Harden Great Tables renderer

Author: uriahf

src/rtichoke/performance_table_great_tables.py

@@ -22,8 +22,8 @@ def render_performance_table_great_tables(
 ) -> GT:
     """Render prepared performance data as a Great Tables table.
 
-    This renderer is intended for Marimo and other HTML notebook/document
-    contexts. Reactable remains available when expandable row details are wanted.
+    This renderer is intended for Marimo and HTML/Quarto contexts. Reactable
+    remains available when expandable confusion-matrix row details are wanted.
     """
     if performance_data.is_empty():
         raise ValueError("performance_data must contain at least one row")
@@ -33,11 +33,11 @@ def render_performance_table_great_tables(
         raise ValueError("performance_data must contain exactly one stratification")
     stratified_by = stratifications[0]
 
-    columns = [
+    source_columns = [
         "reference_group", "chosen_cutoff", "sensitivity", "specificity", "ppv",
         "npv", "lift", "predicted_positives", "net_benefit", "ppcr",
     ]
-    data = performance_data.select([c for c in columns if c in performance_data.columns])
+    data = performance_data.select([c for c in source_columns if c in performance_data.columns])
     if "reference_group" in data.columns:
         data = data.rename({"reference_group": "Model"})
 
@@ -50,40 +50,50 @@ def render_performance_table_great_tables(
     if sort_columns:
         data = data.sort(sort_columns)
 
-    # Match the R table's combined predicted-positive display.
+    # Match the R table's count + percentage display for predicted positives.
+    ppcr_percent = (pl.col("ppcr") * 100).round(2).cast(pl.String)
     data = data.with_columns(
-        pl.format(
-            "{} ({:.2f}%)",
-            pl.col("predicted_positives"),
-            pl.col("ppcr") * 100,
+        pl.concat_str(
+            [
+                pl.col("predicted_positives").cast(pl.String),
+                pl.lit(" ("),
+                ppcr_percent,
+                pl.lit("%)"),
+            ]
         ).alias("Predicted Positives")
     ).drop(["predicted_positives", "ppcr"])
 
     if stratified_by != "probability_threshold" and "net_benefit" in data.columns:
         data = data.drop("net_benefit")
 
-    metric_columns = [c for c in ["sensitivity", "specificity", "ppv", "npv", "lift", "net_benefit"] if c in data.columns]
-
-    table = (
-        GT(data)
-        .cols_label(
-            Model="Model",
-            Threshold="Probability Threshold",
-            sensitivity="Sens",
-            specificity="Spec",
-            ppv="PPV",
-            npv="NPV",
-            lift="Lift",
-            net_benefit="Net Benefit",
-        )
-        .tab_spanner(label="Performance Metrics", columns=metric_columns)
-        .fmt_number(columns=[c for c in metric_columns if c != "lift"], decimals=2)
-        .fmt_number(columns=[c for c in ["lift"] if c in data.columns], decimals=2)
+    metric_columns = [
+        c for c in ["sensitivity", "specificity", "ppv", "npv", "lift", "net_benefit"]
+        if c in data.columns
+    ]
+    labels = {
+        "Model": "Model",
+        "Threshold": "Probability Threshold",
+        "sensitivity": "Sens",
+        "specificity": "Spec",
+        "ppv": "PPV",
+        "npv": "NPV",
+        "lift": "Lift",
+        "net_benefit": "Net Benefit",
+        "Predicted Positives": "Predicted Positives",
+    }
+    labels = {column: label for column, label in labels.items() if column in data.columns}
+
+    table = GT(data).cols_label(cases=labels)
+    if metric_columns:
+        table = table.tab_spanner(label="Performance Metrics", columns=metric_columns)
+        table = table.fmt_number(columns=metric_columns, decimals=2)
+
+    # Keep model identification visually consistent with rtichoke's palette.
+    models = (
+        data.get_column("Model").unique(maintain_order=True).to_list()
+        if "Model" in data.columns
+        else []
     )
-
-    # Great Tables is presentation-first, so use gentle cell fills rather than
-    # recreating Reactable's JS data bars. Keep model colors consistent with R.
-    models = data.get_column("Model").unique(maintain_order=True).to_list() if "Model" in data.columns else []
     for index, model in enumerate(models):
         table = table.tab_style(
             style=[