GH-52: feat(pipeline): Add table information to pipeline serialization

src/getml_io/getml/tables.py

@@ -0,0 +1,8 @@
+from pydantic import BaseModel
+
+
+class Table(BaseModel, frozen=True):
+    name: str
+    marker: str
+    target: str
+    importance: float

src/getml_io/metadata/pipeline_information.py

@@ -16,6 +16,7 @@
 from getml_io.getml.predictors import FeatureSelector, Predictor
 from getml_io.getml.preprocessors import Preprocessor
 from getml_io.getml.scores import Scores
+from getml_io.getml.tables import Table
 from getml_io.metadata.data_model_information import DataModelInformation
 from getml_io.metadata.dataframe_information import DataFrameInformationByName
 from getml_io.metadata.placeholder_information import PlaceholderInformation
@@ -47,4 +48,4 @@ class PipelineInformation(BaseModel, frozen=True):
     scores: Scores
     columns: Sequence[Column]
     metadata: PipelineMetaData
-    # tables # TODO @urfoex: #52
+    tables: Sequence[Table]

src/getml_io/serialize/pipeline.py

@@ -18,12 +18,14 @@
 from getml.pipeline import Features as GetMLFeatures
 from getml.pipeline import Pipeline
 from getml.pipeline import Scores as GetMLScores
+from getml.pipeline import Tables as GetMLTables
 from getml.pipeline.column import Column as GetMLColumn
 from getml.pipeline.metadata import AllMetadata
 from getml.pipeline.metadata import Metadata as GetMLMetadata
 from getml.pipeline.score import ClassificationScore as GetMLClassificationScore
 from getml.pipeline.score import RegressionScore as GetMLRegressionScore
 from getml.pipeline.score import Score as GetMLScore
+from getml.pipeline.table import Table as GetMLTable
 from numpy.typing import NDArray
 
 from getml_io.getml.columns import Column
@@ -60,6 +62,7 @@
     TextFieldSplitter,
 )
 from getml_io.getml.scores import ClassificationScore, RegressionScore, Scores
+from getml_io.getml.tables import Table
 from getml_io.metadata.dataframe_information import DataFrameInformationByName
 from getml_io.metadata.pipeline_information import (
     LossFunction,
@@ -149,7 +152,7 @@ def serialize_pipeline(
         scores=serialize_scores(pipeline.scores),
         columns=serialize_columns(pipeline.columns),
         metadata=serialize_all_metadata(pipeline.metadata),
-        # tables # TODO @urfoex: #52
+        tables=serialize_tables(pipeline.tables),
     )
     pipeline_information_json_path = serialize_pipeline_information(
         pipeline_information=pipeline_information,
@@ -472,3 +475,24 @@ def _serialize_metadata(metadata: GetMLMetadata) -> DataFrameMetaData:
         name=metadata.name,
         roles=serialize_roles(metadata.roles),
     )
+
+
+def serialize_tables(tables: GetMLTables) -> list[Table]:
+    """Serialize getML Tables into a list of Table objects.
+
+    Args:
+        tables: The getML Tables to serialize.
+
+    Returns:
+        list[Table]: The serialized Tables information.
+
+    """
+    return [
+        Table(
+            name=table.name,
+            marker=table.marker,
+            target=table.target,
+            importance=table.importance,
+        )
+        for table in cast("Sequence[GetMLTable]", tables.data)
+    ]

tests/integration/data/loans/expected.pipeline.json

@@ -631,9 +631,9 @@
   "feature_learners": [
     {
       "aggregation": [
+        "SUM",
         "MIN",
         "COUNT",
-        "SUM",
         "MAX",
         "AVG"
       ],
@@ -649,19 +649,19 @@
       "num_threads": 0,
       "propositionalization": {
         "aggregation": [
-          "STDDEV",
-          "MIN",
-          "COUNT DISTINCT",
-          "COUNT",
-          "MEDIAN",
           "MAX",
           "SUM",
-          "FIRST",
+          "MODE",
           "LAST",
           "TREND",
-          "AVG",
+          "MIN",
           "COUNT MINUS COUNT DISTINCT",
-          "MODE"
+          "COUNT",
+          "FIRST",
+          "MEDIAN",
+          "AVG",
+          "COUNT DISTINCT",
+          "STDDEV"
         ],
         "delta_t": 0.0,
         "loss_function": "CrossEntropyLoss",
@@ -1365,5 +1365,25 @@
         }
       }
     ]
-  }
+  },
+  "tables": [
+    {
+      "name": "meta",
+      "marker": "[PERIPHERAL]",
+      "target": "default",
+      "importance": 0.2731073049046048
+    },
+    {
+      "name": "trans",
+      "marker": "[PERIPHERAL]",
+      "target": "default",
+      "importance": 0.39642128482217825
+    },
+    {
+      "name": "population",
+      "marker": "[POPULATION]",
+      "target": "default",
+      "importance": 0.330471410273217
+    }
+  ]
 }

tests/integration/data/numerical/expected.pipeline.json

@@ -570,19 +570,19 @@
       "num_threads": 0,
       "propositionalization": {
         "aggregation": [
-          "STDDEV",
-          "MIN",
-          "COUNT DISTINCT",
-          "COUNT",
-          "MEDIAN",
           "MAX",
           "SUM",
-          "FIRST",
+          "MODE",
           "LAST",
           "TREND",
-          "AVG",
+          "MIN",
           "COUNT MINUS COUNT DISTINCT",
-          "MODE"
+          "COUNT",
+          "FIRST",
+          "MEDIAN",
+          "AVG",
+          "COUNT DISTINCT",
+          "STDDEV"
         ],
         "delta_t": 0.0,
         "loss_function": "SquareLoss",
@@ -915,5 +915,19 @@
         }
       }
     ]
-  }
+  },
+  "tables": [
+    {
+      "name": "perph",
+      "marker": "[PERIPHERAL]",
+      "target": "targets",
+      "importance": 0.2364818293321066
+    },
+    {
+      "name": "population",
+      "marker": "[POPULATION]",
+      "target": "targets",
+      "importance": 0.7635181706678932
+    }
+  ]
 }

tests/integration/data/robot/expected.pipeline.json

@@ -7020,19 +7020,19 @@
       "num_threads": 0,
       "propositionalization": {
         "aggregation": [
-          "STDDEV",
-          "MIN",
-          "COUNT DISTINCT",
-          "COUNT",
-          "MEDIAN",
           "MAX",
           "SUM",
-          "FIRST",
+          "MODE",
           "LAST",
           "TREND",
-          "AVG",
+          "MIN",
           "COUNT MINUS COUNT DISTINCT",
-          "MODE"
+          "COUNT",
+          "FIRST",
+          "MEDIAN",
+          "AVG",
+          "COUNT DISTINCT",
+          "STDDEV"
         ],
         "delta_t": 0.0,
         "loss_function": "SquareLoss",
@@ -11546,5 +11546,43 @@
         }
       }
     ]
-  }
+  },
+  "tables": [
+    {
+      "name": "full",
+      "marker": "[PERIPHERAL]",
+      "target": "f_x",
+      "importance": 0.6157352560942454
+    },
+    {
+      "name": "population",
+      "marker": "[POPULATION]",
+      "target": "f_x",
+      "importance": 0.3842647439057541
+    },
+    {
+      "name": "full",
+      "marker": "[PERIPHERAL]",
+      "target": "f_y",
+      "importance": 0.6632312212854787
+    },
+    {
+      "name": "population",
+      "marker": "[POPULATION]",
+      "target": "f_y",
+      "importance": 0.33676877871452054
+    },
+    {
+      "name": "full",
+      "marker": "[PERIPHERAL]",
+      "target": "f_z",
+      "importance": 0.5290353310682591
+    },
+    {
+      "name": "population",
+      "marker": "[POPULATION]",
+      "target": "f_z",
+      "importance": 0.470964668931741
+    }
+  ]
 }

tests/unit/conftest.py

@@ -26,12 +26,14 @@
 from getml.pipeline import Features as GetMLFeatures
 from getml.pipeline import Pipeline
 from getml.pipeline import Scores as GetMLScores
+from getml.pipeline import Tables as GetMLTables
 from getml.pipeline.columns import Column as GetMLColumn
 from getml.pipeline.feature import Feature as GetMLFeature
 from getml.pipeline.metadata import AllMetadata as GetMLAllMetadata
 from getml.pipeline.metadata import Metadata as GetMLMetadata
 from getml.pipeline.score import ClassificationScore as GetMLClassificationScore
 from getml.pipeline.score import RegressionScore as GetMLRegressionScore
+from getml.pipeline.table import Table as GetMLTable
 from numpy.typing import NDArray
 
 from getml_io.getml.columns import Column
@@ -48,6 +50,7 @@
 from getml_io.getml.relationships import Relationship
 from getml_io.getml.roles import Role, Roles
 from getml_io.getml.scores import ClassificationScore, Scores
+from getml_io.getml.tables import Table
 from getml_io.metadata.container_information import ContainerInformation
 from getml_io.metadata.data_model_information import DataModelInformation
 from getml_io.metadata.dataframe_information import (
@@ -445,6 +448,7 @@ def pipeline_information_empty(
         scores=[],
         columns=[],
         metadata=PipelineMetaData(population=None, peripheral=[]),
+        tables=[],
     )
 
 
@@ -541,6 +545,19 @@ def getml_all_metadata() -> GetMLAllMetadata:
     )
 
 
+@pytest.fixture
+def mock_tables(mocker: pytest_mock.MockerFixture) -> GetMLTables:
+    table = GetMLTable(
+        name="test_table",
+        marker="test_marker",
+        target="target0",
+        importance=0.0,
+    )
+    tables = mocker.MagicMock(spec=GetMLTables)
+    tables.data = [table]
+    return tables
+
+
 @pytest.fixture
 def mock_pipeline(  # noqa: PLR0913
     mocker: pytest_mock.MockerFixture,
@@ -550,6 +567,7 @@ def mock_pipeline(  # noqa: PLR0913
     mock_scores_regression: GetMLScores,
     mock_columns: GetMLColumns,
     getml_all_metadata: GetMLAllMetadata,
+    mock_tables: GetMLTables,
 ) -> Pipeline:
     pipeline = mocker.Mock()
     pipeline.id = "mock_pipeline_id"
@@ -588,6 +606,7 @@ def pipeline_transform(_: DataFrame | View | Subset, *, df_name: str) -> DataFra
     pipeline.scores = mock_scores_regression
     pipeline.columns = mock_columns
     pipeline.metadata = getml_all_metadata
+    pipeline.tables = mock_tables
     return pipeline
 
 
@@ -830,6 +849,18 @@ def meta_data(
     )
 
 
+@pytest.fixture
+def tables() -> list[Table]:
+    return [
+        Table(
+            name="test_table",
+            marker="test_marker",
+            target="target0",
+            importance=0.0,
+        ),
+    ]
+
+
 @pytest.fixture
 def pipeline_information(  # noqa: PLR0913
     data_model_information: DataModelInformation,
@@ -842,6 +873,7 @@ def pipeline_information(  # noqa: PLR0913
     scores: Scores,
     columns: Sequence[Column],
     meta_data: PipelineMetaData,
+    tables: Sequence[Table],
 ) -> PipelineInformation:
     return PipelineInformation(
         id="pipeline_id",
@@ -864,6 +896,7 @@ def pipeline_information(  # noqa: PLR0913
         scores=scores,
         columns=columns,
         metadata=meta_data,
+        tables=tables,
     )
 
 

tests/unit/metadata/test_pipeline_information.py

@@ -65,6 +65,7 @@ def _get_expected_serialized_empty_pipeline_information() -> PipelineInformation
             "population": None,
             "peripheral": [],
         },
+        "tables": [],
     }
 
 
@@ -402,4 +403,12 @@ def _get_expected_serialized_pipeline_information() -> PipelineInformationType:
                 },
             ],
         },
+        "tables": [
+            {
+                "importance": 0.0,
+                "marker": "test_marker",
+                "name": "test_table",
+                "target": "target0",
+            },
+        ],
     }