- implement a new task_group filtering decorator 'with_selected_task_…

…group' in Assigner class

- update all the sub-classes that use task_groups to use the decorator
- update fedeval sample workspace to use default assigner, tasks and aggregator
- use of federated-evaluation/aggregator.yaml for FedEval specific workspace example to use round_number as 1
- removed assigner and tasks yaml from defaults/federated-evaluation, superseded by default assigner/tasks
- added additional checks for assigner sub-classes that might not have task_groups
- Addressing review comments
- Updated existing test cases for Assigner sub-classes
- Remove hard-coded setting in assigner for torch_cnn_mnist ws, refer to default as in other Workspaces
- Use aggregator supplied --task_group to override the assinger selected_task_group
- update existing test cases of aggregator cli
- add test cases for the decorator
- rebased 25-Jan.1
- implemented the support of multiple task_group without selection
- defaulting of selected_task group 'percentage' to 1.0 post successful filtering
- updated test cases for multiple task group support
Signed-off-by: Shailesh Pant <shailesh.pant@intel.com>

openfl-workspace/torch_cnn_mnist/plan/plan.yaml

@@ -10,16 +10,8 @@ aggregator:
     rounds_to_train: 2
     write_logs: false
   template: openfl.component.aggregator.Aggregator
-assigner:
-  settings:
-    task_groups:
-    - name: learning
-      percentage: 1.0
-      tasks:
-      - aggregated_model_validation
-      - train
-      - locally_tuned_model_validation
-  template: openfl.component.RandomGroupedAssigner
+assigner :
+  defaults : plan/defaults/assigner.yaml
 collaborator:
   settings:
     db_store_rounds: 1

openfl-workspace/torch_cnn_mnist_fed_eval/plan/plan.yaml

@@ -32,10 +32,12 @@ network :
   defaults : plan/defaults/network.yaml
 
 assigner :
-  defaults : plan/defaults/federated-evaluation/assigner.yaml
-
+  defaults : plan/defaults/assigner.yaml
+  settings :
+    selected_task_group : evaluation
+
 tasks :
-  defaults : plan/defaults/federated-evaluation/tasks_torch.yaml
+  defaults : plan/defaults/tasks_torch.yaml
 
 compression_pipeline :
   defaults : plan/defaults/compression_pipeline.yaml

openfl-workspace/workspace/plan/defaults/aggregator.yaml

@@ -3,3 +3,4 @@ settings :
     db_store_rounds : 2
     persist_checkpoint: True
     persistent_db_path: local_state/tensor.db
+

openfl-workspace/workspace/plan/defaults/assigner.yaml

@@ -7,3 +7,7 @@ settings :
         - aggregated_model_validation
         - train
         - locally_tuned_model_validation
+    - name       : evaluation
+      percentage : 0
+      tasks      :
+        - aggregated_model_validation

openfl/component/__init__.py

@@ -1,6 +1,7 @@
 # Copyright 2020-2024 Intel Corporation
 # SPDX-License-Identifier: Apache-2.0
 
+"""OpenFL Component Module."""
 
 from openfl.component.aggregator.aggregator import Aggregator
 from openfl.component.assigner.assigner import Assigner

openfl/component/aggregator/aggregator.py

@@ -130,15 +130,25 @@ def __init__(
         self.straggler_handling_policy = (
             straggler_handling_policy or CutoffTimeBasedStragglerHandling()
         )
-        self._end_of_round_check_done = [False] * rounds_to_train
-        self.stragglers = []
 
         self.rounds_to_train = rounds_to_train
+        if self.task_group == "evaluation":
+            self.rounds_to_train = 1
+            logger.info(
+                f"task_group is {self.task_group}, setting rounds_to_train = {self.rounds_to_train}"
+            )
+
+        self._end_of_round_check_done = [False] * rounds_to_train
+        self.stragglers = []
 
         # if the collaborator requests a delta, this value is set to true
         self.authorized_cols = authorized_cols
         self.uuid = aggregator_uuid
         self.federation_uuid = federation_uuid
+        # # override the assigner selected_task_group
+        # # FIXME check the case of CustomAssigner as base class Assigner is redefined
+        # # and doesn't have selected_task_group as attribute
+        # assigner.selected_task_group = task_group
         self.assigner = assigner
         self.quit_job_sent_to = []
 

openfl/component/assigner/__init__.py

@@ -1,6 +1,7 @@
 # Copyright 2020-2024 Intel Corporation
 # SPDX-License-Identifier: Apache-2.0
 
+"""OpenFL Assigner Module."""
 
 from openfl.component.assigner.assigner import Assigner
 from openfl.component.assigner.random_grouped_assigner import RandomGroupedAssigner

openfl/component/assigner/assigner.py

@@ -4,6 +4,11 @@
 
 """Assigner module."""
 
+import logging
+from functools import wraps
+
+logger = logging.getLogger(__name__)
+
 
 class Assigner:
     r"""
@@ -35,18 +40,27 @@ class Assigner:
         \* - ``tasks`` argument is taken from ``tasks`` section of FL plan YAML file.
     """
 
-    def __init__(self, tasks, authorized_cols, rounds_to_train, **kwargs):
+    def __init__(
+        self,
+        tasks,
+        authorized_cols,
+        rounds_to_train,
+        selected_task_group: str = None,
+        **kwargs,
+    ):
         """Initializes the Assigner.
 
         Args:
             tasks (list of object): List of tasks to assign.
             authorized_cols (list of str): Collaborators.
             rounds_to_train (int): Number of training rounds.
+            selected_task_group (str, optional): Selected task_group.
             **kwargs: Additional keyword arguments.
         """
         self.tasks = tasks
         self.authorized_cols = authorized_cols
         self.rounds = rounds_to_train
+        self.selected_task_group = selected_task_group
         self.all_tasks_in_groups = []
 
         self.task_group_collaborators = {}
@@ -93,3 +107,50 @@ def get_aggregation_type_for_task(self, task_name):
         if "aggregation_type" not in self.tasks[task_name]:
             return None
         return self.tasks[task_name]["aggregation_type"]
+
+    @classmethod
+    def with_selected_task_group(cls, func):
+        """Decorator to filter task groups based on selected_task_group.
+
+        This decorator should be applied to define_task_assignments() method
+        in Assigner subclasses to handle task_group filtering.
+        """
+
+        @wraps(func)
+        def wrapper(self, *args, **kwargs):
+            # Check if selection of task_group is applicable
+            if hasattr(self, "selected_task_group") and self.selected_task_group is not None:
+                # Verify task_groups exists before attempting filtering
+                if not hasattr(self, "task_groups"):
+                    logger.warning(
+                        "Task group specified for selection but no task_groups found. "
+                        "Skipping filtering. This might be intentional for custom assigners."
+                    )
+                    return func(self, *args, **kwargs)
+
+                assert self.task_groups, "No task_groups defined in assigner."
+
+                # Perform the filtering
+                selected_task_groups = [
+                    group for group in self.task_groups if group["name"] == self.selected_task_group
+                ]
+
+                assert len(selected_task_groups) == 1, (
+                    f"Only one task group with name {self.selected_task_group} should exist"
+                )
+
+                # Since we have filtered to one of the task_groups, we need to ensure that
+                # the selected_task_group percentage compute allocation is defaulted to 1.0
+                current_percentage = selected_task_groups[0]["percentage"]
+                logger.info(
+                    f"`percentage` for task_group {self.selected_task_group} is "
+                    f"{current_percentage}, setting it to 1.0"
+                )
+                selected_task_groups[0]["percentage"] = 1.0
+
+                self.task_groups = selected_task_groups
+
+            # Call the original method
+            return func(self, *args, **kwargs)
+
+        return wrapper

openfl/component/assigner/random_grouped_assigner.py

@@ -6,7 +6,7 @@
 
 import numpy as np
 
-from openfl.component.assigner.assigner import Assigner
+from openfl.component.assigner import Assigner
 
 
 class RandomGroupedAssigner(Assigner):
@@ -33,16 +33,19 @@ class RandomGroupedAssigner(Assigner):
         \* - Plan setting.
     """
 
+    with_selected_task_group = Assigner.with_selected_task_group
+
     def __init__(self, task_groups, **kwargs):
         """Initializes the RandomGroupedAssigner.
 
         Args:
             task_groups (list of object): Task groups to assign.
-            **kwargs: Additional keyword arguments.
+            **kwargs: Additional keyword arguments, including mode.
         """
         self.task_groups = task_groups
         super().__init__(**kwargs)
 
+    @with_selected_task_group
     def define_task_assignments(self):
         """Define task assignments for each round and collaborator.
 

openfl/component/assigner/static_grouped_assigner.py

@@ -32,6 +32,8 @@ class StaticGroupedAssigner(Assigner):
         \* - Plan setting.
     """
 
+    with_selected_task_group = Assigner.with_selected_task_group
+
     def __init__(self, task_groups, **kwargs):
         """Initializes the StaticGroupedAssigner.
 
@@ -42,6 +44,7 @@ def __init__(self, task_groups, **kwargs):
         self.task_groups = task_groups
         super().__init__(**kwargs)
 
+    @with_selected_task_group
     def define_task_assignments(self):
         """Define task assignments for each round and collaborator.
 

openfl/interface/aggregator.py

@@ -70,8 +70,7 @@ def aggregator(context):
 @option(
     "--task_group",
     required=False,
-    default="learning",
-    help="Selected task-group for assignment - defaults to learning",
+    help="Selected task-group for assignment",
 )
 def start_(plan, authorized_cols, task_group):
     """Start the aggregator service.
@@ -94,11 +93,13 @@ def start_(plan, authorized_cols, task_group):
         cols_config_path=Path(authorized_cols).absolute(),
     )
 
-    # Set task_group in aggregator settings
-    if "settings" not in parsed_plan.config["aggregator"]:
-        parsed_plan.config["aggregator"]["settings"] = {}
-    parsed_plan.config["aggregator"]["settings"]["task_group"] = task_group
-    logger.info(f"Setting aggregator to assign: {task_group} task_group")
+    # Set task_group in aggregator and assigner settings if provided
+    if task_group:
+        if "settings" not in parsed_plan.config["aggregator"]:
+            parsed_plan.config["aggregator"]["settings"] = {}
+        parsed_plan.config["aggregator"]["settings"]["task_group"] = task_group
+        parsed_plan.config["assigner"]["settings"]["selected_task_group"] = task_group
+        logger.info(f"Setting aggregator to assign: {task_group} task_group")
 
     logger.info("🧿 Starting the Aggregator Service.")
 

tests/openfl/component/assigner/test_assigner.py

@@ -19,7 +19,7 @@ def assigner():
 
 def test_get_aggregation_type_for_task_none(assigner):
     """Assert that aggregation type of custom task is None."""
-    task_name = 'test_name'
+    task_name = "test_name"
     tasks = {task_name: {}}
 
     assigner = assigner(tasks, None, None)
@@ -31,11 +31,9 @@ def test_get_aggregation_type_for_task_none(assigner):
 
 def test_get_aggregation_type_for_task(assigner):
     """Assert that aggregation type of task is getting correctly."""
-    task_name = 'test_name'
-    test_aggregation_type = 'test_aggregation_type'
-    tasks = {task_name: {
-        'aggregation_type': test_aggregation_type
-    }}
+    task_name = "test_name"
+    test_aggregation_type = "test_aggregation_type"
+    tasks = {task_name: {"aggregation_type": test_aggregation_type}}
     assigner = assigner(tasks, None, None)
 
     aggregation_type = assigner.get_aggregation_type_for_task(task_name)
@@ -46,13 +44,23 @@ def test_get_aggregation_type_for_task(assigner):
 def test_get_all_tasks_for_round(assigner):
     """Assert that assigner tasks object is list."""
     assigner = Assigner(None, None, None)
-    tasks = assigner.get_all_tasks_for_round('test')
+    tasks = assigner.get_all_tasks_for_round("test")
 
     assert isinstance(tasks, list)
 
+def test_default_task_group(assigner):
+    """Assert that by default learning task_group is assigned."""
+    assigner = Assigner(None,None,None)
+    assert assigner.selected_task_group == None
 
-class TestNotImplError(TestCase):
+def test_task_group_filtering_no_task_groups(assigner):
+    """Assert that task_group_filtering does not filter when no task_groups are defined."""
+    assigner = Assigner(None,None,None)
+    assigner.selected_task_group = "test_group"
+    assigner.define_task_assignments()
+    assert not hasattr(assigner, "task_groups")
 
+class TestNotImplError(TestCase):
     def test_define_task_assignments(self):
         # TODO: define_task_assignments is defined as a mock in multiple fixtures,
         # which leads the function to behave as a mock here and other tests.
@@ -61,9 +69,9 @@ def test_define_task_assignments(self):
     def test_get_tasks_for_collaborator(self):
         with self.assertRaises(NotImplementedError):
             assigner = Assigner(None, None, None)
-            assigner.get_tasks_for_collaborator('col1', 0)
+            assigner.get_tasks_for_collaborator("col1", 0)
 
     def test_get_collaborators_for_task(self):
         with self.assertRaises(NotImplementedError):
             assigner = Assigner(None, None, None)
-            assigner.get_collaborators_for_task('task_name', 0)
+            assigner.get_collaborators_for_task("task_name", 0)