Checkpoint frontend UI/UX overhaul and ingestion embed benchmark work

apps/ai-service/tests/test_manifest.py

@@ -0,0 +1,50 @@
+"""F-05: the service must refuse to start on a corpus/model manifest
+mismatch, not silently search with vectors the collection wasn't built from.
+"""
+import pytest
+
+from rag.manifest import ManifestMismatch, check_manifest, manifest_collection
+
+
+def test_manifest_collection_naming():
+    assert manifest_collection("duocthu_v1") == "duocthu_v1__manifest"
+
+
+def test_matching_manifest_passes():
+    check_manifest(
+        {"model_id": "cohere.embed-v4:0", "dimensions": 1024},
+        "duocthu_v1", "cohere.embed-v4:0", 1024,
+    )  # no raise
+
+
+def test_missing_manifest_refuses():
+    with pytest.raises(ManifestMismatch, match="no corpus manifest"):
+        check_manifest(None, "duocthu_v1", "cohere.embed-v4:0", 1024)
+
+
+def test_wrong_model_id_refuses_even_with_matching_dimensions():
+    """The exact scenario the finding names: two unrelated models can both
+    produce 1024-dim vectors."""
+    with pytest.raises(ManifestMismatch, match="model_id"):
+        check_manifest(
+            {"model_id": "amazon.titan-embed-text-v2:0", "dimensions": 1024},
+            "duocthu_v1", "cohere.embed-v4:0", 1024,
+        )
+
+
+def test_wrong_dimensions_refuses():
+    with pytest.raises(ManifestMismatch, match="dimensions"):
+        check_manifest(
+            {"model_id": "cohere.embed-v4:0", "dimensions": 768},
+            "duocthu_v1", "cohere.embed-v4:0", 1024,
+        )
+
+
+def test_both_mismatched_reports_both():
+    with pytest.raises(ManifestMismatch) as excinfo:
+        check_manifest(
+            {"model_id": "other-model", "dimensions": 768},
+            "duocthu_v1", "cohere.embed-v4:0", 1024,
+        )
+    assert "model_id" in str(excinfo.value)
+    assert "dimensions" in str(excinfo.value)