From 3f2e89108220b26c5e9fb18cf3e63a79055e9c8f Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Tue, 15 Oct 2024 21:33:26 -0700
Subject: [PATCH 01/17] start of model class

---
 src/aind_data_schema/core/model.py | 72 ++++++++++++++++++++++++++++++
 1 file changed, 72 insertions(+)
 create mode 100644 src/aind_data_schema/core/model.py

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
new file mode 100644
index 000000000..16c6b033b
--- /dev/null
+++ b/src/aind_data_schema/core/model.py
@@ -0,0 +1,72 @@
+""" schema describing an analysis model """
+
+from decimal import Decimal
+from enum import Enum
+from typing import List, Literal, Optional
+
+from pydantic import Field
+
+from aind_data_schema_models.modalities import Modality
+
+from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AwareDatetimeWithDefault
+from aind_data_schema.components.devices import Software
+
+
+class Backbone(str, Enum):
+    """Types of network backbones"""
+
+    ALEXNET = "AlexNet"
+    RESNET = "ResNet"
+    VGGNET = "VGGNet"
+
+
+class ModelArchitecture(AindModel):
+    """Description of model architecture"""
+
+    backbone: Backbone = Field(..., title="Backbone")
+    layers: int = Field(..., title="Layers")
+    parameters: AindGenericType = Field(..., title="Parameters")
+
+
+class ModelPerformance(AindModel):
+    """Description of performance metrics"""
+
+    precision: Decimal = Field(..., title="Precision")
+    recall: Decimal = Field(..., title="Recall")
+    f1_score: Decimal = Field(..., title="F1 score")
+
+
+class ModelTraining(AindModel):
+    """Description of model training"""
+
+    training_data: str = Field(..., title="Path to training data")
+    training_data_description: Optional[str] = Field(default=None, title="Description of training data") 
+    validation_folds: int = Field(..., title="Validation folds") #is the validation methods x-fold? or are there other validations? Enum?
+    performance: ModelPerformance = Field(..., title="Training performance")
+
+
+class ModelEvaluation(AindModel):
+    """Description of model evaluation"""
+
+
+
+class Model(AindCoreModel):
+    """Description of an analysis model"""
+
+    _DESCRIBED_BY_URL = AindCoreModel._DESCRIBED_BY_BASE_URL.default + "aind_data_schema/core/model.py"
+    describedBy: str = Field(_DESCRIBED_BY_URL, json_schema_extra={"const": _DESCRIBED_BY_URL})
+    schema_version: Literal["1.0.1"] = Field("0.0.1")
+
+    name: str = Field(..., title="Name")
+    developer_full_name: str = Field(..., title="Name of developer")
+    #developer_institution: 
+    #shared by
+    modality: Modality.ONE_OF = Field(..., title="Modality")
+    date_trained: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
+    model_architecture: ModelArchitecture = Field(..., title="Model architecture")
+    software: Software = Field(..., title="software")
+    direct_use: str = Field(..., title="Intended model use", description="Semantic description of intended use")
+    limitations: Optional[str] = Field(default=None, title="Model limitations")
+    training: List[ModelTraining] = Field(..., title="Training")
+
+

From 9d6932920a79ca8abd936dd9e42d29f5b78f7755 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Wed, 16 Oct 2024 16:01:12 -0700
Subject: [PATCH 02/17] added more to model

---
 src/aind_data_schema/core/model.py | 38 ++++++++++++++++++++----------
 1 file changed, 26 insertions(+), 12 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 16c6b033b..9f73d4ca4 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -2,11 +2,12 @@
 
 from decimal import Decimal
 from enum import Enum
-from typing import List, Literal, Optional
+from typing import List, Literal, Optional, Union
 
 from pydantic import Field
 
 from aind_data_schema_models.modalities import Modality
+from aind_data_schema_models.organizations import Organization
 
 from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AwareDatetimeWithDefault
 from aind_data_schema.components.devices import Software
@@ -26,14 +27,22 @@ class ModelArchitecture(AindModel):
     backbone: Backbone = Field(..., title="Backbone")
     layers: int = Field(..., title="Layers")
     parameters: AindGenericType = Field(..., title="Parameters")
+    #notes?
 
 
-class ModelPerformance(AindModel):
+class ScoreStatistics(AindModel):
+    """Statistics for x-fold validation scores"""
+
+    mean: Decimal = Field(..., title="Mean")
+    std: Decimal = Field(..., title="Standard deviation")
+
+
+class PerformanceScore(AindModel):
     """Description of performance metrics"""
 
-    precision: Decimal = Field(..., title="Precision")
-    recall: Decimal = Field(..., title="Recall")
-    f1_score: Decimal = Field(..., title="F1 score")
+    precision: Union[Decimal, ScoreStatistics] = Field(..., title="Precision")
+    recall: Union[Decimal, ScoreStatistics] = Field(..., title="Recall")
+    f1_score: Union[Decimal, ScoreStatistics] = Field(..., title="F1 score")
 
 
 class ModelTraining(AindModel):
@@ -41,13 +50,20 @@ class ModelTraining(AindModel):
 
     training_data: str = Field(..., title="Path to training data")
     training_data_description: Optional[str] = Field(default=None, title="Description of training data") 
+    training_date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
     validation_folds: int = Field(..., title="Validation folds") #is the validation methods x-fold? or are there other validations? Enum?
-    performance: ModelPerformance = Field(..., title="Training performance")
+    performance: PerformanceScore = Field(..., title="Training performance")
+    notes: Optional[str] = Field(default=None, title="Notes")
 
 
 class ModelEvaluation(AindModel):
     """Description of model evaluation"""
 
+    evaluation_data: str = Field(..., title="Path to evaluation data")
+    evaluation_data_description: Optional[str] = Field(default=None, title="Description of evaluation data")
+    evaluation_date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
+    performance: PerformanceScore = Field(..., title="Evaluation performance")
+    notes: Optional[str] = Field(default=None, title="Notes")
 
 
 class Model(AindCoreModel):
@@ -58,15 +74,13 @@ class Model(AindCoreModel):
     schema_version: Literal["1.0.1"] = Field("0.0.1")
 
     name: str = Field(..., title="Name")
-    developer_full_name: str = Field(..., title="Name of developer")
-    #developer_institution: 
-    #shared by
+    developer_full_name: Optional[str] = Field(default=None, title="Name of developer")
+    developer_institution: Organization.ONE_OF = Field(default=None, title="Institute where developed")
     modality: Modality.ONE_OF = Field(..., title="Modality")
-    date_trained: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
     model_architecture: ModelArchitecture = Field(..., title="Model architecture")
     software: Software = Field(..., title="software")
     direct_use: str = Field(..., title="Intended model use", description="Semantic description of intended use")
     limitations: Optional[str] = Field(default=None, title="Model limitations")
     training: List[ModelTraining] = Field(..., title="Training")
-
-
+    evaluations: Optional[List[ModelEvaluation]] = Field(default=[], title="Evaluations")
+    notes: Optional[str] = Field(default=None, title="Notes")

From 90357d5c4802fcecd0e9a18228e3740129a48d42 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Wed, 16 Oct 2024 16:02:14 -0700
Subject: [PATCH 03/17] fixed literal

---
 src/aind_data_schema/core/model.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 9f73d4ca4..714bda435 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -71,7 +71,7 @@ class Model(AindCoreModel):
 
     _DESCRIBED_BY_URL = AindCoreModel._DESCRIBED_BY_BASE_URL.default + "aind_data_schema/core/model.py"
     describedBy: str = Field(_DESCRIBED_BY_URL, json_schema_extra={"const": _DESCRIBED_BY_URL})
-    schema_version: Literal["1.0.1"] = Field("0.0.1")
+    schema_version: Literal["0.0.1"] = Field("0.0.1")
 
     name: str = Field(..., title="Name")
     developer_full_name: Optional[str] = Field(default=None, title="Name of developer")

From 6776d47a7b0e38d2dcf7bbca122c384460f910cc Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Wed, 16 Oct 2024 16:06:02 -0700
Subject: [PATCH 04/17] added cumulative evaluation

---
 src/aind_data_schema/core/model.py | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 714bda435..b23cd7f84 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -66,6 +66,14 @@ class ModelEvaluation(AindModel):
     notes: Optional[str] = Field(default=None, title="Notes")
 
 
+class CumulativeEvaluation(AindModel):
+    """Description of cumulative evaluation performances"""
+
+    evaluation_date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
+    performance: PerformanceScore = Field(..., title="Evaluation performance")
+    notes: Optional[str] = Field(default=None, title="Notes")
+
+
 class Model(AindCoreModel):
     """Description of an analysis model"""
 
@@ -75,7 +83,7 @@ class Model(AindCoreModel):
 
     name: str = Field(..., title="Name")
     developer_full_name: Optional[str] = Field(default=None, title="Name of developer")
-    developer_institution: Organization.ONE_OF = Field(default=None, title="Institute where developed")
+    developer_institution: Optional[Organization.ONE_OF] = Field(default=None, title="Institute where developed")
     modality: Modality.ONE_OF = Field(..., title="Modality")
     model_architecture: ModelArchitecture = Field(..., title="Model architecture")
     software: Software = Field(..., title="software")
@@ -83,4 +91,5 @@ class Model(AindCoreModel):
     limitations: Optional[str] = Field(default=None, title="Model limitations")
     training: List[ModelTraining] = Field(..., title="Training")
     evaluations: Optional[List[ModelEvaluation]] = Field(default=[], title="Evaluations")
+    cumulative_performance: Optional[List[CumulativeEvaluation]] = Field(default=[], title="Cumulative performance")
     notes: Optional[str] = Field(default=None, title="Notes")

From 05074660f8b101ed67d9481ace82b99f181a8ee1 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Thu, 17 Oct 2024 11:58:02 -0700
Subject: [PATCH 05/17] line length

---
 src/aind_data_schema/core/model.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index b23cd7f84..0f1b17727 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -51,7 +51,7 @@ class ModelTraining(AindModel):
     training_data: str = Field(..., title="Path to training data")
     training_data_description: Optional[str] = Field(default=None, title="Description of training data") 
     training_date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
-    validation_folds: int = Field(..., title="Validation folds") #is the validation methods x-fold? or are there other validations? Enum?
+    validation_folds: int = Field(..., title="Validation folds") #is the validation methods always x-fold? or Enum?
     performance: PerformanceScore = Field(..., title="Training performance")
     notes: Optional[str] = Field(default=None, title="Notes")
 

From b2a68e9b5723446d47f1e26edef6708b72ad4f83 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Fri, 18 Oct 2024 10:42:59 -0700
Subject: [PATCH 06/17] cleanup based on feedback from team

---
 src/aind_data_schema/core/model.py | 69 ++++++++----------------------
 1 file changed, 18 insertions(+), 51 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 0f1b17727..4340639ee 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -2,75 +2,42 @@
 
 from decimal import Decimal
 from enum import Enum
-from typing import List, Literal, Optional, Union
+from typing import Any, Optional, List, Literal
 
 from pydantic import Field
 
 from aind_data_schema_models.modalities import Modality
 from aind_data_schema_models.organizations import Organization
+from aind_data_schema_models.system_architecture import ModelBackbone
 
 from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AwareDatetimeWithDefault
 from aind_data_schema.components.devices import Software
 
 
-class Backbone(str, Enum):
-    """Types of network backbones"""
-
-    ALEXNET = "AlexNet"
-    RESNET = "ResNet"
-    VGGNET = "VGGNet"
-
-
 class ModelArchitecture(AindModel):
     """Description of model architecture"""
 
-    backbone: Backbone = Field(..., title="Backbone")
+    backbone: ModelBackbone = Field(..., title="Backbone")
     layers: int = Field(..., title="Layers")
     parameters: AindGenericType = Field(..., title="Parameters")
-    #notes?
-
-
-class ScoreStatistics(AindModel):
-    """Statistics for x-fold validation scores"""
-
-    mean: Decimal = Field(..., title="Mean")
-    std: Decimal = Field(..., title="Standard deviation")
-
-
-class PerformanceScore(AindModel):
-    """Description of performance metrics"""
-
-    precision: Union[Decimal, ScoreStatistics] = Field(..., title="Precision")
-    recall: Union[Decimal, ScoreStatistics] = Field(..., title="Recall")
-    f1_score: Union[Decimal, ScoreStatistics] = Field(..., title="F1 score")
-
-
-class ModelTraining(AindModel):
-    """Description of model training"""
-
-    training_data: str = Field(..., title="Path to training data")
-    training_data_description: Optional[str] = Field(default=None, title="Description of training data") 
-    training_date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
-    validation_folds: int = Field(..., title="Validation folds") #is the validation methods always x-fold? or Enum?
-    performance: PerformanceScore = Field(..., title="Training performance")
     notes: Optional[str] = Field(default=None, title="Notes")
 
 
-class ModelEvaluation(AindModel):
-    """Description of model evaluation"""
+class PerformanceMetric(AindModel):
+    """Description of a performance metric"""
 
-    evaluation_data: str = Field(..., title="Path to evaluation data")
-    evaluation_data_description: Optional[str] = Field(default=None, title="Description of evaluation data")
-    evaluation_date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
-    performance: PerformanceScore = Field(..., title="Evaluation performance")
-    notes: Optional[str] = Field(default=None, title="Notes")
+    name: str = Field(..., title="Metric name")
+    value: Any = Field(..., title="Metric value")
 
 
-class CumulativeEvaluation(AindModel):
-    """Description of cumulative evaluation performances"""
+class ModelEvaluation(AindModel):
+    """Description of model training"""
 
-    evaluation_date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
-    performance: PerformanceScore = Field(..., title="Evaluation performance")
+    data: str = Field(..., title="Path to training data")
+    data_description: Optional[str] = Field(default=None, title="Description of training data") 
+    date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
+    validation_folds: Optional[int] = Field(default=None, title="Validation folds")
+    performance: List[PerformanceMetric] = Field(..., title="Training performance")
     notes: Optional[str] = Field(default=None, title="Notes")
 
 
@@ -82,14 +49,14 @@ class Model(AindCoreModel):
     schema_version: Literal["0.0.1"] = Field("0.0.1")
 
     name: str = Field(..., title="Name")
-    developer_full_name: Optional[str] = Field(default=None, title="Name of developer")
+    license: str = Field(..., title="License")
+    developer_full_name: Optional[List[str]] = Field(default=None, title="Name of developer")
     developer_institution: Optional[Organization.ONE_OF] = Field(default=None, title="Institute where developed")
     modality: Modality.ONE_OF = Field(..., title="Modality")
     model_architecture: ModelArchitecture = Field(..., title="Model architecture")
-    software: Software = Field(..., title="software")
+    software: List[Software] = Field(..., title="software")
     direct_use: str = Field(..., title="Intended model use", description="Semantic description of intended use")
     limitations: Optional[str] = Field(default=None, title="Model limitations")
-    training: List[ModelTraining] = Field(..., title="Training")
+    training: List[ModelEvaluation] = Field(..., title="Training")
     evaluations: Optional[List[ModelEvaluation]] = Field(default=[], title="Evaluations")
-    cumulative_performance: Optional[List[CumulativeEvaluation]] = Field(default=[], title="Cumulative performance")
     notes: Optional[str] = Field(default=None, title="Notes")

From d38cf13244a81883a9492c23d42e80edec75d9f5 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Wed, 6 Nov 2024 16:37:08 -0800
Subject: [PATCH 07/17] small edits

---
 src/aind_data_schema/core/model.py | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 4340639ee..b97b51d20 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -12,13 +12,14 @@
 
 from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AwareDatetimeWithDefault
 from aind_data_schema.components.devices import Software
+from aind_data_schema.core.processing import DataProcess
 
 
 class ModelArchitecture(AindModel):
     """Description of model architecture"""
 
     backbone: ModelBackbone = Field(..., title="Backbone")
-    layers: int = Field(..., title="Layers")
+    layers: Optional[int] = Field(default=None, title="Layers")
     parameters: AindGenericType = Field(..., title="Parameters")
     notes: Optional[str] = Field(default=None, title="Notes")
 
@@ -33,11 +34,11 @@ class PerformanceMetric(AindModel):
 class ModelEvaluation(AindModel):
     """Description of model training"""
 
-    data: str = Field(..., title="Path to training data")
-    data_description: Optional[str] = Field(default=None, title="Description of training data") 
-    date: AwareDatetimeWithDefault = Field(..., title="Date trained") #not sure we need datetime
+    data: str = Field(..., title="Path to evaluation data")
+    data_description: Optional[str] = Field(default=None, title="Description of evaluation data") 
+    date: AwareDatetimeWithDefault = Field(..., title="Date") #not sure we need datetime
     validation_folds: Optional[int] = Field(default=None, title="Validation folds")
-    performance: List[PerformanceMetric] = Field(..., title="Training performance")
+    performance: List[PerformanceMetric] = Field(..., title="Evaluation performance")
     notes: Optional[str] = Field(default=None, title="Notes")
 
 
@@ -54,9 +55,9 @@ class Model(AindCoreModel):
     developer_institution: Optional[Organization.ONE_OF] = Field(default=None, title="Institute where developed")
     modality: Modality.ONE_OF = Field(..., title="Modality")
     model_architecture: ModelArchitecture = Field(..., title="Model architecture")
-    software: List[Software] = Field(..., title="software")
+    software: List[Software] = Field(..., title="Software")
     direct_use: str = Field(..., title="Intended model use", description="Semantic description of intended use")
     limitations: Optional[str] = Field(default=None, title="Model limitations")
-    training: List[ModelEvaluation] = Field(..., title="Training")
+    training: Optional[List[ModelEvaluation]] = Field(default=[], title="Training")
     evaluations: Optional[List[ModelEvaluation]] = Field(default=[], title="Evaluations")
     notes: Optional[str] = Field(default=None, title="Notes")

From a0ef7f3bfda167bf998e449e21ae46d006ffc407 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Sat, 16 Nov 2024 19:49:17 -0800
Subject: [PATCH 08/17] model validation

---
 src/aind_data_schema/core/model.py | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index b97b51d20..f8a97d53a 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -32,16 +32,21 @@ class PerformanceMetric(AindModel):
 
 
 class ModelEvaluation(AindModel):
-    """Description of model training"""
+    """Description of model evaluation"""
 
-    data: str = Field(..., title="Path to evaluation data")
+    data: Optional[str] = Field(default=None, title="Path to evaluation data")
     data_description: Optional[str] = Field(default=None, title="Description of evaluation data") 
-    date: AwareDatetimeWithDefault = Field(..., title="Date") #not sure we need datetime
-    validation_folds: Optional[int] = Field(default=None, title="Validation folds")
+    date: AwareDatetimeWithDefault = Field(..., title="Date")
     performance: List[PerformanceMetric] = Field(..., title="Evaluation performance")
     notes: Optional[str] = Field(default=None, title="Notes")
 
 
+class ModelTraining(ModelEvaluation):
+    """Description of model training"""
+
+    cross_validation_method: str = Field(..., title="Cross validation method")
+
+
 class Model(AindCoreModel):
     """Description of an analysis model"""
 
@@ -56,8 +61,8 @@ class Model(AindCoreModel):
     modality: Modality.ONE_OF = Field(..., title="Modality")
     model_architecture: ModelArchitecture = Field(..., title="Model architecture")
     software: List[Software] = Field(..., title="Software")
-    direct_use: str = Field(..., title="Intended model use", description="Semantic description of intended use")
+    intended_use: str = Field(..., title="Intended model use", description="Semantic description of intended use")
     limitations: Optional[str] = Field(default=None, title="Model limitations")
-    training: Optional[List[ModelEvaluation]] = Field(default=[], title="Training")
+    training: Optional[List[ModelTraining]] = Field(default=[], title="Training")
     evaluations: Optional[List[ModelEvaluation]] = Field(default=[], title="Evaluations")
     notes: Optional[str] = Field(default=None, title="Notes")

From 029efefab79739015822bbb198237220996d4b17 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Sat, 16 Nov 2024 19:51:37 -0800
Subject: [PATCH 09/17] using dataprocess

---
 src/aind_data_schema/core/model.py | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index f8a97d53a..30c1ed2e0 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -31,14 +31,11 @@ class PerformanceMetric(AindModel):
     value: Any = Field(..., title="Metric value")
 
 
-class ModelEvaluation(AindModel):
+class ModelEvaluation(DataProcess):
     """Description of model evaluation"""
 
-    data: Optional[str] = Field(default=None, title="Path to evaluation data")
     data_description: Optional[str] = Field(default=None, title="Description of evaluation data") 
-    date: AwareDatetimeWithDefault = Field(..., title="Date")
     performance: List[PerformanceMetric] = Field(..., title="Evaluation performance")
-    notes: Optional[str] = Field(default=None, title="Notes")
 
 
 class ModelTraining(ModelEvaluation):

From 28072506233ead82ccfff0a541eb8d90e703d068 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Sat, 16 Nov 2024 19:53:08 -0800
Subject: [PATCH 10/17] linting

---
 src/aind_data_schema/core/model.py | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 30c1ed2e0..4c44e9e5c 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -1,7 +1,5 @@
 """ schema describing an analysis model """
 
-from decimal import Decimal
-from enum import Enum
 from typing import Any, Optional, List, Literal
 
 from pydantic import Field
@@ -10,7 +8,7 @@
 from aind_data_schema_models.organizations import Organization
 from aind_data_schema_models.system_architecture import ModelBackbone
 
-from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AwareDatetimeWithDefault
+from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel
 from aind_data_schema.components.devices import Software
 from aind_data_schema.core.processing import DataProcess
 
@@ -34,7 +32,7 @@ class PerformanceMetric(AindModel):
 class ModelEvaluation(DataProcess):
     """Description of model evaluation"""
 
-    data_description: Optional[str] = Field(default=None, title="Description of evaluation data") 
+    data_description: Optional[str] = Field(default=None, title="Description of evaluation data")
     performance: List[PerformanceMetric] = Field(..., title="Evaluation performance")
 
 

From 47cfc63476007e35eeabdebef562104cb21d5b12 Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Sat, 16 Nov 2024 21:02:36 -0800
Subject: [PATCH 11/17] remove dataprocess and add test

---
 src/aind_data_schema/core/model.py | 10 ++--
 tests/test_model.py                | 96 ++++++++++++++++++++++++++++++
 2 files changed, 102 insertions(+), 4 deletions(-)
 create mode 100644 tests/test_model.py

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 4c44e9e5c..3ab05eb68 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -8,9 +8,8 @@
 from aind_data_schema_models.organizations import Organization
 from aind_data_schema_models.system_architecture import ModelBackbone
 
-from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel
+from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AwareDatetimeWithDefault
 from aind_data_schema.components.devices import Software
-from aind_data_schema.core.processing import DataProcess
 
 
 class ModelArchitecture(AindModel):
@@ -29,11 +28,14 @@ class PerformanceMetric(AindModel):
     value: Any = Field(..., title="Metric value")
 
 
-class ModelEvaluation(DataProcess):
+class ModelEvaluation(AindModel):
     """Description of model evaluation"""
 
-    data_description: Optional[str] = Field(default=None, title="Description of evaluation data")
+    data: Optional[str] = Field(default=None, title="Path to evaluation data")
+    data_description: Optional[str] = Field(default=None, title="Description of evaluation data") 
+    date: AwareDatetimeWithDefault = Field(..., title="Date")
     performance: List[PerformanceMetric] = Field(..., title="Evaluation performance")
+    notes: Optional[str] = Field(default=None, title="Notes")
 
 
 class ModelTraining(ModelEvaluation):
diff --git a/tests/test_model.py b/tests/test_model.py
new file mode 100644
index 000000000..999fd981e
--- /dev/null
+++ b/tests/test_model.py
@@ -0,0 +1,96 @@
+""" tests for Model """
+
+import datetime
+import unittest
+
+import pydantic
+
+from aind_data_schema_models.modalities import Modality
+from aind_data_schema_models.organizations import Organization
+from aind_data_schema_models.process_names import ProcessName
+from aind_data_schema_models.system_architecture import ModelBackbone
+
+from aind_data_schema.core.model import Model, ModelArchitecture, ModelEvaluation, ModelTraining, PerformanceMetric
+from aind_data_schema.components.devices import Software
+
+class ModelTests(unittest.TestCase):
+    """tests for model"""
+
+    def test_constructors(self):
+        """try building model"""
+
+        with self.assertRaises(pydantic.ValidationError):
+            Model()
+
+        now = datetime.datetime.now()
+
+        m = Model(
+            name="2024_01_01_ResNet18_SmartSPIM.h5",
+            license="CC-BY-4.0",
+            developer_full_name="Joe Schmoe",
+            developer_institution=Organization.AIND,
+            modality=Modality.SPIM,
+            model_architecture=ModelArchitecture(
+                backbone=ModelBackbone.RESNET,
+                layers=18,
+                parameters={
+                    "downsample": 1,
+                    "input_shape": [
+                    14,
+                    14,
+                    26
+                    ],
+                    "learning_rate": 0.0001,
+                    "train_test_split": 0.8,
+                    "batch_size": 32,
+                    "augmentation": True,
+                    "finetuning": True
+                },
+            ),
+            software=[Software(
+                name="tensorflow",
+                version="2.11.0",
+                )
+            ],
+            intended_use="Cell counting for 488 channel of SmartSPIM data",
+            limitations="Only trained on 488 channel",
+            training=[
+                ModelTraining(
+                    data="path to training set",
+                    data_description="description of training set",
+                    date=now,
+                    performance=[
+                        PerformanceMetric(
+                            name="precision",
+                            value=0.9
+                        ),
+                        PerformanceMetric(
+                            name="recall",
+                            value=0.85
+                        )
+                    ],
+                    cross_validation_method="5-fold"
+                )
+            ],
+            evaluations=[
+                ModelEvaluation(
+                    data="path to evaluation data",
+                    data_description="description of evaluation set",
+                    date=now,
+                    performance=[
+                        PerformanceMetric(
+                            name="precision",
+                            value="0.8"
+                        )
+                    ]
+                )
+            ]
+        )
+    
+        Model.model_validate_json(m.model_dump_json())
+
+        self.assertIsNotNone(m)
+
+
+if __name__ == "__main__":
+    unittest.main()
\ No newline at end of file

From b6d07b28feac0c3ba4dfa2b5b89e515c0820a33f Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Sat, 16 Nov 2024 21:05:24 -0800
Subject: [PATCH 12/17] linting

---
 src/aind_data_schema/core/model.py |  2 +-
 tests/test_model.py                | 13 +++++++------
 2 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 3ab05eb68..af696f66d 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -32,7 +32,7 @@ class ModelEvaluation(AindModel):
     """Description of model evaluation"""
 
     data: Optional[str] = Field(default=None, title="Path to evaluation data")
-    data_description: Optional[str] = Field(default=None, title="Description of evaluation data") 
+    data_description: Optional[str] = Field(default=None, title="Description of evaluation data")
     date: AwareDatetimeWithDefault = Field(..., title="Date")
     performance: List[PerformanceMetric] = Field(..., title="Evaluation performance")
     notes: Optional[str] = Field(default=None, title="Notes")
diff --git a/tests/test_model.py b/tests/test_model.py
index 999fd981e..f1af9ece5 100644
--- a/tests/test_model.py
+++ b/tests/test_model.py
@@ -7,12 +7,12 @@
 
 from aind_data_schema_models.modalities import Modality
 from aind_data_schema_models.organizations import Organization
-from aind_data_schema_models.process_names import ProcessName
 from aind_data_schema_models.system_architecture import ModelBackbone
 
 from aind_data_schema.core.model import Model, ModelArchitecture, ModelEvaluation, ModelTraining, PerformanceMetric
 from aind_data_schema.components.devices import Software
 
+
 class ModelTests(unittest.TestCase):
     """tests for model"""
 
@@ -36,9 +36,9 @@ def test_constructors(self):
                 parameters={
                     "downsample": 1,
                     "input_shape": [
-                    14,
-                    14,
-                    26
+                        14,
+                        14,
+                        26
                     ],
                     "learning_rate": 0.0001,
                     "train_test_split": 0.8,
@@ -86,11 +86,12 @@ def test_constructors(self):
                 )
             ]
         )
-    
+
         Model.model_validate_json(m.model_dump_json())
 
         self.assertIsNotNone(m)
 
 
 if __name__ == "__main__":
-    unittest.main()
\ No newline at end of file
+    unittest.main()
+    
\ No newline at end of file

From 3343a9622af3913ffcbc8c0ded38dc7e02ea11eb Mon Sep 17 00:00:00 2001
From: Saskia de Vries <sejdevries@gmail.com>
Date: Sat, 16 Nov 2024 21:06:26 -0800
Subject: [PATCH 13/17] more lint

---
 tests/test_model.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tests/test_model.py b/tests/test_model.py
index f1af9ece5..bf6c7d6c2 100644
--- a/tests/test_model.py
+++ b/tests/test_model.py
@@ -94,4 +94,3 @@ def test_constructors(self):
 
 if __name__ == "__main__":
     unittest.main()
-    
\ No newline at end of file

From d4c13bb5bcff35c5c9c90e3747dfcf17f6aa8051 Mon Sep 17 00:00:00 2001
From: Tom Chartrand <tom.chartrand@alleninstitute.org>
Date: Mon, 25 Nov 2024 13:50:45 -0800
Subject: [PATCH 14/17] fix: backward compatible tweaks to processing for
 broader usage

makes software_version, parameters optional, allows list for input_location
---
 src/aind_data_schema/core/processing.py | 11 ++++++-----
 tests/test_processing.py                |  8 +-------
 2 files changed, 7 insertions(+), 12 deletions(-)

diff --git a/src/aind_data_schema/core/processing.py b/src/aind_data_schema/core/processing.py
index 9a1e8edd3..0c0107107 100644
--- a/src/aind_data_schema/core/processing.py
+++ b/src/aind_data_schema/core/processing.py
@@ -1,7 +1,7 @@
 """ schema for processing """
 
 from enum import Enum
-from typing import List, Literal, Optional
+from typing import List, Literal, Optional, Union
 
 from aind_data_schema_models.process_names import ProcessName
 from aind_data_schema_models.units import MemoryUnit, UnitlessUnit
@@ -57,15 +57,16 @@ class DataProcess(AindModel):
     """Description of a single processing step"""
 
     name: ProcessName = Field(..., title="Name")
-    software_version: str = Field(..., description="Version of the software used", title="Version")
+    software_version: Optional[str] = Field(default=None, description="Version of the software used", title="Version")
     start_date_time: AwareDatetimeWithDefault = Field(..., title="Start date time")
     end_date_time: AwareDatetimeWithDefault = Field(..., title="End date time")
-    input_location: str = Field(..., description="Path to data inputs", title="Input location")
+    # allowing multiple input locations, to be replaced by CompositeData object in future
+    input_location: Union[str, List[str]] = Field(..., description="Path(s) to data inputs", title="Input location")
     output_location: str = Field(..., description="Path to data outputs", title="Output location")
     code_url: str = Field(..., description="Path to code repository", title="Code URL")
     code_version: Optional[str] = Field(default=None, description="Version of the code", title="Code version")
-    parameters: AindGenericType = Field(..., title="Parameters")
-    outputs: AindGenericType = Field(AindGeneric(), description="Output parameters", title="Outputs")
+    parameters: AindGenericType = Field(default=AindGeneric(), title="Parameters")
+    outputs: AindGenericType = Field(default=AindGeneric(), description="Output parameters", title="Outputs")
     notes: Optional[str] = Field(default=None, title="Notes", validate_default=True)
     resources: Optional[ResourceUsage] = Field(default=None, title="Process resource usage")
 
diff --git a/tests/test_processing.py b/tests/test_processing.py
index 81d7d9d8e..928ddff6f 100644
--- a/tests/test_processing.py
+++ b/tests/test_processing.py
@@ -37,10 +37,7 @@ def test_constructors(self):
             DataProcess(name="Other", notes="")
 
         expected_exception = (
-            "8 validation errors for DataProcess\n"
-            "software_version\n"
-            "  Field required [type=missing, input_value={'name': 'Other', 'notes': ''}, input_type=dict]\n"
-            f"    For further information visit https://errors.pydantic.dev/{PYD_VERSION}/v/missing\n"
+            "6 validation errors for DataProcess\n"
             "start_date_time\n"
             "  Field required [type=missing, input_value={'name': 'Other', 'notes': ''}, input_type=dict]\n"
             f"    For further information visit https://errors.pydantic.dev/{PYD_VERSION}/v/missing\n"
@@ -56,9 +53,6 @@ def test_constructors(self):
             "code_url\n"
             "  Field required [type=missing, input_value={'name': 'Other', 'notes': ''}, input_type=dict]\n"
             f"    For further information visit https://errors.pydantic.dev/{PYD_VERSION}/v/missing\n"
-            "parameters\n"
-            "  Field required [type=missing, input_value={'name': 'Other', 'notes': ''}, input_type=dict]\n"
-            f"    For further information visit https://errors.pydantic.dev/{PYD_VERSION}/v/missing\n"
             "notes\n"
             "  Value error, Notes cannot be empty if 'name' is Other. Describe the process name in the notes field."
             " [type=value_error, input_value='', input_type=str]\n"

From 91a3785c186de99f5cfa97ca00fe3794b5d43306 Mon Sep 17 00:00:00 2001
From: Tom Chartrand <tom.chartrand@alleninstitute.org>
Date: Mon, 25 Nov 2024 14:56:38 -0800
Subject: [PATCH 15/17] fix: use DataProcess in model training/eval

---
 src/aind_data_schema/core/model.py | 30 +++++++++-------
 tests/test_model.py                | 58 +++++++++++++++++++-----------
 2 files changed, 55 insertions(+), 33 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index af696f66d..9cea2b544 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -8,16 +8,18 @@
 from aind_data_schema_models.organizations import Organization
 from aind_data_schema_models.system_architecture import ModelBackbone
 
-from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AwareDatetimeWithDefault
+from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AindGeneric
 from aind_data_schema.components.devices import Software
+from aind_data_schema.core.processing import DataProcess, ProcessName
 
 
 class ModelArchitecture(AindModel):
     """Description of model architecture"""
 
-    backbone: ModelBackbone = Field(..., title="Backbone")
+    backbone: ModelBackbone = Field(..., title="Backbone", description="Core network architecture")
+    software: List[Software] = Field(default=[], title="Software frameworks")
     layers: Optional[int] = Field(default=None, title="Layers")
-    parameters: AindGenericType = Field(..., title="Parameters")
+    parameters: AindGenericType = Field(default=AindGeneric(), title="Parameters")
     notes: Optional[str] = Field(default=None, title="Notes")
 
 
@@ -28,20 +30,24 @@ class PerformanceMetric(AindModel):
     value: Any = Field(..., title="Metric value")
 
 
-class ModelEvaluation(AindModel):
+class ModelEvaluation(DataProcess):
     """Description of model evaluation"""
 
-    data: Optional[str] = Field(default=None, title="Path to evaluation data")
-    data_description: Optional[str] = Field(default=None, title="Description of evaluation data")
-    date: AwareDatetimeWithDefault = Field(..., title="Date")
+    name: ProcessName = Field(ProcessName.MODEL_EVALUATION, title="Process name")
     performance: List[PerformanceMetric] = Field(..., title="Evaluation performance")
-    notes: Optional[str] = Field(default=None, title="Notes")
 
 
-class ModelTraining(ModelEvaluation):
+class ModelTraining(DataProcess):
     """Description of model training"""
 
-    cross_validation_method: str = Field(..., title="Cross validation method")
+    name: ProcessName = Field(ProcessName.MODEL_TRAINING, title="Process name")
+    train_performance: List[PerformanceMetric] = Field(..., title="Training performance", description="Performance on training set")
+    test_performance: Optional[List[PerformanceMetric]] = Field(
+        default=None, title="Test performance", description="Performance on untrained data, evaluated during training"
+    )
+    test_data: Optional[str] = Field(
+        default=None, title="Test data", description="Path or cross-validation/split approach"
+    )
 
 
 class Model(AindCoreModel):
@@ -56,10 +62,10 @@ class Model(AindCoreModel):
     developer_full_name: Optional[List[str]] = Field(default=None, title="Name of developer")
     developer_institution: Optional[Organization.ONE_OF] = Field(default=None, title="Institute where developed")
     modality: Modality.ONE_OF = Field(..., title="Modality")
-    model_architecture: ModelArchitecture = Field(..., title="Model architecture")
-    software: List[Software] = Field(..., title="Software")
+    architecture: ModelArchitecture = Field(..., title="Model architecture")
     intended_use: str = Field(..., title="Intended model use", description="Semantic description of intended use")
     limitations: Optional[str] = Field(default=None, title="Model limitations")
+    pretrained_source_url: Optional[str] = Field(default=None, title="Pretrained source URL")
     training: Optional[List[ModelTraining]] = Field(default=[], title="Training")
     evaluations: Optional[List[ModelEvaluation]] = Field(default=[], title="Evaluations")
     notes: Optional[str] = Field(default=None, title="Notes")
diff --git a/tests/test_model.py b/tests/test_model.py
index bf6c7d6c2..d9dd0fba7 100644
--- a/tests/test_model.py
+++ b/tests/test_model.py
@@ -27,10 +27,11 @@ def test_constructors(self):
         m = Model(
             name="2024_01_01_ResNet18_SmartSPIM.h5",
             license="CC-BY-4.0",
-            developer_full_name="Joe Schmoe",
+            developer_full_name=["Joe Schmoe"],
             developer_institution=Organization.AIND,
             modality=Modality.SPIM,
-            model_architecture=ModelArchitecture(
+            pretrained_source_url="url pretrained weights are from",
+            architecture=ModelArchitecture(
                 backbone=ModelBackbone.RESNET,
                 layers=18,
                 parameters={
@@ -40,26 +41,23 @@ def test_constructors(self):
                         14,
                         26
                     ],
-                    "learning_rate": 0.0001,
-                    "train_test_split": 0.8,
-                    "batch_size": 32,
-                    "augmentation": True,
-                    "finetuning": True
                 },
+                software=[Software(
+                    name="tensorflow",
+                    version="2.11.0",
+                    )
+                ],
             ),
-            software=[Software(
-                name="tensorflow",
-                version="2.11.0",
-                )
-            ],
             intended_use="Cell counting for 488 channel of SmartSPIM data",
             limitations="Only trained on 488 channel",
             training=[
                 ModelTraining(
-                    data="path to training set",
-                    data_description="description of training set",
-                    date=now,
-                    performance=[
+                    input_location=["s3 path to eval 1", "s3 path to eval 2"],
+                    output_location="s3 path to trained model asset",
+                    code_url="url for training code repo",
+                    start_date_time=now,
+                    end_date_time=now,
+                    train_performance=[
                         PerformanceMetric(
                             name="precision",
                             value=0.9
@@ -69,18 +67,36 @@ def test_constructors(self):
                             value=0.85
                         )
                     ],
-                    cross_validation_method="5-fold"
+                    test_performance=[
+                        PerformanceMetric(
+                            name="precision",
+                            value=0.8
+                        ),
+                        PerformanceMetric(
+                            name="recall",
+                            value=0.8
+                        )
+                    ],
+                    test_data="4:1 train/test split",
+                    parameters={
+                        "learning_rate": 0.0001,
+                        "batch_size": 32,
+                        "augmentation": True,
+                    },
+                    notes="note on training data selection"
                 )
             ],
             evaluations=[
                 ModelEvaluation(
-                    data="path to evaluation data",
-                    data_description="description of evaluation set",
-                    date=now,
+                    input_location=["s3 path to eval 1", "s3 path to eval 2"],
+                    output_location="s3 path (output asset or trained model asset if no output)",
+                    code_url="url for evaluation code repo (or capsule?)",
+                    start_date_time=now,
+                    end_date_time=now,
                     performance=[
                         PerformanceMetric(
                             name="precision",
-                            value="0.8"
+                            value=0.8
                         )
                     ]
                 )

From 0a8e986eb0c9679fb935097b752244d6fdd295eb Mon Sep 17 00:00:00 2001
From: Tom Chartrand <tom.chartrand@alleninstitute.org>
Date: Mon, 25 Nov 2024 15:57:42 -0800
Subject: [PATCH 16/17] linting

---
 src/aind_data_schema/core/model.py | 11 +++----
 tests/test_model.py                | 47 +++++++++---------------------
 2 files changed, 19 insertions(+), 39 deletions(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index 9cea2b544..a215e8924 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -1,14 +1,13 @@
 """ schema describing an analysis model """
 
-from typing import Any, Optional, List, Literal
-
-from pydantic import Field
+from typing import Any, List, Literal, Optional
 
 from aind_data_schema_models.modalities import Modality
 from aind_data_schema_models.organizations import Organization
 from aind_data_schema_models.system_architecture import ModelBackbone
+from pydantic import Field
 
-from aind_data_schema.base import AindCoreModel, AindGenericType, AindModel, AindGeneric
+from aind_data_schema.base import AindCoreModel, AindGeneric, AindGenericType, AindModel
 from aind_data_schema.components.devices import Software
 from aind_data_schema.core.processing import DataProcess, ProcessName
 
@@ -41,7 +40,9 @@ class ModelTraining(DataProcess):
     """Description of model training"""
 
     name: ProcessName = Field(ProcessName.MODEL_TRAINING, title="Process name")
-    train_performance: List[PerformanceMetric] = Field(..., title="Training performance", description="Performance on training set")
+    train_performance: List[PerformanceMetric] = Field(
+        ..., title="Training performance", description="Performance on training set"
+    )
     test_performance: Optional[List[PerformanceMetric]] = Field(
         default=None, title="Test performance", description="Performance on untrained data, evaluated during training"
     )
diff --git a/tests/test_model.py b/tests/test_model.py
index d9dd0fba7..d57b2cf3a 100644
--- a/tests/test_model.py
+++ b/tests/test_model.py
@@ -4,13 +4,12 @@
 import unittest
 
 import pydantic
-
 from aind_data_schema_models.modalities import Modality
 from aind_data_schema_models.organizations import Organization
 from aind_data_schema_models.system_architecture import ModelBackbone
 
-from aind_data_schema.core.model import Model, ModelArchitecture, ModelEvaluation, ModelTraining, PerformanceMetric
 from aind_data_schema.components.devices import Software
+from aind_data_schema.core.model import Model, ModelArchitecture, ModelEvaluation, ModelTraining, PerformanceMetric
 
 
 class ModelTests(unittest.TestCase):
@@ -36,15 +35,12 @@ def test_constructors(self):
                 layers=18,
                 parameters={
                     "downsample": 1,
-                    "input_shape": [
-                        14,
-                        14,
-                        26
-                    ],
+                    "input_shape": [14, 14, 26],
                 },
-                software=[Software(
-                    name="tensorflow",
-                    version="2.11.0",
+                software=[
+                    Software(
+                        name="tensorflow",
+                        version="2.11.0",
                     )
                 ],
             ),
@@ -58,24 +54,12 @@ def test_constructors(self):
                     start_date_time=now,
                     end_date_time=now,
                     train_performance=[
-                        PerformanceMetric(
-                            name="precision",
-                            value=0.9
-                        ),
-                        PerformanceMetric(
-                            name="recall",
-                            value=0.85
-                        )
+                        PerformanceMetric(name="precision", value=0.9),
+                        PerformanceMetric(name="recall", value=0.85),
                     ],
                     test_performance=[
-                        PerformanceMetric(
-                            name="precision",
-                            value=0.8
-                        ),
-                        PerformanceMetric(
-                            name="recall",
-                            value=0.8
-                        )
+                        PerformanceMetric(name="precision", value=0.8),
+                        PerformanceMetric(name="recall", value=0.8),
                     ],
                     test_data="4:1 train/test split",
                     parameters={
@@ -83,7 +67,7 @@ def test_constructors(self):
                         "batch_size": 32,
                         "augmentation": True,
                     },
-                    notes="note on training data selection"
+                    notes="note on training data selection",
                 )
             ],
             evaluations=[
@@ -93,14 +77,9 @@ def test_constructors(self):
                     code_url="url for evaluation code repo (or capsule?)",
                     start_date_time=now,
                     end_date_time=now,
-                    performance=[
-                        PerformanceMetric(
-                            name="precision",
-                            value=0.8
-                        )
-                    ]
+                    performance=[PerformanceMetric(name="precision", value=0.8)],
                 )
-            ]
+            ],
         )
 
         Model.model_validate_json(m.model_dump_json())

From e59a2918bb59aad4d3ae3d988a1ed7641b86aeb1 Mon Sep 17 00:00:00 2001
From: Tom Chartrand <tom.chartrand@alleninstitute.org>
Date: Mon, 9 Dec 2024 12:53:11 -0800
Subject: [PATCH 17/17] wording change

---
 src/aind_data_schema/core/model.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/aind_data_schema/core/model.py b/src/aind_data_schema/core/model.py
index a215e8924..5de0259cd 100644
--- a/src/aind_data_schema/core/model.py
+++ b/src/aind_data_schema/core/model.py
@@ -44,7 +44,7 @@ class ModelTraining(DataProcess):
         ..., title="Training performance", description="Performance on training set"
     )
     test_performance: Optional[List[PerformanceMetric]] = Field(
-        default=None, title="Test performance", description="Performance on untrained data, evaluated during training"
+        default=None, title="Test performance", description="Performance on test data, evaluated during training"
     )
     test_data: Optional[str] = Field(
         default=None, title="Test data", description="Path or cross-validation/split approach"