Rename CoresetSupportingModule. Fix test

integrationtests/selector/integrationtest_selector.py

@@ -910,8 +910,8 @@ def test_get_available_labels(reset_after_trigger: bool):
         GetAvailableLabelsRequest(pipeline_id=pipeline_id)
     ).available_labels
 
-    assert len(available_labels) == 2
-    assert 0 in available_labels and 1 in available_labels
+    # here we expect to have 0 labels since it's before the first trigger
+    assert len(available_labels) == 0
 
     selector.inform_data_and_trigger(
         DataInformRequest(
@@ -922,6 +922,14 @@ def test_get_available_labels(reset_after_trigger: bool):
         )
     )
 
+    available_labels = selector.get_available_labels(
+        GetAvailableLabelsRequest(pipeline_id=pipeline_id)
+    ).available_labels
+
+    # we want all the labels belonging to the first trigger
+    assert len(available_labels) == 3
+    assert sorted(available_labels) == [0, 1, 189]
+
     selector.inform_data_and_trigger(
         DataInformRequest(
             pipeline_id=pipeline_id,
@@ -931,14 +939,26 @@ def test_get_available_labels(reset_after_trigger: bool):
         )
     )
 
+    # this label (99) should not appear in the available labels since it belongs to a future trigger.
+    selector.inform_data(
+        DataInformRequest(
+            pipeline_id=pipeline_id,
+            keys=[99],
+            timestamps=[7],
+            labels=[99],
+        )
+    )
+
     available_labels = selector.get_available_labels(
         GetAvailableLabelsRequest(pipeline_id=pipeline_id)
     ).available_labels
 
     if reset_after_trigger:
+        # only the last trigger must be considered but not point99
         assert len(available_labels) == 3
         assert sorted(available_labels) == [7, 10, 45]
     else:
+        # every past point must be considered. Only point99 is excluded.
         assert len(available_labels) == 6
         assert sorted(available_labels) == [0, 1, 7, 10, 45, 189]
 

modyn/models/articlenet/articlenet.py

@@ -1,7 +1,7 @@
 from typing import Any
 
 import torch
-from modyn.models.coreset_methods_support import CoresetMethodsSupport
+from modyn.models.coreset_methods_support import CoresetSupportingModule
 from torch import nn
 from transformers import DistilBertModel
 
@@ -44,7 +44,7 @@ def __call__(self, data: torch.Tensor) -> torch.Tensor:
         return pooled_output
 
 
-class ArticleNetwork(CoresetMethodsSupport):
+class ArticleNetwork(CoresetSupportingModule):
     def __init__(self, num_classes: int) -> None:
         super().__init__()
         self.featurizer = DistilBertFeaturizer.from_pretrained("distilbert-base-uncased")

modyn/models/coreset_methods_support.py

@@ -25,7 +25,7 @@ def end_recording(self) -> None:
         self.embedding = None
 
 
-class CoresetMethodsSupport(nn.Module, ABC):
+class CoresetSupportingModule(nn.Module, ABC):
     """
     This class is used to support some Coreset Methods.
     Embeddings, here defined as the activation before the last layer, are often used to estimate the importance of

modyn/models/dlrm/dlrm.py

@@ -2,7 +2,7 @@
 
 import numpy as np
 import torch
-from modyn.models.coreset_methods_support import CoresetMethodsSupport, EmbeddingRecorder
+from modyn.models.coreset_methods_support import CoresetSupportingModule, EmbeddingRecorder
 from modyn.models.dlrm.nn.factories import create_interaction
 from modyn.models.dlrm.nn.parts import DlrmBottom, DlrmTop
 from modyn.models.dlrm.utils.install_lib import install_cuda_extensions_if_not_present
@@ -17,7 +17,7 @@ def __init__(self, model_configuration: dict[str, Any], device: str, amp: bool)
         self.model.to(device)
 
 
-class DlrmModel(CoresetMethodsSupport):
+class DlrmModel(CoresetSupportingModule):
     # pylint: disable=too-many-instance-attributes
     def __init__(self, model_configuration: dict[str, Any], device: str, amp: bool) -> None:
         super().__init__()

modyn/models/dlrm/nn/parts.py

@@ -19,7 +19,7 @@
 from typing import Optional, Sequence, Tuple
 
 import torch
-from modyn.models.coreset_methods_support import CoresetMethodsSupport
+from modyn.models.coreset_methods_support import CoresetSupportingModule
 from modyn.models.dlrm.nn.embeddings import Embeddings
 from modyn.models.dlrm.nn.factories import create_embeddings, create_mlp
 from modyn.models.dlrm.nn.interactions import Interaction
@@ -105,7 +105,7 @@ def forward(self, numerical_input, categorical_inputs) -> Tuple[torch.Tensor, Op
         return torch.cat(bottom_output, dim=1), bottom_mlp_output
 
 
-class DlrmTop(CoresetMethodsSupport):
+class DlrmTop(CoresetSupportingModule):
     def __init__(self, top_mlp_sizes: Sequence[int], interaction: Interaction, device: str, use_cpp_mlp: bool = False):
         super().__init__()
 

modyn/models/fmownet/fmownet.py

@@ -2,7 +2,7 @@
 
 import torch
 import torch.nn.functional as F
-from modyn.models.coreset_methods_support import CoresetMethodsSupport
+from modyn.models.coreset_methods_support import CoresetSupportingModule
 from torch import nn
 from torchvision.models import densenet121
 
@@ -20,7 +20,7 @@ def __init__(self, model_configuration: dict[str, Any], device: str, amp: bool)
         self.model.to(device)
 
 
-class FmowNetModel(CoresetMethodsSupport):
+class FmowNetModel(CoresetSupportingModule):
     def __init__(self, num_classes: int) -> None:
         super().__init__()
         self.num_classes = num_classes

modyn/models/resnet18/resnet18.py

@@ -1,7 +1,7 @@
 from typing import Any
 
 import torch
-from modyn.models.coreset_methods_support import CoresetMethodsSupport
+from modyn.models.coreset_methods_support import CoresetSupportingModule
 from torch import Tensor, nn
 from torchvision.models.resnet import BasicBlock, ResNet
 
@@ -17,7 +17,7 @@ def __init__(self, model_configuration: dict[str, Any], device: str, amp: bool)
 # torchvision https://github.com/pytorch/vision/blob/main/torchvision/models/resnet.py
 
 
-class ResNet18Modyn(ResNet, CoresetMethodsSupport):
+class ResNet18Modyn(ResNet, CoresetSupportingModule):
     def __init__(self, model_configuration: dict[str, Any]) -> None:
         super().__init__(BasicBlock, [2, 2, 2, 2], **model_configuration)
 

modyn/models/yearbooknet/yearbooknet.py

@@ -1,7 +1,7 @@
 from typing import Any
 
 import torch
-from modyn.models.coreset_methods_support import CoresetMethodsSupport
+from modyn.models.coreset_methods_support import CoresetSupportingModule
 from torch import nn
 
 
@@ -18,7 +18,7 @@ def __init__(self, model_configuration: dict[str, Any], device: str, amp: bool)
         self.model.to(device)
 
 
-class YearbookNetModel(CoresetMethodsSupport):
+class YearbookNetModel(CoresetSupportingModule):
     def __init__(self, num_input_channels: int, num_classes: int) -> None:
         super().__init__()
         self.enc = nn.Sequential(

modyn/selector/internal/selector_strategies/abstract_selection_strategy.py

@@ -186,6 +186,7 @@ def trigger(self) -> tuple[int, int, int]:
             tuple[int, int, int]: Trigger ID, how many keys are in the trigger, number of overall partitions
         """
         # TODO(#276) Unify AbstractSelection Strategy and LocalDatasetWriter
+
         trigger_id = self._next_trigger_id
         total_keys_in_trigger = 0
 
@@ -404,6 +405,7 @@ def get_available_labels(self) -> list[int]:
                 database.session.query(SelectorStateMetadata.label)
                 .filter(
                     SelectorStateMetadata.pipeline_id == self._pipeline_id,
+                    SelectorStateMetadata.seen_in_trigger_id < self._next_trigger_id,
                     SelectorStateMetadata.seen_in_trigger_id >= self._next_trigger_id - self.tail_triggers - 1
                     if self.tail_triggers is not None
                     else True,

modyn/tests/selector/internal/selector_strategies/test_abstract_selection_strategy.py

@@ -329,7 +329,7 @@ def test_get_available_labels_reset():
 @patch.multiple(AbstractSelectionStrategy, __abstractmethods__=set())
 def test_get_available_labels_no_reset():
     with MetadataDatabaseConnection(get_minimal_modyn_config()) as database:
-        # first trigger
+        # first batch of data
         database.session.add(
             SelectorStateMetadata(pipeline_id=1, sample_key=0, seen_in_trigger_id=0, timestamp=0, label=1)
         )
@@ -346,10 +346,13 @@ def test_get_available_labels_no_reset():
 
     abstr = AbstractSelectionStrategy({"limit": -1, "reset_after_trigger": False}, get_minimal_modyn_config(), 1, 1000)
 
+    assert sorted(abstr.get_available_labels()) == []
+    # simulate a trigger
+    abstr._next_trigger_id +=1
     assert sorted(abstr.get_available_labels()) == [0, 1, 18]
 
     with MetadataDatabaseConnection(get_minimal_modyn_config()) as database:
-        # second trigger
+        # another batch of data is inserted with just one more class
         database.session.add(
             SelectorStateMetadata(pipeline_id=1, sample_key=4, seen_in_trigger_id=1, timestamp=0, label=0)
         )
@@ -358,5 +361,7 @@ def test_get_available_labels_no_reset():
         )
         database.session.commit()
 
+    assert sorted(abstr.get_available_labels()) == [0, 1, 18]
+    # simulate a trigger
     abstr._next_trigger_id += 1
     assert sorted(abstr.get_available_labels()) == [0, 1, 18, 890]