feat: add CLI control for Milvus force-merge size and toggle

tests/test_fts_cli_user_control.py

@@ -80,6 +80,25 @@ def test_cli_applies_bm25_overrides_after_routing_vector_config_to_fts():
     }
 
 
+def test_cli_preserves_milvus_force_merge_options_when_routing_vector_config_to_fts():
+    from vectordb_bench.backend.clients.milvus.config import MilvusIndexConfig
+
+    selected = select_cli_db_case_config(
+        DB.Milvus,
+        MilvusIndexConfig(
+            index=IndexType.HNSW,
+            force_merge_enabled=False,
+            force_merge_target_size_mb=512,
+        ),
+        "FTSBm25Performance",
+        {},
+    )
+
+    assert isinstance(selected, MilvusFtsConfig)
+    assert selected.force_merge_enabled is False
+    assert selected.force_merge_target_size_mb == 512
+
+
 def test_cli_leaves_fts_bm25_defaults_when_options_are_omitted():
     selected = select_cli_db_case_config(
         DB.Vespa,

tests/test_milvus.py

@@ -128,19 +128,42 @@ def test_milvus_rejects_existing_collection_with_incompatible_properties(monkeyp
         _create_milvus_with_collection_properties(monkeypatch, collection_exists=True)
 
 
+def test_milvus_force_merge_config_defaults_and_validation():
+    from vectordb_bench.backend.clients.milvus.config import AutoIndexConfig, MilvusFtsConfig
+
+    assert AutoIndexConfig().force_merge_enabled is True
+    assert AutoIndexConfig().force_merge_target_size_mb is None
+    assert AutoIndexConfig(force_merge_target_size_mb=1024).force_merge_target_size_mb == 1024
+    assert MilvusFtsConfig().force_merge_enabled is True
+    assert MilvusFtsConfig().force_merge_target_size_mb is None
+
+    with pytest.raises(ValueError, match="positive integer"):
+        AutoIndexConfig(force_merge_target_size_mb=0)
+    with pytest.raises(ValueError, match="positive integer"):
+        AutoIndexConfig(force_merge_target_size_mb=-5)
+    with pytest.raises(ValueError, match="positive integer"):
+        MilvusFtsConfig(force_merge_target_size_mb=0)
+
+
 class TestMilvusOptimize:
     def _milvus(
         self,
         *,
         compact_side_effect: Exception | None = None,
         is_fts: bool = False,
         is_gpu_index: bool = False,
+        force_merge_enabled: bool = True,
+        force_merge_target_size_mb: int | None = None,
     ):
         milvus = Milvus.__new__(Milvus)
         milvus.name = "Milvus"
         milvus.collection_name = "test_collection"
         milvus._is_fts = is_fts
-        milvus.case_config = SimpleNamespace(is_gpu_index=is_gpu_index)
+        milvus.case_config = SimpleNamespace(
+            is_gpu_index=is_gpu_index,
+            force_merge_enabled=force_merge_enabled,
+            force_merge_target_size_mb=force_merge_target_size_mb,
+        )
         milvus.client = MagicMock()
         milvus.client.compact.side_effect = compact_side_effect
         milvus.client.compact.return_value = 42
@@ -180,6 +203,24 @@ def test_optimize_flushes_and_runs_normal_compaction_before_force_merge(self):
         assert milvus._wait_for_compaction.call_args_list == [call(41), call(42)]
         milvus.client.refresh_load.assert_called_once_with("test_collection")
 
+    def test_optimize_uses_configured_force_merge_target_size(self):
+        milvus = self._milvus(force_merge_target_size_mb=512)
+
+        milvus._optimize()
+
+        milvus.client.compact.assert_any_call("test_collection", target_size=512)
+        milvus.client.refresh_load.assert_called_once_with("test_collection")
+
+    def test_optimize_skips_force_merge_when_disabled_but_keeps_normal_compaction(self):
+        milvus = self._milvus(force_merge_enabled=False)
+
+        milvus._optimize()
+
+        milvus.client.flush.assert_called_once_with("test_collection")
+        assert milvus.client.compact.call_args_list == [call("test_collection")]
+        assert milvus._wait_for_compaction.call_args_list == [call(42)]
+        milvus.client.refresh_load.assert_called_once_with("test_collection")
+
     def test_optimize_retries_when_compacting_segments_are_missing_from_force_merge_plan(
         self, monkeypatch: pytest.MonkeyPatch
     ):
@@ -353,6 +394,7 @@ def get_compaction_plans(self, _compaction_id: int):
         milvus.name = "Milvus"
         milvus.collection_name = "test_collection"
         milvus._main_index_name = "vector_idx"
+        milvus.case_config = SimpleNamespace()
         milvus.client = FakeMilvusClient()
         monkeypatch.setattr("vectordb_bench.backend.clients.milvus.milvus.time.sleep", lambda _seconds: None)
 

tests/test_milvus_zilliz_cli.py

@@ -57,6 +57,73 @@ def fake_run(**kwargs):
     assert captured["db_case_config"].use_partition_key is True
 
 
+def test_milvus_autoindex_cli_defaults_force_merge_to_current_behavior(monkeypatch: MonkeyPatch) -> None:
+    captured = {}
+
+    def fake_run(**kwargs):
+        captured.update(kwargs)
+
+    monkeypatch.setattr(milvus_cli, "run", fake_run)
+
+    result = CliRunner().invoke(
+        milvus_cli.MilvusAutoIndex,
+        ["--uri", "http://localhost:19530", "--dry-run"],
+    )
+
+    assert result.exit_code == 0, result.output
+    assert captured["db_case_config"].force_merge_enabled is True
+    assert captured["db_case_config"].force_merge_target_size_mb is None
+
+
+def test_milvus_autoindex_cli_accepts_force_merge_flags(monkeypatch: MonkeyPatch) -> None:
+    captured = {}
+
+    def fake_run(**kwargs):
+        captured.update(kwargs)
+
+    monkeypatch.setattr(milvus_cli, "run", fake_run)
+
+    result = CliRunner().invoke(
+        milvus_cli.MilvusAutoIndex,
+        [
+            "--uri",
+            "http://localhost:19530",
+            "--force-merge-target-size-mb",
+            "512",
+            "--no-force-merge-enabled",
+            "--dry-run",
+        ],
+    )
+
+    assert result.exit_code == 0, result.output
+    assert captured["db_case_config"].force_merge_enabled is False
+    assert captured["db_case_config"].force_merge_target_size_mb == 512
+
+
+def test_milvus_fts_cli_accepts_force_merge_flags(monkeypatch: MonkeyPatch) -> None:
+    captured = {}
+
+    def fake_run(**kwargs):
+        captured.update(kwargs)
+
+    monkeypatch.setattr(milvus_cli, "run", fake_run)
+
+    result = CliRunner().invoke(
+        milvus_cli.MilvusFTS,
+        [
+            "--uri",
+            "http://localhost:19530",
+            "--force-merge-target-size-mb",
+            "256",
+            "--dry-run",
+        ],
+    )
+
+    assert result.exit_code == 0, result.output
+    assert captured["db_case_config"].force_merge_enabled is True
+    assert captured["db_case_config"].force_merge_target_size_mb == 256
+
+
 @pytest.mark.parametrize(
     ("level_args", "expected_level"),
     [(["--level", "2"], 2), ([], None)],
@@ -134,6 +201,69 @@ def fake_run(**kwargs):
     assert captured["db_case_config"].use_partition_key is True
 
 
+def test_milvus_autoindex_cli_rejects_non_positive_force_merge_target_size() -> None:
+    result = CliRunner().invoke(
+        milvus_cli.MilvusAutoIndex,
+        ["--uri", "http://localhost:19530", "--force-merge-target-size-mb", "0", "--dry-run"],
+    )
+
+    assert result.exit_code != 0
+    assert "positive integer" in result.output
+
+    result = CliRunner().invoke(
+        milvus_cli.MilvusAutoIndex,
+        ["--uri", "http://localhost:19530", "--force-merge-target-size-mb", "-5", "--dry-run"],
+    )
+
+    assert result.exit_code != 0
+    assert "positive integer" in result.output
+
+
+def test_zilliz_autoindex_cli_rejects_non_positive_force_merge_target_size() -> None:
+    result = CliRunner().invoke(
+        zilliz_cli.ZillizAutoIndex,
+        [
+            "--uri",
+            "https://example.api.gcp-us-west1.zillizcloud.com",
+            "--token",
+            "secret",
+            "--force-merge-target-size-mb",
+            "0",
+            "--dry-run",
+        ],
+    )
+
+    assert result.exit_code != 0
+    assert "positive integer" in result.output
+
+
+def test_zilliz_autoindex_cli_accepts_force_merge_flags(monkeypatch: MonkeyPatch) -> None:
+    captured = {}
+
+    def fake_run(**kwargs):
+        captured.update(kwargs)
+
+    monkeypatch.setattr(zilliz_cli, "run", fake_run)
+
+    result = CliRunner().invoke(
+        zilliz_cli.ZillizAutoIndex,
+        [
+            "--uri",
+            "https://example.api.gcp-us-west1.zillizcloud.com",
+            "--token",
+            "secret",
+            "--force-merge-target-size-mb",
+            "1024",
+            "--no-force-merge-enabled",
+            "--dry-run",
+        ],
+    )
+
+    assert result.exit_code == 0, result.output
+    assert captured["db_case_config"].force_merge_enabled is False
+    assert captured["db_case_config"].force_merge_target_size_mb == 1024
+
+
 def test_milvus_autoindex_cli_nq(monkeypatch: MonkeyPatch) -> None:
     captured = {}