Merge pull request #149 from weaviate/jose/improve-tenant-ops

Improve tenants logic.

Author: jfrancoa

.claude/skills/contributing-to-weaviate-cli/references/code-review.md

@@ -94,6 +94,23 @@ Commands must exit with code 1 on error for proper scripting support. Don't just
 
 Every manager method that produces output should accept and use `json_output`.
 
+### 6. Unstripped Comma-Separated Tenant Input
+
+Wrong:
+```python
+tenants_list = tenants.split(",") if tenants else None
+```
+
+Right:
+```python
+tenants_list = (
+    [t.strip() for t in tenants.split(",") if t.strip()] if tenants else None
+)
+```
+
+This prevents invalid tenant names like `" b"` or `""` when users pass `--tenants "a, b,"`.
+Apply the same pattern inside manager methods when splitting a `tenants` string argument.
+
 ## Code Conventions
 
 ### Formatting

.claude/skills/operating-weaviate-cli/SKILL.md

@@ -167,6 +167,7 @@ See [references/data.md](references/data.md) and [references/search.md](referenc
 
 ```bash
 weaviate-cli create tenants --collection Movies --number_tenants 100 --tenant_suffix "Tenant" --state active --json
+weaviate-cli create tenants --collection Movies --tenants "Alice,Bob,Charlie" --state active --json
 weaviate-cli get tenants --collection Movies --json
 weaviate-cli get tenants --collection Movies --tenant_id "Tenant-1" --verbose --json
 weaviate-cli update tenants --collection Movies --state cold --number_tenants 100 --json

.claude/skills/operating-weaviate-cli/references/tenants.md

@@ -3,6 +3,8 @@
 Manage tenants in multi-tenant Weaviate collections.
 
 ## Create Tenants
+
+**Auto-generate N tenants** (named `Tenant-0` through `Tenant-99`):
 ```bash
 weaviate-cli create tenants \
   --collection "Movies" \
@@ -11,7 +13,16 @@ weaviate-cli create tenants \
   --state active \
   --json
 ```
-Creates tenants named `Tenant-0` through `Tenant-99`.
+
+**Create specific tenants by name** (comma-separated):
+```bash
+weaviate-cli create tenants \
+  --collection "Movies" \
+  --tenants "Alice,Bob,Charlie" \
+  --state active \
+  --json
+```
+When `--tenants` is provided it overrides `--tenant_suffix` and `--number_tenants`.
 
 Options: `--tenant_batch_size N` for batched creation.
 

test/unittests/test_managers/test_data_manager.py

@@ -4,6 +4,225 @@
 import weaviate.classes.config as wvc
 
 
+# ---------------------------------------------------------------------------
+# _resolve_tenants_for_ingestion
+# ---------------------------------------------------------------------------
+
+
+def _make_col(name: str = "TestCollection") -> MagicMock:
+    """Return a minimal mock Collection with configurable tenants.get()."""
+    col = MagicMock()
+    col.name = name
+    return col
+
+
+def _make_manager() -> DataManager:
+    return DataManager(MagicMock())
+
+
+class TestResolveTenants:
+    """Unit tests for DataManager._resolve_tenants_for_ingestion."""
+
+    # ------------------------------------------------------------------
+    # Case 1: non-MT collection
+    # ------------------------------------------------------------------
+
+    def test_non_mt_no_options_returns_sentinel(self) -> None:
+        col = _make_col()
+        manager = _make_manager()
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=False,
+            auto_tenant_creation_enabled=False,
+            tenant_suffix="Tenant",
+            auto_tenants=0,
+            tenants_list=None,
+        )
+        assert result == ["None"]
+
+    def test_non_mt_with_tenants_list_raises(self) -> None:
+        col = _make_col()
+        manager = _make_manager()
+        with pytest.raises(Exception, match="does not have multi-tenancy enabled"):
+            manager._resolve_tenants_for_ingestion(
+                col=col,
+                mt_enabled=False,
+                auto_tenant_creation_enabled=False,
+                tenant_suffix="Tenant",
+                auto_tenants=0,
+                tenants_list=["Tenant-0"],
+            )
+
+    def test_non_mt_with_auto_tenants_raises(self) -> None:
+        col = _make_col()
+        manager = _make_manager()
+        with pytest.raises(Exception, match="does not have multi-tenancy enabled"):
+            manager._resolve_tenants_for_ingestion(
+                col=col,
+                mt_enabled=False,
+                auto_tenant_creation_enabled=False,
+                tenant_suffix="Tenant",
+                auto_tenants=3,
+                tenants_list=None,
+            )
+
+    # ------------------------------------------------------------------
+    # Case 2: auto_tenants > 0
+    # ------------------------------------------------------------------
+
+    def test_auto_tenants_without_auto_creation_raises(self) -> None:
+        col = _make_col()
+        manager = _make_manager()
+        with pytest.raises(Exception, match="Auto tenant creation is not enabled"):
+            manager._resolve_tenants_for_ingestion(
+                col=col,
+                mt_enabled=True,
+                auto_tenant_creation_enabled=False,
+                tenant_suffix="Tenant",
+                auto_tenants=3,
+                tenants_list=None,
+            )
+
+    def test_auto_tenants_no_existing_generates_new(self) -> None:
+        col = _make_col()
+        col.tenants.get.return_value = {}
+        manager = _make_manager()
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=True,
+            auto_tenant_creation_enabled=True,
+            tenant_suffix="Tenant",
+            auto_tenants=3,
+            tenants_list=None,
+        )
+        assert result == ["Tenant-0", "Tenant-1", "Tenant-2"]
+
+    def test_auto_tenants_continues_from_highest_index(self) -> None:
+        col = _make_col()
+        col.tenants.get.return_value = {
+            "Tenant-0": MagicMock(),
+            "Tenant-1": MagicMock(),
+        }
+        manager = _make_manager()
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=True,
+            auto_tenant_creation_enabled=True,
+            tenant_suffix="Tenant",
+            auto_tenants=4,
+            tenants_list=None,
+        )
+        # 2 existing + 2 new starting at index 2
+        assert result == ["Tenant-0", "Tenant-1", "Tenant-2", "Tenant-3"]
+
+    def test_auto_tenants_enough_existing_returns_slice(self) -> None:
+        col = _make_col()
+        col.tenants.get.return_value = {
+            "Tenant-0": MagicMock(),
+            "Tenant-1": MagicMock(),
+            "Tenant-2": MagicMock(),
+        }
+        manager = _make_manager()
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=True,
+            auto_tenant_creation_enabled=True,
+            tenant_suffix="Tenant",
+            auto_tenants=2,
+            tenants_list=None,
+        )
+        assert result == ["Tenant-0", "Tenant-1"]
+
+    def test_auto_tenants_non_numeric_suffix_skipped_in_indexing(self) -> None:
+        """Tenants whose suffix is non-numeric are included but don't affect index."""
+        col = _make_col()
+        col.tenants.get.return_value = {
+            "Tenant-abc": MagicMock(),  # non-numeric, skip for index
+        }
+        manager = _make_manager()
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=True,
+            auto_tenant_creation_enabled=True,
+            tenant_suffix="Tenant",
+            auto_tenants=2,
+            tenants_list=None,
+        )
+        # existing_matching has "Tenant-abc", highest_index stays -1 → new ones start at 0
+        assert "Tenant-abc" in result
+        assert "Tenant-0" in result
+        assert len(result) == 2
+
+    # ------------------------------------------------------------------
+    # Case 3: explicit tenants_list
+    # ------------------------------------------------------------------
+
+    def test_explicit_tenants_list_returned_as_is(self) -> None:
+        col = _make_col()
+        manager = _make_manager()
+        tenants = ["Alpha", "Beta", "Gamma"]
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=True,
+            auto_tenant_creation_enabled=False,
+            tenant_suffix="Tenant",
+            auto_tenants=0,
+            tenants_list=tenants,
+        )
+        assert result is tenants
+
+    # ------------------------------------------------------------------
+    # Case 4: no options – fall back to existing tenants with suffix
+    # ------------------------------------------------------------------
+
+    def test_fallback_existing_tenants_with_suffix_returned(self) -> None:
+        col = _make_col()
+        col.tenants.get.return_value = {
+            "Tenant-0": MagicMock(),
+            "Tenant-1": MagicMock(),
+            "Other-0": MagicMock(),  # different suffix, must be excluded
+        }
+        manager = _make_manager()
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=True,
+            auto_tenant_creation_enabled=False,
+            tenant_suffix="Tenant",
+            auto_tenants=0,
+            tenants_list=None,
+        )
+        assert sorted(result) == ["Tenant-0", "Tenant-1"]
+
+    def test_fallback_no_tenants_no_auto_creation_raises(self) -> None:
+        col = _make_col()
+        col.tenants.get.return_value = {}
+        manager = _make_manager()
+        with pytest.raises(Exception, match="No tenants present"):
+            manager._resolve_tenants_for_ingestion(
+                col=col,
+                mt_enabled=True,
+                auto_tenant_creation_enabled=False,
+                tenant_suffix="Tenant",
+                auto_tenants=0,
+                tenants_list=None,
+            )
+
+    def test_fallback_no_tenants_with_auto_creation_returns_empty(self) -> None:
+        """When auto_tenant_creation is on and no tenants exist, returns empty list."""
+        col = _make_col()
+        col.tenants.get.return_value = {}
+        manager = _make_manager()
+        result = manager._resolve_tenants_for_ingestion(
+            col=col,
+            mt_enabled=True,
+            auto_tenant_creation_enabled=True,
+            tenant_suffix="Tenant",
+            auto_tenants=0,
+            tenants_list=None,
+        )
+        assert result == []
+
+
 def test_ingest_data(mock_client):
     manager = DataManager(mock_client)
     mock_collections = MagicMock()