fix research result handoff and synthesis provenance

tests/test_backend_pipeline_repairs.py

@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 import asyncio
+import json
 from uuid import uuid4
 
 import httpx
@@ -18,6 +19,8 @@ from nsct.providers.metrics import ProviderMetrics
 from nsct.providers.searxng import SearXNGProvider
 from nsct.crawler.normalize import NormalizedDocument
 from nsct.api.rest_research import DEPTH_CONFIGS
+from nsct.models.claim import Claim, ClaimType
+from nsct.stages.stage9_synthesis import _build_evidence_package_json
 
 
 def _config(*, searxng_base_url: str | None = "http://searxng:8080") -> AppSettings:
@@ -206,6 +209,59 @@ def test_depth_time_budgets_allow_local_35b_inference() -> None:
     assert DEPTH_CONFIGS["deep"].max_research_duration_seconds == 7_200
 
 
+def test_synthesis_package_keeps_stage5_uuid_claim_provenance() -> None:
+    """Stage 9 must accept the actual, UUID-bearing Stage 5 claim model."""
+    source_id = uuid4()
+    claim = Claim(
+        research_run_id=uuid4(),
+        source_id=source_id,
+        claim_text="Eine überprüfbare Behauptung.",
+        evidence_span="Die belegende Passage.",
+        claim_type=ClaimType.CLAIM,
+        source_url="https://example.org/source",
+    )
+
+    package = json.loads(_build_evidence_package_json([claim.model_dump()], [], "Thema"))
+
+    assert package["evidence"][0]["claim_id"] == str(claim.id)
+    assert package["evidence"][0]["source_id"] == str(source_id)
+    assert package["sources"][0]["source_id"] == str(source_id)
+
+
+def test_orchestrator_synthesis_returns_a_report_for_uuid_claims() -> None:
+    """The production hand-off to Stage 9 must not trigger UUID JSON errors."""
+    async def run() -> None:
+        orchestrator = ResearchOrchestrator(_config(), uuid4(), "Testthema")
+        await orchestrator.start()
+        orchestrator._state_machine = StateMachine(ResearchRunState.COMPARING)
+        orchestrator._claims = [
+            Claim(
+                research_run_id=uuid4(),
+                source_id=uuid4(),
+                claim_text="Eine überprüfbare Behauptung.",
+                evidence_span="Die belegende Passage.",
+                claim_type=ClaimType.CLAIM,
+                source_url="https://example.org/source",
+            )
+        ]
+        provider = Mock()
+        provider.complete = AsyncMock(return_value=json.dumps({
+            "summary": "Neutraler Bericht.",
+            "confident_findings": [],
+            "uncertain_areas": [],
+            "contradictions": [],
+        }))
+        orchestrator._llm_provider = provider
+
+        response = await orchestrator._step_synthesizing()
+
+        assert response["success"] is True
+        assert response["report"]["summary"] == "Neutraler Bericht."
+        provider.complete.assert_awaited_once()
+
+    asyncio.run(run())
+
+
 def test_start_does_not_charge_a_phantom_planner_request() -> None:
     async def run() -> None:
         orchestrator = ResearchOrchestrator(_config(), uuid4(), "test")

tests/test_rest_research.py

@@ -334,7 +334,18 @@ def test_detail_endpoints_serialize_pipeline_uuid_identifiers(client: TestClient
             }
         ],
         evidence_scores=[{"claim_id": claim_id, "research_run_id": uuid4()}],
-        report={"summary": "Zusammenfassung", "findings": [], "methodology": "Methodik"},
+        report={
+            "summary": "Zusammenfassung",
+            "confident_findings": [
+                {
+                    "claim_text": "Eine überprüfbare Behauptung.",
+                    "confidence": 0.9,
+                    "source_id": source_id,
+                }
+            ],
+            "uncertain_areas": [],
+            "methodology": "Methodik",
+        },
     )
 
     sources = client.get(f"/v1/research/{research_id}/sources")
@@ -348,6 +359,8 @@ def test_detail_endpoints_serialize_pipeline_uuid_identifiers(client: TestClient
     assert claims.json()["claims"][0]["source_id"] == str(source_id)
     assert evidence.json()["evidence"][0]["claim_id"] == str(claim_id)
     assert report.json()["methodology"] == "Methodik"
+    assert report.json()["findings"][0]["text"] == "Eine überprüfbare Behauptung."
+    assert report.json()["findings"][0]["source_ids"] == [str(source_id)]
 
 
 # ---------------------------------------------------------------------------