Merge nucleic/fuzzy-dewy-urchin-hpvd into dev

README.md

@@ -77,6 +77,10 @@ Print the two clean, from-pretrained-base training commands after promotion:
 "$PY" ml/purpose-classifier/rebuild_sol_high.py train-commands
 ```
 
+By default the generated commands use the Python interpreter that invoked
+`rebuild_sol_high.py`, so running this from an activated virtual environment emits that
+environment's concrete Python path. Pass `--python <path>` only to override it.
+
 The purpose-lite command starts from the pinned MiniLM revision. The purpose-deep command
 starts from the pinned ModernBERT base variant; neither command supplies a prior classifier
 checkpoint or continuation flag.

rebuild_sol_high.py

@@ -666,7 +666,11 @@ def build_parser() -> argparse.ArgumentParser:
     commands_parser = subparsers.add_parser(
         "train-commands", help="print clean from-base training commands"
     )
-    commands_parser.add_argument("--python", default="$PY")
+    commands_parser.add_argument(
+        "--python",
+        default=sys.executable,
+        help="Python interpreter to print in each command (default: this interpreter)",
+    )
     return parser
 
 

tests/test_rebuild_sol_high.py

@@ -131,6 +131,16 @@ class RebuildSolHighTests(unittest.TestCase):
         self.assertNotIn("--resume-from", commands)
         self.assertNotIn("--model ", commands)
 
+    def test_training_commands_default_to_invoking_interpreter(self):
+        args = rebuild_sol_high.build_parser().parse_args(["train-commands"])
+
+        self.assertEqual(args.python, sys.executable)
+        self.assertTrue(
+            rebuild_sol_high.train_commands(args.python).startswith(
+                f'"{sys.executable}" ml/purpose-classifier/train.py'
+            )
+        )
+
     def test_complete_stage_promotes_public_and_combined_datasets(self):
         with tempfile.TemporaryDirectory() as directory:
             root = Path(directory)