Raise the clustering to 100 groups

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

tools/src/pop_fem_audit_tools/commands/cluster_keywords.py

@@ -40,7 +40,6 @@ from typing import Any
 from ..utils import format_duration
 
 MODEL: str = "sentence-transformers/all-mpnet-base-v2"
-DEFAULT_CLUSTERS: int = 50
 CLUSTER_EXTRA_MESSAGE: str = (
     "cluster-keywords requires the optional \"cluster\""
     " dependency group; install it with"
@@ -101,8 +100,9 @@ def parse_args(argv: list[str] | None) -> argparse.Namespace:
         "--revision", default=None,
         help="the model revision to pin (default: unpinned)")
     parser.add_argument(
-        "--clusters", type=int, default=DEFAULT_CLUSTERS,
-        help=f"the number of clusters (default {DEFAULT_CLUSTERS})")
+        "--clusters", type=int, required=True,
+        help="the number of clusters; required, so that the\n"
+             "group count is stated on every invocation")
     parser.add_argument(
         "--extra-keyword", dest="extra_keywords", action="append",
         default=None,

tools/tests/test_cluster_keywords.py

@@ -478,6 +478,15 @@ class TestClusterKeywords(unittest.TestCase):
             ["a-center", "aaa-extra", "b-middle", "zzz-extra"])
         self.assertEqual(keywords, sorted(keywords))
 
+    def test_missing_clusters_option_rejected(self) -> None:
+        """Test that omitting --clusters fails the run."""
+        with self.assertRaises(SystemExit) as caught, \
+                redirect_stderr(io.StringIO()):
+            cluster_keywords.parse_args(
+                [str(self.__run1), str(self.__run2),
+                 str(self.__output_dir)])
+        self.assertNotEqual(caught.exception.code, 0)
+
     def test_duplicate_extra_keyword_rejected(self) -> None:
         """Test that repeating the same ``--extra-keyword`` value
         fails the run without writing any output file."""