Merge pull request #839 from CitrineInformatics/PLA-12272/deprecate-t…

…raining-data

PLA-12272: Deprecate training_data on non-graphs

src/citrine/__version__.py

@@ -1 +1 @@
-__version__ = '2.11.0'
+__version__ = '2.11.1'

src/citrine/informatics/predictors/auto_ml_predictor.py

@@ -89,8 +89,10 @@ def __init__(self,
         self.description: str = description
         self.inputs: List[Descriptor] = inputs
         self.estimators: Set[AutoMLEstimator] = estimators or {AutoMLEstimator.RANDOM_FOREST}
-        self.training_data: List[DataSource] = training_data or []
         self.outputs = outputs
 
+        self._check_deprecated_training_data(training_data)
+        self.training_data: List[DataSource] = training_data or []
+
     def __str__(self):
         return '<AutoMLPredictor {!r}>'.format(self.name)

src/citrine/informatics/predictors/mean_property_predictor.py

@@ -105,6 +105,8 @@ def __init__(self,
         self.impute_properties: bool = impute_properties
         self.label: Optional[str] = label
         self.default_properties: Optional[Mapping[str, Union[str, float]]] = default_properties
+
+        self._check_deprecated_training_data(training_data)
         self.training_data: List[DataSource] = training_data or []
 
     def __str__(self):

src/citrine/informatics/predictors/predictor.py

@@ -1,10 +1,12 @@
-from typing import Optional, Type
+import warnings
+from typing import Optional, Type, List
 from uuid import UUID
 
 from citrine._rest.asynchronous_object import AsynchronousObject
 from citrine._serialization import properties
 from citrine._serialization.polymorphic_serializable import PolymorphicSerializable
 from citrine._session import Session
+from citrine.informatics.data_sources import DataSource
 from citrine.resources.report import ReportResource
 from citrine.informatics.predictors.single_predict_request import SinglePredictRequest
 from citrine.informatics.predictors.single_prediction import SinglePrediction
@@ -110,3 +112,14 @@ def predict(self,
         path = self._path() + '/predict'
         res = self._session.post_resource(path, predict_request.dump(), version=self._api_version)
         return SinglePrediction.build(res)
+
+    @classmethod
+    def _check_deprecated_training_data(cls, training_data: Optional[List[DataSource]]) -> None:
+        if training_data is not None:
+            warnings.warn(
+                f"The field `training_data` on {cls.__name__} predictors is deprecated "
+                "and will be removed in version 3.0.0. Include training data for all "
+                "sub-predictors on the parent GraphPredictor. Existing training data "
+                "on this predictor will be moved to the parent GraphPredictor upon registration.",
+                DeprecationWarning
+            )

src/citrine/informatics/predictors/simple_mixture_predictor.py

@@ -44,6 +44,8 @@ def __init__(self,
                  training_data: Optional[List[DataSource]] = None):
         self.name: str = name
         self.description: str = description
+
+        self._check_deprecated_training_data(training_data)
         self.training_data: List[DataSource] = training_data or []
 
         if input_descriptor is not None:

tests/conftest.py

@@ -216,7 +216,7 @@ def valid_auto_ml_predictor_data(valid_gem_data_source_dict):
         inputs=[x.dump()],
         outputs=[z.dump()],
         estimators=[AutoMLEstimator.RANDOM_FOREST.value],
-        training_data=[valid_gem_data_source_dict]
+        training_data=[]
     )
     return PredictorEntityDataFactory(data=PredictorDataDataFactory(instance=instance))
 
@@ -395,7 +395,7 @@ def valid_predictor_report_data(example_categorical_pva_metrics, example_f1_metr
 @pytest.fixture
 def valid_ing_formulation_predictor_data():
     """Produce valid data used for tests."""
-    from citrine.informatics.descriptors import FormulationDescriptor, RealDescriptor
+    from citrine.informatics.descriptors import RealDescriptor
     instance = dict(
         type='IngredientsToSimpleMixture',
         name='Ingredients to formulation predictor',
@@ -416,7 +416,6 @@ def valid_ing_formulation_predictor_data():
 def valid_generalized_mean_property_predictor_data():
     """Produce valid data used for tests."""
     from citrine.informatics.descriptors import FormulationDescriptor
-    from citrine.informatics.data_sources import GemTableDataSource
     formulation_descriptor = FormulationDescriptor.hierarchical()
     instance = dict(
         type='GeneralizedMeanProperty',
@@ -425,7 +424,6 @@ def valid_generalized_mean_property_predictor_data():
         input=formulation_descriptor.dump(),
         properties=['density'],
         p=2,
-        training_data=[GemTableDataSource(table_id=uuid.uuid4(), table_version=0).dump()],
         impute_properties=True,
         default_properties={'density': 1.0},
         label='solvent'
@@ -437,7 +435,6 @@ def valid_generalized_mean_property_predictor_data():
 def valid_mean_property_predictor_data():
     """Produce valid data used for tests."""
     from citrine.informatics.descriptors import FormulationDescriptor, RealDescriptor
-    from citrine.informatics.data_sources import GemTableDataSource
     formulation_descriptor = FormulationDescriptor.flat()
     density = RealDescriptor(key='density', lower_bound=0, upper_bound=100, units='g/cm^3')
     instance = dict(
@@ -447,10 +444,10 @@ def valid_mean_property_predictor_data():
         input=formulation_descriptor.dump(),
         properties=[density.dump()],
         p=2,
-        training_data=[GemTableDataSource(table_id=uuid.uuid4(), table_version=0).dump()],
         impute_properties=True,
         default_properties={'density': 1.0},
-        label='solvent'
+        label='solvent',
+        training_data=[]
     )
     return PredictorEntityDataFactory(data=PredictorDataDataFactory(instance=instance))
 
@@ -524,12 +521,11 @@ def invalid_predictor_data():
 @pytest.fixture
 def valid_simple_mixture_predictor_data():
     """Produce valid data used for tests."""
-    from citrine.informatics.data_sources import GemTableDataSource
     instance = dict(
         type='SimpleMixture',
         name='Simple mixture predictor',
         description='simple mixture description',
-        training_data=[GemTableDataSource(table_id=uuid.uuid4(), table_version=0).dump()]
+        training_data=[]
     )
     return PredictorEntityDataFactory(data=PredictorDataDataFactory(instance=instance))
 

tests/informatics/test_predictors.py

@@ -91,8 +91,7 @@ def auto_ml() -> AutoMLPredictor:
         name='AutoML Predictor',
         description='Predicts z from inputs w and x',
         inputs=[w, x],
-        outputs=[z],
-        training_data=[data_source]
+        outputs=[z]
     )
 
 
@@ -102,8 +101,7 @@ def auto_ml_no_outputs() -> AutoMLPredictor:
         name='AutoML Predictor',
         description='Predicts z from inputs w and x',
         inputs=[w, x],
-        outputs=[],
-        training_data=[data_source]
+        outputs=[]
     )
 
 
@@ -113,8 +111,7 @@ def auto_ml_multiple_outputs() -> AutoMLPredictor:
         name='AutoML Predictor',
         description='Predicts z from inputs w and x',
         inputs=[w, x],
-        outputs=[z, y],
-        training_data=[data_source]
+        outputs=[z, y]
     )
 
 
@@ -125,7 +122,7 @@ def graph_predictor() -> GraphPredictor:
         name='Graph predictor',
         description='description',
         predictors=[uuid.uuid4(), uuid.uuid4()],
-        training_data=[data_source]
+        training_data=[data_source, formulation_data_source]
     )
 
 
@@ -169,7 +166,6 @@ def mean_property_predictor() -> MeanPropertyPredictor:
         input_descriptor=flat_formulation,
         properties=[density, chain_type],
         p=2.5,
-        training_data=[formulation_data_source],
         impute_properties=True,
         default_properties={'density': 1.0, 'Chain Type': 'Gaussian Coil'},
         label='solvent'
@@ -182,7 +178,6 @@ def simple_mixture_predictor() -> SimpleMixturePredictor:
     return SimpleMixturePredictor(
         name='Simple mixture predictor',
         description='Computes mean ingredient properties',
-        training_data=[formulation_data_source]
     )
 
 
@@ -229,7 +224,7 @@ def test_graph_initialization(graph_predictor):
     assert graph_predictor.name == 'Graph predictor'
     assert graph_predictor.description == 'description'
     assert len(graph_predictor.predictors) == 2
-    assert graph_predictor.training_data == [data_source]
+    assert graph_predictor.training_data == [data_source, formulation_data_source]
     assert str(graph_predictor) == '<GraphPredictor \'Graph predictor\'>'
 
 
@@ -286,7 +281,6 @@ def test_auto_ml(auto_ml):
     assert auto_ml.name == "AutoML Predictor"
     assert auto_ml.description == "Predicts z from inputs w and x"
     assert auto_ml.inputs == [w, x]
-    assert auto_ml.training_data == [data_source]
     assert auto_ml.dump()['instance']['outputs'] == [z.dump()]
 
     assert str(auto_ml) == "<AutoMLPredictor 'AutoML Predictor'>"
@@ -353,7 +347,6 @@ def test_mean_property_initialization(mean_property_predictor):
     assert mean_property_predictor.properties == [density, chain_type]
     assert mean_property_predictor.p == 2.5
     assert mean_property_predictor.impute_properties == True
-    assert mean_property_predictor.training_data == [formulation_data_source]
     assert mean_property_predictor.default_properties == {'density': 1.0, 'Chain Type': 'Gaussian Coil'}
     assert mean_property_predictor.label == 'solvent'
     expected_str = '<MeanPropertyPredictor \'Mean property predictor\'>'
@@ -385,7 +378,6 @@ def test_simple_mixture_predictor_initialization(simple_mixture_predictor):
     assert simple_mixture_predictor.name == 'Simple mixture predictor'
     assert simple_mixture_predictor.input_descriptor.key == FormulationKey.HIERARCHICAL.value
     assert simple_mixture_predictor.output_descriptor.key == FormulationKey.FLAT.value
-    assert simple_mixture_predictor.training_data == [formulation_data_source]
     expected_str = '<SimpleMixturePredictor \'Simple mixture predictor\'>'
     assert str(simple_mixture_predictor) == expected_str
 
@@ -407,6 +399,7 @@ def test_status(valid_label_fractions_predictor_data, auto_ml):
     predictor = LabelFractionsPredictor.build(valid_label_fractions_predictor_data)
     assert predictor.succeeded() and not predictor.in_progress() and not predictor.failed()
 
+
 def test_single_predict(graph_predictor):
     """Ensures we get a prediction back from a simple predict call"""
     session = mock.Mock()
@@ -426,6 +419,35 @@ def test_single_predict(graph_predictor):
     assert session.post_resource.call_count == 1
 
 
+def test_deprecated_training_data():
+    with pytest.warns(DeprecationWarning):
+        AutoMLPredictor(
+            name="AutoML",
+            description="",
+            inputs=[x, y],
+            outputs=[z],
+            training_data=[data_source]
+        )
+
+    with pytest.warns(DeprecationWarning):
+        MeanPropertyPredictor(
+            name="SimpleMixture",
+            description="",
+            input_descriptor=flat_formulation,
+            properties=[x, y, z],
+            p=1.0,
+            impute_properties=True,
+            training_data=[data_source]
+        )
+
+    with pytest.warns(DeprecationWarning):
+        SimpleMixturePredictor(
+            name="Warning",
+            description="Description",
+            training_data=[data_source]
+        )
+
+
 def test_formulation_deprecations():
     with pytest.warns(DeprecationWarning):
         SimpleMixturePredictor(

tests/serialization/test_predictors.py

@@ -9,7 +9,7 @@
 from citrine.informatics.predictors import *
 
 
-def test_simple_legacy_deserialization(valid_auto_ml_predictor_data):
+def test_auto_ml_deserialization(valid_auto_ml_predictor_data):
     """Ensure that a deserialized SimplePredictor looks sane."""
     predictor: AutoMLPredictor = AutoMLPredictor.build(valid_auto_ml_predictor_data)
     assert predictor.name == 'AutoML predictor'
@@ -18,11 +18,10 @@ def test_simple_legacy_deserialization(valid_auto_ml_predictor_data):
     assert predictor.inputs[0] == RealDescriptor("x", lower_bound=0, upper_bound=100, units="")
     assert len(predictor.outputs) == 1
     assert predictor.outputs[0] == RealDescriptor("z", lower_bound=0, upper_bound=100, units="")
-    assert len(predictor.training_data) == 1
-    assert predictor.training_data[0].table_id == UUID('e5c51369-8e71-4ec6-b027-1f92bdc14762')
+    assert len(predictor.training_data) == 0
 
 
-def test_polymorphic_legacy_deserialization(valid_auto_ml_predictor_data):
+def test_polymorphic_auto_ml_deserialization(valid_auto_ml_predictor_data):
     """Ensure that a polymorphically deserialized SimplePredictor looks sane."""
     predictor: AutoMLPredictor = Predictor.build(valid_auto_ml_predictor_data)
     assert predictor.name == 'AutoML predictor'
@@ -31,8 +30,7 @@ def test_polymorphic_legacy_deserialization(valid_auto_ml_predictor_data):
     assert predictor.inputs[0] == RealDescriptor("x", lower_bound=0, upper_bound=100, units="")
     assert len(predictor.outputs) == 1
     assert predictor.outputs[0] == RealDescriptor("z", lower_bound=0, upper_bound=100, units="")
-    assert len(predictor.training_data) == 1
-    assert predictor.training_data[0].table_id == UUID('e5c51369-8e71-4ec6-b027-1f92bdc14762')
+    assert len(predictor.training_data) == 0
 
 
 def test_legacy_serialization(valid_auto_ml_predictor_data):