workbench 0.8.192__py3-none-any.whl → 0.8.197__py3-none-any.whl

workbench/utils/model_utils.py

@@ -113,9 +113,16 @@ def proximity_model_local(model: "Model"):
     fs = FeatureSet(model.get_input())
     id_column = fs.id_column
 
-    # Create the Proximity Model from our Training Data
-    df = model.training_view().pull_dataframe()
-    return Proximity(df, id_column, features, target, track_columns=features)
+    # Create the Proximity Model from both the full FeatureSet and the Model training data
+    full_df = fs.pull_dataframe()
+    model_df = model.training_view().pull_dataframe()
+
+    # Mark rows that are in the model
+    model_ids = set(model_df[id_column])
+    full_df["in_model"] = full_df[id_column].isin(model_ids)
+
+    # Create and return the Proximity Model
+    return Proximity(full_df, id_column, features, target, track_columns=features)
 
 
 def proximity_model(model: "Model", prox_model_name: str, track_columns: list = None) -> "Model":
@@ -165,9 +172,6 @@ def uq_model(model: "Model", uq_model_name: str, train_all_data: bool = False) -
     """
     from workbench.api import Model, ModelType, FeatureSet  # noqa: F401 (avoid circular import)
 
-    # Get the custom script path for the UQ model
-    script_path = get_custom_script_path("uq_models", "mapie.template")
-
     # Get Feature and Target Columns from the existing given Model
     features = model.features()
     target = model.target()
@@ -182,12 +186,25 @@ def uq_model(model: "Model", uq_model_name: str, train_all_data: bool = False) -
         description=f"UQ Model for {model.name}",
         tags=["uq", model.name],
         train_all_data=train_all_data,
-        custom_script=script_path,
         custom_args={"id_column": fs.id_column, "track_columns": [target]},
     )
     return uq_model
 
 
+def safe_extract_tarfile(tar_path: str, extract_path: str) -> None:
+    """
+    Extract a tarball safely, using data filter if available.
+
+    The filter parameter was backported to Python 3.8+, 3.9+, 3.10.13+, 3.11+
+    as a security patch, but may not be present in older patch versions.
+    """
+    with tarfile.open(tar_path, "r:gz") as tar:
+        if hasattr(tarfile, "data_filter"):
+            tar.extractall(path=extract_path, filter="data")
+        else:
+            tar.extractall(path=extract_path)
+
+
 def load_category_mappings_from_s3(model_artifact_uri: str) -> Optional[dict]:
     """
     Download and extract category mappings from a model artifact in S3.
@@ -206,9 +223,7 @@ def load_category_mappings_from_s3(model_artifact_uri: str) -> Optional[dict]:
         wr.s3.download(path=model_artifact_uri, local_file=local_tar_path)
 
         # Extract tarball
-        with tarfile.open(local_tar_path, "r:gz") as tar:
-            # Note: For 3.12+, can use filter="data" argument
-            tar.extractall(path=tmpdir)
+        safe_extract_tarfile(local_tar_path, tmpdir)
 
         # Look for category mappings in base directory only
         mappings_path = os.path.join(tmpdir, "category_mappings.json")

workbench/utils/xgboost_model_utils.py

@@ -3,17 +3,15 @@
 import logging
 import os
 import tempfile
-import tarfile
 import joblib
 import pickle
 import glob
 import awswrangler as wr
-from typing import Optional, List, Tuple
+from typing import Optional, List, Tuple, Any
 import hashlib
 import pandas as pd
 import numpy as np
 import xgboost as xgb
-from typing import Dict, Any
 from sklearn.model_selection import KFold, StratifiedKFold
 from sklearn.metrics import (
     precision_recall_fscore_support,
@@ -21,13 +19,14 @@ from sklearn.metrics import (
     mean_absolute_error,
     r2_score,
     median_absolute_error,
+    roc_auc_score,
 )
 from scipy.stats import spearmanr
 from sklearn.preprocessing import LabelEncoder
 
 # Workbench Imports
-from workbench.utils.model_utils import load_category_mappings_from_s3
-from workbench.utils.pandas_utils import convert_categorical_types
+from workbench.utils.model_utils import load_category_mappings_from_s3, safe_extract_tarfile
+from workbench.utils.pandas_utils import convert_categorical_types, expand_proba_column
 
 # Set up the log
 log = logging.getLogger("workbench")
@@ -50,9 +49,7 @@ def xgboost_model_from_s3(model_artifact_uri: str):
         wr.s3.download(path=model_artifact_uri, local_file=local_tar_path)
 
         # Extract tarball
-        with tarfile.open(local_tar_path, "r:gz") as tar:
-            # Note: For 3.12+, can use filter="data" argument
-            tar.extractall(path=tmpdir)
+        safe_extract_tarfile(local_tar_path, tmpdir)
 
         # Define model file patterns to search for (in order of preference)
         patterns = [
@@ -261,7 +258,7 @@ def leaf_stats(df: pd.DataFrame, target_col: str) -> pd.DataFrame:
     return result_df
 
 
-def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[str, Any], pd.DataFrame]:
+def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[pd.DataFrame, pd.DataFrame]:
     """
     Performs K-fold cross-validation with detailed metrics.
     Args:
@@ -269,10 +266,8 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
         nfolds: Number of folds for cross-validation (default is 5)
     Returns:
         Tuple of:
-            - Dictionary containing:
-                - folds: Dictionary of formatted strings for each fold
-                - summary_metrics: Summary metrics across folds
-            - DataFrame with columns: id, target, prediction (out-of-fold predictions for all samples)
+            - DataFrame with per-class metrics (and 'all' row for overall metrics)
+            - DataFrame with columns: id, target, prediction, and *_proba columns (for classifiers)
     """
     from workbench.api import FeatureSet
 
@@ -281,12 +276,22 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
     loaded_model = xgboost_model_from_s3(model_artifact_uri)
     if loaded_model is None:
         log.error("No XGBoost model found in the artifact.")
-        return {}, pd.DataFrame()
+        return pd.DataFrame(), pd.DataFrame()
 
     # Check if we got a full sklearn model or need to create one
     if isinstance(loaded_model, (xgb.XGBClassifier, xgb.XGBRegressor)):
-        xgb_model = loaded_model
-        is_classifier = isinstance(xgb_model, xgb.XGBClassifier)
+        is_classifier = isinstance(loaded_model, xgb.XGBClassifier)
+
+        # Get the model's hyperparameters and ensure enable_categorical=True
+        params = loaded_model.get_params()
+        params["enable_categorical"] = True
+
+        # Create new model with same params but enable_categorical=True
+        if is_classifier:
+            xgb_model = xgb.XGBClassifier(**params)
+        else:
+            xgb_model = xgb.XGBRegressor(**params)
+
     elif isinstance(loaded_model, xgb.Booster):
         # Legacy: got a booster, need to wrap it
         log.warning("Deprecated: Loaded model is a Booster, wrapping in sklearn model.")
@@ -297,7 +302,7 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
         xgb_model._Booster = loaded_model
     else:
         log.error(f"Unexpected model type: {type(loaded_model)}")
-        return {}, pd.DataFrame()
+        return pd.DataFrame(), pd.DataFrame()
 
     # Prepare data
     fs = FeatureSet(workbench_model.get_input())
@@ -308,10 +313,12 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
     target_col = workbench_model.target()
     feature_cols = workbench_model.features()
 
-    # Convert string features to categorical
+    # Convert string[python] to object, then to category for XGBoost compatibility
+    # This avoids XGBoost's issue with pandas 2.x string[python] dtype in categorical categories
     for col in feature_cols:
-        if df[col].dtype in ["object", "string"]:
-            df[col] = df[col].astype("category")
+        if pd.api.types.is_string_dtype(df[col]):
+            # Double conversion: string[python] -> object -> category
+            df[col] = df[col].astype("object").astype("category")
 
     X = df[feature_cols]
     y = df[target_col]
@@ -326,12 +333,12 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
         y_for_cv = y
 
     # Prepare KFold
+    # Note: random_state=42 seems to not actually give us reproducible results
     kfold = (StratifiedKFold if is_classifier else KFold)(n_splits=nfolds, shuffle=True, random_state=42)
 
     # Initialize results collection
     fold_metrics = []
-    predictions_df = pd.DataFrame({id_col: ids, target_col: y})  # Keep original values
-    # Note: 'prediction' column will be created automatically with correct dtype
+    predictions_df = pd.DataFrame({id_col: ids, target_col: y})
 
     # Perform cross-validation
     for fold_idx, (train_idx, val_idx) in enumerate(kfold.split(X, y_for_cv), 1):
@@ -346,6 +353,8 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
         val_indices = X_val.index
         if is_classifier:
             predictions_df.loc[val_indices, "prediction"] = label_encoder.inverse_transform(preds.astype(int))
+            y_proba = xgb_model.predict_proba(X_val)
+            predictions_df.loc[val_indices, "pred_proba"] = pd.Series(y_proba.tolist(), index=val_indices)
         else:
             predictions_df.loc[val_indices, "prediction"] = preds
 
@@ -353,10 +362,34 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
         if is_classifier:
             y_val_orig = label_encoder.inverse_transform(y_val)
             preds_orig = label_encoder.inverse_transform(preds.astype(int))
+
+            # Overall weighted metrics
             prec, rec, f1, _ = precision_recall_fscore_support(
                 y_val_orig, preds_orig, average="weighted", zero_division=0
             )
-            fold_metrics.append({"fold": fold_idx, "precision": prec, "recall": rec, "fscore": f1})
+
+            # Per-class F1
+            prec_per_class, rec_per_class, f1_per_class, _ = precision_recall_fscore_support(
+                y_val_orig, preds_orig, average=None, zero_division=0, labels=label_encoder.classes_
+            )
+
+            # ROC-AUC (overall and per-class)
+            roc_auc_overall = roc_auc_score(y_val, y_proba, multi_class="ovr", average="macro")
+            roc_auc_per_class = roc_auc_score(y_val, y_proba, multi_class="ovr", average=None)
+
+            fold_metrics.append(
+                {
+                    "fold": fold_idx,
+                    "precision": prec,
+                    "recall": rec,
+                    "f1": f1,
+                    "roc_auc": roc_auc_overall,
+                    "precision_per_class": prec_per_class,
+                    "recall_per_class": rec_per_class,
+                    "f1_per_class": f1_per_class,
+                    "roc_auc_per_class": roc_auc_per_class,
+                }
+            )
         else:
             spearman_corr, _ = spearmanr(y_val, preds)
             fold_metrics.append(
@@ -370,32 +403,67 @@ def cross_fold_inference(workbench_model: Any, nfolds: int = 5) -> Tuple[Dict[st
                 }
             )
 
-    # Calculate summary metrics (mean ± std)
+    # Calculate summary metrics
     fold_df = pd.DataFrame(fold_metrics)
-    metric_names = ["precision", "recall", "fscore"] if is_classifier else ["rmse", "mae", "medae", "r2", "spearmanr"]
-    summary_metrics = {metric: f"{fold_df[metric].mean():.3f} ±{fold_df[metric].std():.3f}" for metric in metric_names}
 
-    # Format fold results for display
-    formatted_folds = {}
-    for _, row in fold_df.iterrows():
-        fold_key = f"Fold {int(row['fold'])}"
-        if is_classifier:
-            formatted_folds[fold_key] = (
-                f"precision: {row['precision']:.3f}  " f"recall: {row['recall']:.3f}  " f"fscore: {row['fscore']:.3f}"
-            )
-        else:
-            formatted_folds[fold_key] = (
-                f"rmse: {row['rmse']:.3f}  "
-                f"mae: {row['mae']:.3f}  "
-                f"medae: {row['medae']:.3f}  "
-                f"r2: {row['r2']:.3f}  "
-                f"spearmanr: {row['spearmanr']:.3f}"
+    if is_classifier:
+        # Expand the *_proba columns into separate columns for easier handling
+        predictions_df = expand_proba_column(predictions_df, label_encoder.classes_)
+
+        # Build per-class metrics DataFrame
+        metric_rows = []
+
+        # Per-class rows
+        for idx, class_name in enumerate(label_encoder.classes_):
+            prec_scores = np.array([fold["precision_per_class"][idx] for fold in fold_metrics])
+            rec_scores = np.array([fold["recall_per_class"][idx] for fold in fold_metrics])
+            f1_scores = np.array([fold["f1_per_class"][idx] for fold in fold_metrics])
+            roc_auc_scores = np.array([fold["roc_auc_per_class"][idx] for fold in fold_metrics])
+
+            y_orig = label_encoder.inverse_transform(y_for_cv)
+            support = int((y_orig == class_name).sum())
+
+            metric_rows.append(
+                {
+                    "class": class_name,
+                    "precision": prec_scores.mean(),
+                    "recall": rec_scores.mean(),
+                    "f1": f1_scores.mean(),
+                    "roc_auc": roc_auc_scores.mean(),
+                    "support": support,
+                }
             )
 
-    # Build return dictionary
-    metrics_dict = {"summary_metrics": summary_metrics, "folds": formatted_folds}
+        # Overall 'all' row
+        metric_rows.append(
+            {
+                "class": "all",
+                "precision": fold_df["precision"].mean(),
+                "recall": fold_df["recall"].mean(),
+                "f1": fold_df["f1"].mean(),
+                "roc_auc": fold_df["roc_auc"].mean(),
+                "support": len(y_for_cv),
+            }
+        )
+
+        metrics_df = pd.DataFrame(metric_rows)
+
+    else:
+        # Regression metrics
+        metrics_df = pd.DataFrame(
+            [
+                {
+                    "rmse": fold_df["rmse"].mean(),
+                    "mae": fold_df["mae"].mean(),
+                    "medae": fold_df["medae"].mean(),
+                    "r2": fold_df["r2"].mean(),
+                    "spearmanr": fold_df["spearmanr"].mean(),
+                    "support": len(y_for_cv),
+                }
+            ]
+        )
 
-    return metrics_dict, predictions_df
+    return metrics_df, predictions_df
 
 
 def leave_one_out_inference(workbench_model: Any) -> pd.DataFrame:
@@ -440,10 +508,12 @@ def leave_one_out_inference(workbench_model: Any) -> pd.DataFrame:
     target_col = workbench_model.target()
     feature_cols = workbench_model.features()
 
-    # Convert string features to categorical
+    # Convert string[python] to object, then to category for XGBoost compatibility
+    # This avoids XGBoost's issue with pandas 2.x string[python] dtype in categorical categories
     for col in feature_cols:
-        if df[col].dtype in ["object", "string"]:
-            df[col] = df[col].astype("category")
+        if pd.api.types.is_string_dtype(df[col]):
+            # Double conversion: string[python] -> object -> category
+            df[col] = df[col].astype("object").astype("category")
 
     # Determine which samples to run LOO on
     if len(df) > 1000:

workbench/web_interface/components/model_plot.py

@@ -39,7 +39,13 @@ class ModelPlot(ComponentInterface):
             # Calculate the distance from the diagonal for each point
             target = model.target()
             df["error"] = abs(df["prediction"] - df[target])
-            return ScatterPlot().update_properties(df, color="error", regression_line=True)[0]
+            return ScatterPlot().update_properties(
+                df,
+                color="error",
+                regression_line=True,
+                x=target,
+                y="prediction",
+            )[0]
         else:
             return self.display_text(f"Model Type: {model.model_type}\n\n Awesome Plot Coming Soon!")
 

workbench/web_interface/components/plugin_unit_test.py

@@ -156,10 +156,13 @@ class PluginUnitTest:
         """Run the Dash server for the plugin, handling common errors gracefully."""
         while self.is_port_in_use(self.port):
             log.info(f"Port {self.port} is in use. Trying the next one...")
-            self.port += 1  # Increment the port number until an available one is found
+            self.port += 1
 
         log.info(f"Starting Dash server on port {self.port}...")
-        self.app.run(debug=True, use_reloader=False, port=self.port)
+        try:
+            self.app.run(debug=True, use_reloader=False, port=self.port)
+        except KeyboardInterrupt:
+            log.info("Shutting down Dash server...")
 
     @staticmethod
     def is_port_in_use(port):

workbench/web_interface/components/plugins/model_details.py

@@ -45,8 +45,6 @@ class ModelDetails(PluginInterface):
                 html.H5(children="Inference Metrics", style={"marginTop": "20px"}),
                 dcc.Dropdown(id=f"{self.component_id}-dropdown", className="dropdown"),
                 dcc.Markdown(id=f"{self.component_id}-metrics"),
-                html.H5(children="Cross Fold Metrics", style={"marginTop": "20px"}),
-                dcc.Markdown(id=f"{self.component_id}-cross-metrics", dangerously_allow_html=True),
             ],
         )
 
@@ -57,7 +55,6 @@ class ModelDetails(PluginInterface):
             (f"{self.component_id}-dropdown", "options"),
             (f"{self.component_id}-dropdown", "value"),
             (f"{self.component_id}-metrics", "children"),
-            (f"{self.component_id}-cross-metrics", "children"),
         ]
         self.signals = [(f"{self.component_id}-dropdown", "value")]
 
@@ -84,10 +81,9 @@ class ModelDetails(PluginInterface):
         # Populate the inference runs dropdown
         inference_runs, default_run = self.get_inference_runs()
         metrics = self.inference_metrics(default_run)
-        cross_metrics = self.cross_metrics()
 
         # Return the updated property values for the plugin
-        return [header, details, inference_runs, default_run, metrics, cross_metrics]
+        return [header, details, inference_runs, default_run, metrics]
 
     def register_internal_callbacks(self):
         @callback(
@@ -225,6 +221,7 @@ class ModelDetails(PluginInterface):
 
     def cross_metrics(self) -> str:
         # Get cross fold metrics if they exist
+        # Note: Currently not used since we show cross fold metrics in the dropdown
         model_name = self.current_model.name
         cross_fold_data = self.params.get(f"/workbench/models/{model_name}/inference/cross_fold", warn=False)
         if not cross_fold_data:
@@ -249,8 +246,13 @@ class ModelDetails(PluginInterface):
         if not inference_runs:
             return [], None
 
-        # Set "auto_inference" as the default, if that doesn't exist, set the first
-        default_inference_run = "auto_inference" if "auto_inference" in inference_runs else inference_runs[0]
+        # Default inference run (full_cross_fold if it exists, then auto_inference, then first)
+        if "full_cross_fold" in inference_runs:
+            default_inference_run = "full_cross_fold"
+        elif "auto_inference" in inference_runs:
+            default_inference_run = "auto_inference"
+        else:
+            default_inference_run = inference_runs[0]
 
         # Return the options for the dropdown and the selected value
         return inference_runs, default_inference_run

{workbench-0.8.192.dist-info → workbench-0.8.197.dist-info}/METADATA

@@ -1,9 +1,30 @@
 Metadata-Version: 2.4
 Name: workbench
-Version: 0.8.192
+Version: 0.8.197
 Summary: Workbench: A Dashboard and Python API for creating and deploying AWS SageMaker Model Pipelines
 Author-email: SuperCowPowers LLC <support@supercowpowers.com>
-License-Expression: MIT
+License: MIT License
+        
+        Copyright (c) 2021-2026 SuperCowPowers LLC
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+        
 Project-URL: Homepage, https://github.com/SuperCowPowers/workbench
 Keywords: SageMaker,Machine Learning,AWS,Python,Utilities
 Classifier: Development Status :: 4 - Beta

{workbench-0.8.192.dist-info → workbench-0.8.197.dist-info}/RECORD

@@ -1,12 +1,12 @@
 workbench/__init__.py,sha256=Kbp7lpicM-LH4ODhViZyas4uuvlDUzZQW8Dioks19Dc,1241
 workbench/algorithms/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 workbench/algorithms/dataframe/Readme.md,sha256=ZoCEi2BRJ4ZH0wlkFELTV_njIvYt7Wnhanuv4eoFluw,378
-workbench/algorithms/dataframe/__init__.py,sha256=n5dRcEXK6URKXeWoCf8XhMWxDalUdEWAK6YYBQfe8NY,455
+workbench/algorithms/dataframe/__init__.py,sha256=AXpwD4WYOk7odNS9vaKSO0DM-bMRuC93faq0JAltQ54,419
 workbench/algorithms/dataframe/data_source_eda.py,sha256=WgVL6tzBCw1tznQr8RQ6daQnTxQ0-DQUiMwbjztVMSU,1606
 workbench/algorithms/dataframe/feature_space_proximity.py,sha256=6RxzvbpLdDkHMm1D49Nv59SFcyYUj8bisd6_5EpBEGI,3515
-workbench/algorithms/dataframe/fingerprint_proximity.py,sha256=zqvk2WfY_nRFGC5bYcBbooGfSZsAhZBS8HNQ0GSMpwg,5838
+workbench/algorithms/dataframe/fingerprint_proximity.py,sha256=nGxfmYQ3bfMtvs90s4p7gaY9DN4gijdDU7R6B2lRHgo,5825
 workbench/algorithms/dataframe/projection_2d.py,sha256=zK4hc0OQrySmfcfFg8y0GxEL34uDNqvZL4OgttB9vRs,7834
-workbench/algorithms/dataframe/proximity.py,sha256=P8f3GHRhuc4QHj5KkKW0JMrHhIo2QdBiFG-JituTV1U,14633
+workbench/algorithms/dataframe/proximity.py,sha256=AkzcXvSEtbJdJwHYWoitaGOrTMUJZbrPzTvY8q-Fr8s,15383
 workbench/algorithms/dataframe/storage/aggregation.py,sha256=VuTb7A6Vh6IS5djZeItvOLnnEOlf7tzMQ8OaYIuftvU,2852
 workbench/algorithms/dataframe/storage/feature_resolution.py,sha256=w_iLf8EFTg7Jc5laH-bsq8MEtZVqcg05W-GihCqR-r4,9450
 workbench/algorithms/dataframe/storage/feature_spider.py,sha256=uIZ4JHIKuhpy08wBFReSrohb5DGxx8vGroHUbjPm1jE,14353
@@ -17,7 +17,7 @@ workbench/algorithms/graph/__init__.py,sha256=Hi2vBi-qw2OPLZnCu5qVTDdorRMOd3wNIt
 workbench/algorithms/graph/heavy/Readme.md,sha256=QzefnMevGXHBtaP3umzMNn2KEdNQMvJkJNbhkrUw7QE,96
 workbench/algorithms/graph/light/Readme.md,sha256=8kP_hRrRC6JtKtl5qyTwrV6gdO5rz6uq54c8WW1fTLg,122
 workbench/algorithms/graph/light/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-workbench/algorithms/graph/light/proximity_graph.py,sha256=cwPqyswX4l4my8yGyC3LDZ0Jc5ZkVFx3N6W8xFxfzEY,9651
+workbench/algorithms/graph/light/proximity_graph.py,sha256=CdHQ7-Tt_Ab7KrKTsnXR-p8OoX52RiiM6zD0OS4dGK0,9631
 workbench/algorithms/spark/Readme.md,sha256=18bPoFISlT3Ls5t1cBGb5N5Z6lOWyJupQkQxab1wcO4,615
 workbench/algorithms/sql/Readme.md,sha256=fzm4jQ-unJWT-fp5JhIjpApYSAqHUGSiuHE0eNfbF4A,685
 workbench/algorithms/sql/__init__.py,sha256=TbOZQwCfx6Tjc3pCCLCiM31wpCX26j5MBNQ6yG11EwY,462
@@ -31,11 +31,11 @@ workbench/api/__init__.py,sha256=kvrP70ypDOMdPGj_Eeftdh8J0lu_1qQVne6GXMkD4_E,102
 workbench/api/compound.py,sha256=kf5EaM5qjWwsZutcxqj9IC_MPnDV1uVHDMns9OA_GOo,2545
 workbench/api/data_source.py,sha256=Ngz36YZWxFfpJbmURhM1LQPYjh5kdpZNGo6_fCRePbA,8321
 workbench/api/df_store.py,sha256=Wybb3zO-jPpAi2Ns8Ks1-lagvXAaBlRpBZHhnnl3Lms,6131
-workbench/api/endpoint.py,sha256=RDQs78OFRSp9GYGHHph0gLtAfF4p3We96bzdJ4-9bzw,3926
+workbench/api/endpoint.py,sha256=spLse2UoAsZdu_ZxmAvMJX_aX-zutAsQ5_SPm9Xt-nA,3839
 workbench/api/feature_set.py,sha256=Yxei3tvWR4gSLcdJnNndux07dNeKNu1HKgsChJtHxEM,6633
 workbench/api/graph_store.py,sha256=LremJyPrQFgsHb7hxsctuCsoxx3p7TKtaY5qALHe6pc,4372
 workbench/api/meta.py,sha256=1_9989cPvf3hd3tA-83hLijOGNnhwXAF8aZF45adeDQ,8596
-workbench/api/model.py,sha256=coIl2QwbfyeVBM_Y2SaU587ioBh2TTDC9r9bOuYme9U,4605
+workbench/api/model.py,sha256=apDRhbnbQvSu-krQJm0zQWvbKNs7DUcslMwLKq4edHk,4442
 workbench/api/monitor.py,sha256=Cez89Uac7Tzt47FxkjoX-YDGccEhvBcxw3sZFtw4ud8,4506
 workbench/api/parameter_store.py,sha256=7BObkuATuP6C5AG_46kCWsmuCwuh1vgMJDBSN0gTkwM,4294
 workbench/api/pipeline.py,sha256=MSYGrDSXrRB_oQELtAlOwBfxSBTw3REAkHy5XBHau0Y,6261
@@ -54,9 +54,9 @@ workbench/core/artifacts/cached_artifact_mixin.py,sha256=ngqFLZ4cQx_TFouXZgXZQsv
 workbench/core/artifacts/data_capture_core.py,sha256=q8f79rRTYiZ7T4IQRWXl8ZvPpcvZyNxYERwvo8o0OQc,14858
 workbench/core/artifacts/data_source_abstract.py,sha256=5IRCzFVK-17cd4NXPMRfx99vQAmQ0WHE5jcm5RfsVTg,10619
 workbench/core/artifacts/data_source_factory.py,sha256=YL_tA5fsgubbB3dPF6T4tO0rGgz-6oo3ge4i_YXVC-M,2380
-workbench/core/artifacts/endpoint_core.py,sha256=VH-q-R4pfKyjCOXl2Gq1pxMNf4Ir0YdMw9YIHqd7CVU,51974
-workbench/core/artifacts/feature_set_core.py,sha256=6qOJoJ9_qwtyz4neFY6vMn73Ujjeut7E0dy_e8nYfSE,31462
-workbench/core/artifacts/model_core.py,sha256=vVERAD7cmYW8Y9tG8FmT7GMCWiudj8rJ4wfmucWrHUI,52287
+workbench/core/artifacts/endpoint_core.py,sha256=3e7GOWviFvVKU32bKzHGaDAk-wWxuglnZ0nPEXB0LXE,52200
+workbench/core/artifacts/feature_set_core.py,sha256=-4_FR4lLGHoeOkPRsAliEubVqZgai2RJJr1qJMo9Wao,36645
+workbench/core/artifacts/model_core.py,sha256=9UBuIm0xjFZlIlJ5YX945y5tGaAYz0YzlT5LWk_oG98,51156
 workbench/core/artifacts/monitor_core.py,sha256=M307yz7tEzOEHgv-LmtVy9jKjSbM98fHW3ckmNYrwlU,27897
 workbench/core/cloud_platform/cloud_meta.py,sha256=-g4-LTC3D0PXb3VfaXdLR1ERijKuHdffeMK_zhD-koQ,8809
 workbench/core/cloud_platform/aws/README.md,sha256=QT5IQXoUHbIA0qQ2wO6_2P2lYjYQFVYuezc22mWY4i8,97
@@ -102,7 +102,7 @@ workbench/core/transforms/features_to_features/__init__.py,sha256=47DEQpj8HBSa-_
 workbench/core/transforms/features_to_features/heavy/emr/Readme.md,sha256=YtQgCEQeKe0CQXQkhzMTYq9xOtCsCYb5P5LW2BmRKWQ,68
 workbench/core/transforms/features_to_features/heavy/glue/Readme.md,sha256=TuyCatWfoDr99zUwvOcxf-TqMkQzaMqXlj5nmFcRzfo,48
 workbench/core/transforms/features_to_model/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-workbench/core/transforms/features_to_model/features_to_model.py,sha256=qdE7sKMWPERpmDBsBOAAkZ94FzNZftDmUMNu0Mjq6ns,20115
+workbench/core/transforms/features_to_model/features_to_model.py,sha256=cExOKz6lrpSMQk2TgUDhiF9jxzAiOtrKSmasKlWVCO4,20110
 workbench/core/transforms/model_to_endpoint/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 workbench/core/transforms/model_to_endpoint/model_to_endpoint.py,sha256=TIYXvuK0s383PwJ4iS6fCRhuif6oIxsoWb4CpMGJjY4,6358
 workbench/core/transforms/pandas_transforms/__init__.py,sha256=xL4MT8-fZ1SFqDbTLc8XyxjupHtB1YR6Ej0AC2nwd7I,894
@@ -118,8 +118,8 @@ workbench/core/views/create_view.py,sha256=2Ykzb2NvJGoD4PP4k2Bka46GDog9iGG5SWnAc
 workbench/core/views/display_view.py,sha256=9K4O77ZnKOh93aMRhxcQJQ1lqScLhuJnU_tHtYZ_U4E,2598
 workbench/core/views/inference_view.py,sha256=9s70M0dFdGq0tWvzMZfgUK7EPKtuvcQhux0uyRZuuLM,3293
 workbench/core/views/pandas_to_view.py,sha256=20uCsnG2iMh-U1VxqVUUtnrWAY98SeuHjmfJK_wcq1I,6422
-workbench/core/views/training_view.py,sha256=Wk5fWvnoapngGCHEkqYzRvSNw2f-ZRvhsrzRMYGZfuE,6274
-workbench/core/views/view.py,sha256=QdUSGcrNzpzcwkSkutyptqrlXJZZReWzd2Uc6xQRd8A,13524
+workbench/core/views/training_view.py,sha256=7HwhbQhDBhT3Zo_gssS-b4eueJ0h9nqqT8YGFSuaEcU,9016
+workbench/core/views/view.py,sha256=8pZSVDhOFMnAh49ccvnvjQs0dWpiA5IeHaYrztRcqkM,13532
 workbench/core/views/view_utils.py,sha256=CwOlpqXpumCr6REi-ey7Qjz5_tpg-s4oWHmlOVu8POQ,12270
 workbench/core/views/storage/mdq_view.py,sha256=qf_ep1KwaXOIfO930laEwNIiCYP7VNOqjE3VdHfopRE,5195
 workbench/model_scripts/script_generation.py,sha256=dLxVRrvrrI_HQatJRAXta6UEbFFbkgITNvDJllQZyCM,7905
@@ -133,7 +133,7 @@ workbench/model_scripts/custom_models/meta_endpoints/example.py,sha256=hzOAuLhIG
 workbench/model_scripts/custom_models/network_security/Readme.md,sha256=Z2gtiu0hLHvEJ1x-_oFq3qJZcsK81sceBAGAGltpqQ8,222
 workbench/model_scripts/custom_models/proximity/Readme.md,sha256=RlMFAJZgAT2mCgDk-UwR_R0Y_NbCqeI5-8DUsxsbpWQ,289
 workbench/model_scripts/custom_models/proximity/feature_space_proximity.template,sha256=eOllmqB20BWtTiV53dgpIqXKtgSbPFDW_zf8PvM3oF0,4813
-workbench/model_scripts/custom_models/proximity/proximity.py,sha256=P8f3GHRhuc4QHj5KkKW0JMrHhIo2QdBiFG-JituTV1U,14633
+workbench/model_scripts/custom_models/proximity/proximity.py,sha256=AkzcXvSEtbJdJwHYWoitaGOrTMUJZbrPzTvY8q-Fr8s,15383
 workbench/model_scripts/custom_models/proximity/requirements.txt,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 workbench/model_scripts/custom_models/uq_models/Readme.md,sha256=UVpL-lvtTrLqwBeQFinLhd_uNrEw4JUlggIdUSDrd-w,188
 workbench/model_scripts/custom_models/uq_models/bayesian_ridge.template,sha256=ca3CaAk6HVuNv1HnPgABTzRY3oDrRxomjgD4V1ZDwoc,6448
@@ -141,22 +141,25 @@ workbench/model_scripts/custom_models/uq_models/ensemble_xgb.template,sha256=xlK
 workbench/model_scripts/custom_models/uq_models/gaussian_process.template,sha256=3nMlCi8nEbc4N-MQTzjfIcljfDQkUmWeLBfmd18m5fg,6632
 workbench/model_scripts/custom_models/uq_models/meta_uq.template,sha256=XTfhODRaHlI1jZGo9pSe-TqNsk2_nuSw0xMO2fKzDv8,14011
 workbench/model_scripts/custom_models/uq_models/ngboost.template,sha256=v1rviYTJGJnQRGgAyveXhOQlS-WFCTlc2vdnWq6HIXk,8241
-workbench/model_scripts/custom_models/uq_models/proximity.py,sha256=P8f3GHRhuc4QHj5KkKW0JMrHhIo2QdBiFG-JituTV1U,14633
+workbench/model_scripts/custom_models/uq_models/proximity.py,sha256=AkzcXvSEtbJdJwHYWoitaGOrTMUJZbrPzTvY8q-Fr8s,15383
 workbench/model_scripts/custom_models/uq_models/requirements.txt,sha256=fw7T7t_YJAXK3T6Ysbesxh_Agx_tv0oYx72cEBTqRDY,98
 workbench/model_scripts/custom_script_example/custom_model_script.py,sha256=T8aydawgRVAdSlDimoWpXxG2YuWWQkbcjBVjAeSG2_0,6408
 workbench/model_scripts/custom_script_example/requirements.txt,sha256=jWlGc7HH7vqyukTm38LN4EyDi8jDUPEay4n45z-30uc,104
 workbench/model_scripts/ensemble_xgb/ensemble_xgb.template,sha256=pWmuo-EVz0owvkRI-h9mUTYt1-ouyD-_yyQu6SQbYZ4,10350
 workbench/model_scripts/ensemble_xgb/requirements.txt,sha256=jWlGc7HH7vqyukTm38LN4EyDi8jDUPEay4n45z-30uc,104
+workbench/model_scripts/pytorch_model/generated_model_script.py,sha256=PG9MbcJrWcOxS9kl4tAZqfCjDKxvZpG4LGbHyPBRsmo,26231
 workbench/model_scripts/pytorch_model/pytorch.template,sha256=_gRp6DH294FLxF21UpSTq7s9RFfrLjViKvjXQ4yDfBQ,21999
 workbench/model_scripts/pytorch_model/requirements.txt,sha256=ICS5nW0wix44EJO2tJszJSaUrSvhSfdedn6FcRInGx4,181
 workbench/model_scripts/scikit_learn/requirements.txt,sha256=aVvwiJ3LgBUhM_PyFlb2gHXu_kpGPho3ANBzlOkfcvs,107
 workbench/model_scripts/scikit_learn/scikit_learn.template,sha256=QQvqx-eX9ZTbYmyupq6R6vIQwosmsmY_MRBPaHyfjdk,12586
-workbench/model_scripts/uq_models/mapie.template,sha256=2HIwB_658IsZiLIV1RViIZBIGgXxDsJPZinDUu8SchU,18961
+workbench/model_scripts/uq_models/generated_model_script.py,sha256=TGel0KUZyUmJvBhEXh7unpqq4I8KV-NDeVU8gLvGlvQ,27194
+workbench/model_scripts/uq_models/mapie.template,sha256=lq_kG9aRE_7_Or_jVfM4M5zkn9A1fEatneKI7_2zLQs,22784
 workbench/model_scripts/uq_models/requirements.txt,sha256=fw7T7t_YJAXK3T6Ysbesxh_Agx_tv0oYx72cEBTqRDY,98
+workbench/model_scripts/xgb_model/generated_model_script.py,sha256=jp4OCWdH_j5dc2ZxUMdlxSumZohbV02w9HWrCRt-8kc,18083
 workbench/model_scripts/xgb_model/requirements.txt,sha256=jWlGc7HH7vqyukTm38LN4EyDi8jDUPEay4n45z-30uc,104
 workbench/model_scripts/xgb_model/xgb_model.template,sha256=0uXknIEqgUaIFUfu2gfkxa3WHUr8HBBqBepGUTDvrhQ,17917
 workbench/repl/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-workbench/repl/workbench_shell.py,sha256=Gw7-tWCc6k7nwWrEimg-YuQOnWCsOzc2QMiX3Rt-Wpk,22131
+workbench/repl/workbench_shell.py,sha256=cbFaKaRyAJd79i8m6-3OIVK0f-W_wARwJbjxhqC11_s,22135
 workbench/resources/open_source_api.key,sha256=3S0OTblsmC0msUPdE_dbBmI83xJNmYscuwLJ57JmuOc,433
 workbench/resources/signature_verify_pub.pem,sha256=V3-u-3_z2PH-805ybkKvzDOBwAbvHxcKn0jLBImEtzM,272
 workbench/scripts/check_double_bond_stereo.py,sha256=p5hnL54Weq77ES0HCELq9JeoM-PyUGkvVSeWYF2dKyo,7776
@@ -213,7 +216,7 @@ workbench/utils/lambda_utils.py,sha256=7GhGRPyXn9o-toWb9HBGSnI8-DhK9YRkwhCSk_mNK
 workbench/utils/license_manager.py,sha256=lNE9zZIglmX3zqqCKBdN1xqTgHCEZgJDxavF6pdG7fc,6825
 workbench/utils/log_utils.py,sha256=7n1NJXO_jUX82e6LWAQug6oPo3wiPDBYsqk9gsYab_A,3167
 workbench/utils/markdown_utils.py,sha256=4lEqzgG4EVmLcvvKKNUwNxVCySLQKJTJmWDiaDroI1w,8306
-workbench/utils/model_utils.py,sha256=THs0GXkZdJLkk-sMxEo_wu7-ouGrJZdlN99yDUzO5xk,13515
+workbench/utils/model_utils.py,sha256=3hh4ZiZ58RZ8ahAGxz2BcvPMj3mxYIDCOw-ZiBIPlCI,14029
 workbench/utils/monitor_utils.py,sha256=kVaJ7BgUXs3VPMFYfLC03wkIV4Dq-pEhoXS0wkJFxCc,7858
 workbench/utils/pandas_utils.py,sha256=uTUx-d1KYfjbS9PMQp2_9FogCV7xVZR6XLzU5YAGmfs,39371
 workbench/utils/performance_utils.py,sha256=WDNvz-bOdC99cDuXl0urAV4DJ7alk_V3yzKPwvqgST4,1329
@@ -235,7 +238,7 @@ workbench/utils/workbench_cache.py,sha256=IQchxB81iR4eVggHBxUJdXxUCRkqWz1jKe5gxN
 workbench/utils/workbench_event_bridge.py,sha256=z1GmXOB-Qs7VOgC6Hjnp2DI9nSEWepaSXejACxTIR7o,4150
 workbench/utils/workbench_logging.py,sha256=WCuMWhQwibrvcGAyj96h2wowh6dH7zNlDJ7sWUzdCeI,10263
 workbench/utils/workbench_sqs.py,sha256=RwM80z7YWwdtMaCKh7KWF8v38f7eBRU7kyC7ZhTRuI0,2072
-workbench/utils/xgboost_model_utils.py,sha256=K0lbeb6LpS5oyN8cmK5Mfzz89n9q3p6ikvN39U9rvQw,22090
+workbench/utils/xgboost_model_utils.py,sha256=WXSDMt4Ae6ClihYanywzduMJ8FdaT_bWKS_1ut-Z28Q,24853
 workbench/utils/chem_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 workbench/utils/chem_utils/fingerprints.py,sha256=Qvs8jaUwguWUq3Q3j695MY0t0Wk3BvroW-oWBwalMUo,5255
 workbench/utils/chem_utils/misc.py,sha256=Nevf8_opu-uIPrv_1_0ubuFVVo2_fGUkMoLAHB3XAeo,7372
@@ -251,9 +254,9 @@ workbench/web_interface/components/component_interface.py,sha256=QCPWqiZLkVsAEzQ
 workbench/web_interface/components/correlation_matrix.py,sha256=Lv4vRta5-TdxBsu0G8Ea7hyyR3XyPes-k5AfL6qZWEc,6376
 workbench/web_interface/components/data_details_markdown.py,sha256=axDs6eXniglBmvFwIKjpJ5oyT-3D4FO9IcfA_cl-EJ8,9706
 workbench/web_interface/components/endpoint_metric_plots.py,sha256=H0cXuj9UQrrh_2JvRHtq7O8pMXFXKs7o9XpzySENylw,3441
-workbench/web_interface/components/model_plot.py,sha256=fgmOOxdrRkXN2-uNhY3K-pOsY34I-wcfy2MAnhr9ehY,2415
+workbench/web_interface/components/model_plot.py,sha256=Rojx_ZED4P9gvgeEsUm6xnwMNPoeOyn0evw45BWTITc,2536
 workbench/web_interface/components/plugin_interface.py,sha256=jGRq4igUTVXUT4sDqqsKKI2yjilV0ORNBQq6CjEWE84,9563
-workbench/web_interface/components/plugin_unit_test.py,sha256=UBZtGztLk2oJMDXfExfxkhHdmXr6ayv4NS0RpwGc8ro,7704
+workbench/web_interface/components/plugin_unit_test.py,sha256=Lx3HhIMHzrwDUYs2bADSFYzQq3sFHS9RyA415hyUOdc,7747
 workbench/web_interface/components/regression_plot.py,sha256=k18Bd0fcH7ig6kL5GqC_dINci3_YLle_fSEM32zXtzY,3342
 workbench/web_interface/components/violin_plots.py,sha256=3_T85hIs_R_WZpfFkSrqY2eYXmYzWsywDqsLhB7W1RQ,5320
 workbench/web_interface/components/experiments/dashboard_metric_plots.py,sha256=DPIw13tO9XOGxA6IeRPLgl-C3XUJ2N287JkSEg73Rjg,2984
@@ -266,7 +269,7 @@ workbench/web_interface/components/plugins/endpoint_details.py,sha256=0A7g_Lx5-3
 workbench/web_interface/components/plugins/generated_compounds.py,sha256=A6JGlkl7buZUugPK21YgufVFDRoGlHJowaqf8PAmz_s,8056
 workbench/web_interface/components/plugins/graph_plot.py,sha256=JFzuSH_CkEmlaLAgFpzmiEpS3sXov0ycnCfP0VLsK2g,14502
 workbench/web_interface/components/plugins/license_details.py,sha256=UyMSBGxEgdp3m9szDkDUAl_Ua8C5a4RNMdYpYCx354M,5497
-workbench/web_interface/components/plugins/model_details.py,sha256=BBvT3zdGHpRWz4V9SpNO0AFMPaExWMbNAE1_onfCGRI,10294
+workbench/web_interface/components/plugins/model_details.py,sha256=FiEFw3KJkbJ5igUotYbHLSeI0BEi7-umoLdvRapR2BI,10243
 workbench/web_interface/components/plugins/molecule_panel.py,sha256=xGCEI5af8F5lNId5eKUpetdQs_ahnIPdW6U7wKvbz2o,3515
 workbench/web_interface/components/plugins/molecule_viewer.py,sha256=xavixcu4RNzh6Nj_-3-XlK09DgpNx5jGmo3wEPNftiE,4529
 workbench/web_interface/components/plugins/pipeline_details.py,sha256=caiFIakHk-1dGGNW7wlio2X7iAm2_tCNbSjDzoRWGEk,5534
@@ -280,9 +283,9 @@ workbench/web_interface/page_views/main_page.py,sha256=X4-KyGTKLAdxR-Zk2niuLJB2Y
 workbench/web_interface/page_views/models_page_view.py,sha256=M0bdC7bAzLyIaE2jviY12FF4abdMFZmg6sFuOY_LaGI,2650
 workbench/web_interface/page_views/page_view.py,sha256=Gh6YnpOGlUejx-bHZAf5pzqoQ1H1R0OSwOpGhOBO06w,455
 workbench/web_interface/page_views/pipelines_page_view.py,sha256=v2pxrIbsHBcYiblfius3JK766NZ7ciD2yPx0t3E5IJo,2656
-workbench-0.8.192.dist-info/licenses/LICENSE,sha256=z4QMMPlLJkZjU8VOKqJkZiQZCEZ--saIU2Z8-p3aVc0,1080
-workbench-0.8.192.dist-info/METADATA,sha256=dlskxjoY82E7I0Z-JwO802MdU6hTmk5eWD5hwn1ahq8,9261
-workbench-0.8.192.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-workbench-0.8.192.dist-info/entry_points.txt,sha256=o7ohD4D2oygnHp7i9-C0LfcHDuPW5Tv0JXGAg97DpGk,413
-workbench-0.8.192.dist-info/top_level.txt,sha256=Dhy72zTxaA_o_yRkPZx5zw-fwumnjGaeGf0hBN3jc_w,10
-workbench-0.8.192.dist-info/RECORD,,
+workbench-0.8.197.dist-info/licenses/LICENSE,sha256=RTBoTMeEwTgEhS-n8vgQ-VUo5qig0PWVd8xFPKU6Lck,1080
+workbench-0.8.197.dist-info/METADATA,sha256=SBIr8UPVmyTRIItzlTYkNqGDk1zxXLLEb4XpbyO7bn0,10495
+workbench-0.8.197.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+workbench-0.8.197.dist-info/entry_points.txt,sha256=o7ohD4D2oygnHp7i9-C0LfcHDuPW5Tv0JXGAg97DpGk,413
+workbench-0.8.197.dist-info/top_level.txt,sha256=Dhy72zTxaA_o_yRkPZx5zw-fwumnjGaeGf0hBN3jc_w,10
+workbench-0.8.197.dist-info/RECORD,,