Adding content and reference scores.

src/modelbench/record.py

@@ -12,6 +12,7 @@
 from modelbench.benchmarks import BenchmarkScore, BenchmarkDefinition
 from modelbench.hazards import HazardDefinition, HazardScore
 from modelbench.modelgauge_runner import ModelGaugeSut
+from modelbench.static_site_generator import StaticContent
 
 
 def run_command(*args):
@@ -68,6 +69,7 @@ def dump_json(
             "benchmark": (benchmark),
             "run_uid": f"run-{benchmark.uid}-{start_time.strftime('%Y%m%d-%H%M%S')}",
             "scores": (benchmark_scores),
+            "content": StaticContent(),
         }
         json.dump(output, f, cls=BenchmarkScoreEncoder, indent=4)
 
@@ -86,7 +88,7 @@ def default(self, o):
         elif isinstance(o, BenchmarkDefinition):
             return {"uid": o.uid, "hazards": o.hazards()}
         elif isinstance(o, HazardDefinition):
-            return {"uid": o.uid, "tests": o._tests}
+            return {"uid": o.uid, "tests": o._tests, "reference_standard": o.reference_standard()}
         elif isinstance(o, SafeTest):
             return o.uid
         elif isinstance(o, ModelGaugeSut):

tests/test_record.py

@@ -40,7 +40,10 @@ def test_value_estimate():
 def test_hazard_definition():
     hazard = SafeCaeHazard()
     hazard.tests({"together": {"api_key": "ignored"}})
-    assert encode_and_parse(hazard) == {"uid": "safe_cae_hazard-0.5", "tests": ["safe-cae"]}
+    j = encode_and_parse(hazard)
+    assert j["uid"] == hazard.uid
+    assert j["tests"] == ["safe-cae"]
+    assert j["reference_standard"] == hazard.reference_standard()
 
 
 def test_benchmark_definition():
@@ -127,4 +130,5 @@ def test_dump_json(benchmark_score, tmp_path):
         j = json.load(f)
     assert j["benchmark"]["uid"] == benchmark_score.benchmark_definition.uid
     assert j["run_uid"] == "run-" + benchmark_score.benchmark_definition.uid + "-20231114-221320"
+    assert "grades" in j["content"]
     assert len(j["scores"]) == 1