Phase 2 complete: all five configs + eval harness working end to end

src/eval.py

@@ -0,0 +1,59 @@
+"""
+Runs all five RAG configs against a single query and returns structured
+results for comparison. No scoring/taxonomy logic here yet.
+"""
+
+import sys
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).resolve().parent))
+
+from configs.grounded import run as run_grounded
+from configs.hybrid import run as run_hybrid
+from configs.kg_augmented import run as run_kg
+from configs.naive import run as run_naive
+from configs.rerank import run as run_rerank
+
+CONFIG_RUNNERS = {
+    "naive": run_naive,
+    "hybrid": run_hybrid,
+    "rerank": run_rerank,
+    "kg_augmented": run_kg,
+    "grounded": run_grounded,
+}
+
+
+def evaluate_query(query: str, source_filter: list[str] | None = None) -> dict:
+    results = {}
+
+    for config_name, run_fn in CONFIG_RUNNERS.items():
+        try:
+            results[config_name] = run_fn(query, source_filter=source_filter)
+        except Exception as e:
+            results[config_name] = {"error": str(e), "config_name": config_name}
+
+    return {"query": query, "results": results}
+
+
+if __name__ == "__main__":
+    smoke_query = (
+        "What are the obligations of an APRA-regulated entity under CPS 234 "
+        "regarding information security incidents?"
+    )
+
+    evaluation = evaluate_query(smoke_query)
+
+    for config_name, result in evaluation["results"].items():
+        print(f"=== {config_name} ===")
+
+        if "error" in result:
+            print(f"ERROR: {result['error']}")
+            print("-" * 60)
+            continue
+
+        print(result["answer"][:300])
+        print("\nRetrieved paragraph IDs:")
+        for chunk in result["retrieved_chunks"]:
+            print(f"- [{chunk['source']}] {chunk['paragraph_id']}")
+
+        print("-" * 60)