                self.assertEqual(counts[key], sizes[key])
+            with (root / 'analysis/heatmap_pairs.csv').open() as f:
+                cells = list(csv.DictReader(f))
+            self.assertTrue(all(row['verified'] == 'False' for row in cells if row['identity'] == 'True'))
 
 
 if __name__ == '__main__':