Pull request update/240417

1685a40 OS-7482. Fixed bug with removed list values in etcd
1a99bd4 OS-7425. Fixed etcd black and white lists values
a32ea69 OS-7463. Leaderboards improvements

arcee/arcee_receiver/models.py

@@ -147,6 +147,13 @@ class LeaderboardPatchIn(LeaderboardPostIn):
     primary_metric: Optional[str] = None
     group_by_hp: Optional[bool] = None
 
+    @model_validator(mode='after')
+    def set_metrics(self):
+        self.metrics = list(set(self.other_metrics))
+        if self.primary_metric:
+            self.metrics = list(set(self.metrics + [self.primary_metric]))
+        return self
+
 
 class Leaderboard(LeaderboardPostIn):
     id: str = id_
@@ -160,8 +167,8 @@ class Config:
 
 
 class LeaderboardDatasetPatchIn(BaseModel):
-    dataset_ids: Optional[list]
-    name: Optional[str]
+    dataset_ids: Optional[list] = []
+    name: Optional[str] = None
 
     @staticmethod
     def remove_dup_ds_ids(kwargs):

arcee/arcee_receiver/tests/base.py

@@ -21,6 +21,8 @@ async def cluster_secret(self):
 
 class Urls:
     leaderboards = '/arcee/v2/tasks/{}/leaderboards'
+    leaderboard_datasets = '/arcee/v2/leaderboards/{}/leaderboard_datasets'
+    leaderboard_dataset = '/arcee/v2/leaderboard_datasets/{}'
     tasks = '/arcee/v2/tasks'
     task = '/arcee/v2/tasks/{}'
     models = '/arcee/v2/models'
@@ -101,6 +103,68 @@ async def prepare_run(task_id, start, state, number, data):
     return await DB_MOCK['run'].find_one({'_id': run['_id']})
 
 
+async def prepare_leaderboard(primary_metric, task_id, other_metrics=None,
+                              filters=None, group_by_hp=False,
+                              grouping_tags=None):
+    if other_metrics is None:
+        other_metrics = []
+    if filters is None:
+        filters = []
+    if grouping_tags is None:
+        grouping_tags = []
+    lb = {
+        "_id": str(uuid.uuid4()),
+        "primary_metric": primary_metric,
+        "other_metrics": other_metrics,
+        "filters": filters,
+        "group_by_hp": group_by_hp,
+        "grouping_tags": grouping_tags,
+        "task_id": task_id,
+        "token": TOKEN1,
+        "created_at": int(datetime.now(tz=timezone.utc).timestamp()),
+        "deleted_at": 0
+    }
+    await DB_MOCK['leaderboard'].insert_one(lb)
+    return await DB_MOCK['leaderboard'].find_one({'_id': lb['_id']})
+
+
+async def prepare_leaderboard_dataset(leaderboard_id, name=None,
+                                      dataset_ids=None):
+    if dataset_ids is None:
+        dataset_ids = []
+    leaderboard_dataset = {
+        "_id": str(uuid.uuid4()),
+        "leaderboard_id": leaderboard_id,
+        "name": name or 'test',
+        "dataset_ids": dataset_ids,
+        "token": TOKEN1,
+        "created_at": int(datetime.now(tz=timezone.utc).timestamp()),
+        "deleted_at": 0
+    }
+    await DB_MOCK['leaderboard_dataset'].insert_one(leaderboard_dataset)
+    return await DB_MOCK['leaderboard_dataset'].find_one(
+        {'_id': leaderboard_dataset['_id']})
+
+
+async def prepare_dataset(name=None, description=None, labels=None, path=None,
+                          training_set=None, validation_set=None):
+    if labels is None:
+        labels = []
+    dataset = {
+        "_id": str(uuid.uuid4()),
+        "name": name or 'test',
+        "description": description or 'test',
+        "labels": labels,
+        "path": path or 'test',
+        "training_set": training_set,
+        "validation_set": validation_set,
+        "created_at": int(datetime.now(tz=timezone.utc).timestamp()),
+        "deleted_at": 0
+    }
+    await DB_MOCK['dataset'].insert_one(dataset)
+    return await DB_MOCK['dataset'].find_one({'_id': dataset['_id']})
+
+
 async def prepare_model(token=TOKEN1, key="key"):
     model = {
         "_id": str(uuid.uuid4()),

arcee/arcee_receiver/tests/conftest.py

@@ -22,11 +22,13 @@ async def clean_env():
     await DB_MOCK['metric'].drop()
     await DB_MOCK['task'].drop()
     await DB_MOCK['leaderboard'].drop()
+    await DB_MOCK['leaderboard_dataset'].drop()
     await DB_MOCK['run'].drop()
     await DB_MOCK['model'].drop()
     await DB_MOCK['model_version'].drop()
     await DB_MOCK['log'].drop()
     await DB_MOCK['platform'].drop()
+    await DB_MOCK['dataset'].drop()
 
 
 @pytest.fixture(autouse=True)

arcee/arcee_receiver/tests/test_leaderboard.py

@@ -365,6 +365,35 @@ async def test_patch_leaderboard(app):
     assert response.json == lb
 
 
+@pytest.mark.asyncio
+async def test_patch_leaderboard_group_by_hp(app):
+    client = app.asgi_client
+    await prepare_token()
+    metrics = await prepare_metrics()
+    tasks = await prepare_tasks()
+    lb = {
+        "_id": str(uuid.uuid4()),
+        "primary_metric": metrics[0]['_id'],
+        "task_id": tasks[0]['_id'],
+        "other_metrics": [metrics[1]['_id']],
+        "filters": [{"id": metrics[1]['_id'], "min": 0, "max": 100}],
+        "group_by_hp": True,
+        "grouping_tags": [],
+        "deleted_at": 0,
+        "created_at": int(datetime.now(tz=timezone.utc).timestamp()),
+        "token": TOKEN1
+    }
+    await DB_MOCK['leaderboard'].insert_one(lb)
+    lb_update = {
+        "group_by_hp": False
+    }
+    _, response = await client.patch(Urls.leaderboards.format(tasks[0]['_id']),
+                                     headers={"x-api-key": TOKEN1},
+                                     data=json.dumps(lb_update))
+    assert response.status == 200
+    assert response.json['group_by_hp'] is False
+
+
 @pytest.mark.asyncio
 async def test_patch_invalid_params(app):
     client = app.asgi_client