try to create a taskconfig but will be removed after

aramis-lab · Aug 30, 2024 · e027c21 · e027c21
1 parent ad8f37c
commit e027c21
Show file tree

Hide file tree

Showing 2 changed files with 43 additions and 4 deletions.
diff --git a/clinicadl/maps_manager/maps_manager.py b/clinicadl/maps_manager/maps_manager.py
@@ -76,7 +76,15 @@ def __init__(
                 )
             test_parameters = self.get_parameters()
             # test_parameters = path_decoder(test_parameters)
+            from clinicadl.utils.task_manager.task_manager import TaskConfig
+
             self.parameters = add_default_values(test_parameters)
+            self.task_config = TaskConfig(
+                network_task=self.network_task,
+                mode=self.mode,
+                n_classes=self.output_size,
+            )
+
             self.split_name = (
                 self._check_split_wording()
             )  # Used only for retro-compatibility
@@ -417,9 +425,14 @@ def _check_args(self, parameters):
             self.parameters["label"] = None
 
         from clinicadl.utils.enum import Task
-        from clinicadl.utils.task_manager.task_manager import get_default_network
+        from clinicadl.utils.task_manager.task_manager import (
+            TaskConfig,
+            get_default_network,
+        )
 
         self.network_task = Task(self.parameters["network_task"])
+        self.task_config = TaskConfig(self.network_task, self.mode, df=train_df)
+        # self.task_manager = self._init_task_manager(df=train_df)
 
         if self.parameters["architecture"] == "default":
             self.parameters["architecture"] = get_default_network(self.network_task)

diff --git a/clinicadl/utils/task_manager/task_manager.py b/clinicadl/utils/task_manager/task_manager.py
@@ -5,6 +5,11 @@
 import pandas as pd
 import torch
 import torch.distributed as dist
+from pydantic import (
+    BaseModel,
+    ConfigDict,
+    computed_field,
+)
 from torch import Tensor, nn
 from torch.cuda.amp import autocast
 from torch.nn.functional import softmax
@@ -32,6 +37,8 @@
 # elif network_task == Task.RECONSTRUCTION:
 
 
+# This function is not useful anymore since we introduced config class
+# default network will automatically be initialized when running the task
 def get_default_network(network_task: Task) -> str:  # return Network
     """Returns the default network to use when no architecture is specified."""
     if network_task == Task.CLASSIFICATION:
@@ -109,7 +116,7 @@ def handle_reconstruction_loss(criterion, compatible_losses):
 
 def output_size(
     network_task: Union[str, Task],
-    input_size: Sequence[int],
+    input_size: Optional[Sequence[int]],
     df: pd.DataFrame,
     label: str,
 ) -> Union[int, Sequence[int]]:
@@ -719,8 +726,27 @@ def get_sampler(weights):
     return get_sampler(weights)
 
 
-class TaskManager:
-    def __init__(self, mode: str, n_classes: int = None):
+class TaskConfig(BaseModel):
+    mode: str
+    network_task: Task
+    n_classe: Optional[int] = None
+    df: Optional[pd.DataFrame] = None
+    label: Optional[str] = None
+
+    def __init__(
+        self,
+        network_task: Union[str, Task],
+        mode: str,
+        n_classes: Optional[int] = None,
+        df: Optional[pd.DataFrame] = None,
+        label: Optional[str] = None,
+    ):
+        network_task = Task(network_task)
+        if network_task == Task.CLASSIFICATION:
+            if n_classes is None and df is not None:
+                n_classes = output_size(Task.CLASSIFICATION, None, df, label)
+            self.n_classes = n_classes
+
         self.mode = mode
         self.metrics_module = MetricModule(
             evaluation_metrics(network_task), n_classes=n_classes