*res.metrics stored as ConfigsRecord

src/py/flwr/common/recordset_compat.py

@@ -210,9 +210,9 @@ def recordset_to_fitres(recordset: RecordSet, keep_input: bool) -> FitRes:
     num_examples = cast(
         int, recordset.get_metrics(f"{ins_str}.num_examples")["num_examples"]
     )
-    metrics_record = recordset.get_metrics(f"{ins_str}.metrics")
+    configs_record = recordset.get_configs(f"{ins_str}.metrics")
 
-    metrics = _check_mapping_from_recordscalartype_to_scalar(metrics_record.data)
+    metrics = _check_mapping_from_recordscalartype_to_scalar(configs_record.data)
     status = _extract_status_from_recordset(ins_str, recordset)
 
     return FitRes(
@@ -226,8 +226,8 @@ def fitres_to_recordset(fitres: FitRes, keep_input: bool) -> RecordSet:
 
     res_str = "fitres"
 
-    recordset.set_metrics(
-        name=f"{res_str}.metrics", record=MetricsRecord(fitres.metrics)  # type: ignore
+    recordset.set_configs(
+        name=f"{res_str}.metrics", record=ConfigsRecord(fitres.metrics)  # type: ignore
     )
     recordset.set_metrics(
         name=f"{res_str}.num_examples",
@@ -269,9 +269,9 @@ def recordset_to_evaluateres(recordset: RecordSet) -> EvaluateRes:
     num_examples = cast(
         int, recordset.get_metrics(f"{ins_str}.num_examples")["num_examples"]
     )
-    metrics_record = recordset.get_metrics(f"{ins_str}.metrics")
+    configs_record = recordset.get_configs(f"{ins_str}.metrics")
 
-    metrics = _check_mapping_from_recordscalartype_to_scalar(metrics_record.data)
+    metrics = _check_mapping_from_recordscalartype_to_scalar(configs_record.data)
     status = _extract_status_from_recordset(ins_str, recordset)
 
     return EvaluateRes(
@@ -297,9 +297,9 @@ def evaluateres_to_recordset(evaluateres: EvaluateRes) -> RecordSet:
     )
 
     # metrics
-    recordset.set_metrics(
+    recordset.set_configs(
         name=f"{res_str}.metrics",
-        record=MetricsRecord(evaluateres.metrics),  # type: ignore
+        record=ConfigsRecord(evaluateres.metrics),  # type: ignore
     )
 
     # status

src/py/flwr/common/recordset_compat_test.py

@@ -14,12 +14,10 @@
 # ==============================================================================
 """RecordSet from legacy messages tests."""
 
-from contextlib import nullcontext
 from copy import deepcopy
-from typing import Any, Dict
+from typing import Dict
 
 import numpy as np
-import pytest
 
 from .parameter import ndarrays_to_parameters
 from .recordset_compat import (
@@ -75,9 +73,10 @@ def _get_valid_fitins() -> FitIns:
     return FitIns(parameters=ndarrays_to_parameters(arrays), config={"a": 1.0, "b": 0})
 
 
-def _get_valid_fitres_with_config(metrics: Dict[str, Scalar]) -> FitRes:
+def _get_valid_fitres() -> FitRes:
     """Returnn Valid parameters but potentially invalid config."""
     arrays = get_ndarrays()
+    metrics: Dict[str, Scalar] = {"a": 1.0, "b": 0}
     return FitRes(
         parameters=ndarrays_to_parameters(arrays),
         num_examples=1,
@@ -91,8 +90,9 @@ def _get_valid_evaluateins() -> EvaluateIns:
     return EvaluateIns(parameters=fit_ins.parameters, config=fit_ins.config)
 
 
-def _get_valid_evaluateres_with_config(metrics: Dict[str, Scalar]) -> EvaluateRes:
+def _get_valid_evaluateres() -> EvaluateRes:
     """Return potentially invalid config."""
+    metrics: Dict[str, Scalar] = {"a": 1.0, "b": 0}
     return EvaluateRes(
         num_examples=1,
         loss=0.1,
@@ -149,31 +149,16 @@ def test_fitins_to_recordset_and_back() -> None:
     assert fitins_copy == fitins_
 
 
-@pytest.mark.parametrize(
-    "context, metrics",
-    [
-        (nullcontext(), {"a": 1.0, "b": 0}),
-        (
-            pytest.raises(TypeError),
-            {"a": 1.0, "b": 3, "c": True},
-        ),  # fails due to unsupported type for metricsrecord value
-    ],
-)
-def test_fitres_to_recordset_and_back(context: Any, metrics: Dict[str, Scalar]) -> None:
+def test_fitres_to_recordset_and_back() -> None:
     """Test conversion FitRes --> RecordSet --> FitRes."""
-    fitres = _get_valid_fitres_with_config(metrics)
+    fitres = _get_valid_fitres()
 
     fitres_copy = deepcopy(fitres)
 
-    with context:
-        recordset = fitres_to_recordset(fitres, keep_input=False)
-        fitres_ = recordset_to_fitres(recordset, keep_input=False)
+    recordset = fitres_to_recordset(fitres, keep_input=False)
+    fitres_ = recordset_to_fitres(recordset, keep_input=False)
 
-    # only check if we didn't test for an invalid setting. Only in valid settings
-    # makes sense to evaluate the below, since both functions above have succesfully
-    # being executed.
-    if isinstance(context, nullcontext):
-        assert fitres_copy == fitres_
+    assert fitres_copy == fitres_
 
 
 def test_evaluateins_to_recordset_and_back() -> None:
@@ -189,33 +174,16 @@ def test_evaluateins_to_recordset_and_back() -> None:
     assert evaluateins_copy == evaluateins_
 
 
-@pytest.mark.parametrize(
-    "context, metrics",
-    [
-        (nullcontext(), {"a": 1.0, "b": 0}),
-        (
-            pytest.raises(TypeError),
-            {"a": 1.0, "b": 3, "c": True},
-        ),  # fails due to unsupported type for metricsrecord value
-    ],
-)
-def test_evaluateres_to_recordset_and_back(
-    context: Any, metrics: Dict[str, Scalar]
-) -> None:
+def test_evaluateres_to_recordset_and_back() -> None:
     """Test conversion EvaluateRes --> RecordSet --> EvaluateRes."""
-    evaluateres = _get_valid_evaluateres_with_config(metrics)
+    evaluateres = _get_valid_evaluateres()
 
     evaluateres_copy = deepcopy(evaluateres)
 
-    with context:
-        recordset = evaluateres_to_recordset(evaluateres)
-        evaluateres_ = recordset_to_evaluateres(recordset)
+    recordset = evaluateres_to_recordset(evaluateres)
+    evaluateres_ = recordset_to_evaluateres(recordset)
 
-    # only check if we didn't test for an invalid setting. Only in valid settings
-    # makes sense to evaluate the below, since both functions above have succesfully
-    # being executed.
-    if isinstance(context, nullcontext):
-        assert evaluateres_copy == evaluateres_
+    assert evaluateres_copy == evaluateres_
 
 
 def test_get_properties_ins_to_recordset_and_back() -> None: