Allow empty tunable values to represent the defaults (#868)

# Pull Request

## Title

Allows an empty dictionary (object) to represent the default tunable
values for tunable params.

---

## Description

This should make it easier to maintain configs that loop over a chosen
set of tunable values, where some of the tunable values are the default
values without needing to copy/paste those values from the tunable
params definitions.

- See Also: [discussion in
#855](#855 (comment))
---

## Type of Change

- ✨ New feature

---

## Testing

Extends existing testing infrastructure to check that this works.

---

mlos_bench/mlos_bench/config/schemas/tunables/tunable-values-schema.json

@@ -12,7 +12,6 @@
                     "type": ["string", "number", "boolean"]
                 }
             },
-            "minProperties": 1,
             "not": {
                 "required": ["tunable_values"]
             }

mlos_bench/mlos_bench/tests/config/schemas/tunable-values/test-cases/good/full/bare-tunable-values-with-schema.jsonc

@@ -1,4 +1,6 @@
 {
+    "$schema": "https://raw.githubusercontent.com/microsoft/MLOS/main/mlos_bench/mlos_bench/config/schemas/tunables/tunable-values-schema.json",
+
     "foo": "bar",
     "int": 1,
     "float": 1.1,

mlos_bench/mlos_bench/tests/config/schemas/tunable-values/test-cases/good/full/empty-tunable-values-with-schema.jsonc

@@ -0,0 +1,4 @@
+{
+    "$schema": "https://raw.githubusercontent.com/microsoft/MLOS/main/mlos_bench/mlos_bench/config/schemas/tunables/tunable-values-schema.json"
+    // empty tunable values represents the defaults
+}

mlos_bench/mlos_bench/tests/config/schemas/tunable-values/test-cases/good/partial/empty-tunable-values-without-schema.jsonc

@@ -0,0 +1,3 @@
+{
+    // empty tunable values represents the defaults
+}

mlos_bench/mlos_bench/tests/tunables/tunables_assign_test.py

@@ -28,6 +28,32 @@ def test_tunables_assign_unknown_param(tunable_groups: TunableGroups) -> None:
         )
 
 
+def test_tunables_assign_defaults(tunable_groups: TunableGroups) -> None:
+    """Make sure we can assign the default values using an empty dictionary."""
+    tunable_groups_defaults = tunable_groups.copy()
+    assert tunable_groups.is_defaults()
+    # Assign the default values.
+    # Also reset the _is_updated flag to avoid considering that in the comparison.
+    tunable_groups.assign({}).reset()
+    assert tunable_groups_defaults == tunable_groups
+    assert tunable_groups.is_defaults()
+    new_vm_size = "Standard_B2s"
+    assert tunable_groups["vmSize"] != new_vm_size
+    # Change one value.
+    tunable_groups.assign({"vmSize": new_vm_size}).reset()
+    assert tunable_groups["vmSize"] == new_vm_size
+    # Check that the other values are still defaults.
+    idle_tunable, _ = tunable_groups.get_tunable("idle")
+    assert idle_tunable.is_default()
+    assert tunable_groups_defaults != tunable_groups
+    assert not tunable_groups.is_defaults()
+    # Reassign defaults.
+    tunable_groups.assign({}).reset()
+    assert tunable_groups["vmSize"] != new_vm_size
+    assert tunable_groups.is_defaults()
+    assert tunable_groups_defaults == tunable_groups
+
+
 def test_tunables_assign_categorical(tunable_categorical: Tunable) -> None:
     """Regular assignment for categorical tunable."""
     # Must be one of: {"Standard_B2s", "Standard_B2ms", "Standard_B4ms"}

mlos_bench/mlos_bench/tunables/tunable_groups.py

@@ -4,12 +4,15 @@
 #
 """TunableGroups definition."""
 import copy
+import logging
 from typing import Dict, Generator, Iterable, Mapping, Optional, Tuple, Union
 
 from mlos_bench.config.schemas import ConfigSchema
 from mlos_bench.tunables.covariant_group import CovariantTunableGroup
 from mlos_bench.tunables.tunable import Tunable, TunableValue
 
+_LOG = logging.getLogger(__name__)
+
 
 class TunableGroups:
     """A collection of covariant groups of tunable parameters."""
@@ -346,11 +349,20 @@ def assign(self, param_values: Mapping[str, TunableValue]) -> "TunableGroups":
         param_values : Mapping[str, TunableValue]
             Dictionary mapping Tunable parameter names to new values.
 
+            As a special behavior when the mapping is empty the method will restore
+            the default values rather than no-op.
+            This allows an empty dictionary in json configs to be used to reset the
+            tunables to defaults without having to copy the original values from the
+            tunable_params definition.
+
         Returns
         -------
         self : TunableGroups
             Self-reference for chaining.
         """
+        if not param_values:
+            _LOG.info("Empty tunable values set provided. Resetting all tunables to defaults.")
+            return self.restore_defaults()
         for key, value in param_values.items():
             self[key] = value
         return self