ins-pricing 0.4.4__py3-none-any.whl → 0.5.0__py3-none-any.whl

ins_pricing/modelling/plotting/compat.py

@@ -0,0 +1,228 @@
+from __future__ import annotations
+
+from typing import Iterable, List, Mapping, Optional, Sequence, Tuple
+
+import numpy as np
+import pandas as pd
+
+from ins_pricing.modelling.plotting.common import PlotStyle, plt
+from ins_pricing.modelling.plotting.curves import (
+    plot_double_lift_curve,
+    plot_lift_curve,
+)
+
+
+def _safe_tag(value: str) -> str:
+    return "".join(ch if ch.isalnum() or ch in "-_." else "_" for ch in str(value))
+
+
+def _to_array(values: Sequence[float], name: str) -> np.ndarray:
+    if values is None:
+        raise ValueError(f"{name} is required.")
+    return np.asarray(values, dtype=float).reshape(-1)
+
+
+def _resolve_pred_pairs_from_df(
+    df: pd.DataFrame,
+    pred_cols: Optional[Sequence[str] | Mapping[str, object]],
+    pred_labels: Optional[Sequence[str]],
+) -> List[Tuple[str, np.ndarray]]:
+    if pred_cols is None:
+        raise ValueError("pred_cols is required when data is a DataFrame.")
+
+    pairs: List[Tuple[str, np.ndarray]] = []
+    if isinstance(pred_cols, Mapping):
+        for label, col in pred_cols.items():
+            if isinstance(col, str) and col in df.columns:
+                arr = df[col].to_numpy()
+            else:
+                arr = np.asarray(col, dtype=float).reshape(-1)
+            pairs.append((str(label), arr))
+        return pairs
+
+    if isinstance(pred_cols, str):
+        label = pred_cols if not pred_labels else str(pred_labels[0])
+        return [(label, df[pred_cols].to_numpy())]
+
+    labels = list(pred_labels) if pred_labels else [str(c) for c in pred_cols]
+    for col, label in zip(pred_cols, labels):
+        if isinstance(col, str) and col in df.columns:
+            arr = df[col].to_numpy()
+        else:
+            arr = np.asarray(col, dtype=float).reshape(-1)
+        pairs.append((str(label), arr))
+    return pairs
+
+
+def _resolve_pred_pairs_from_arrays(
+    preds: object,
+    pred_labels: Optional[Sequence[str]],
+) -> List[Tuple[str, np.ndarray]]:
+    if isinstance(preds, Mapping):
+        return [(str(k), np.asarray(v, dtype=float).reshape(-1)) for k, v in preds.items()]
+
+    if isinstance(preds, (list, tuple)):
+        labels = list(pred_labels) if pred_labels else [f"model_{idx + 1}" for idx in range(len(preds))]
+        return [
+            (str(label), np.asarray(arr, dtype=float).reshape(-1))
+            for label, arr in zip(labels, preds)
+        ]
+
+    return [("model_1", np.asarray(preds, dtype=float).reshape(-1))]
+
+
+def plot_lift_list(
+    data_or_preds: pd.DataFrame | Mapping[str, object] | Sequence[object],
+    pred_cols: Optional[Sequence[str] | Mapping[str, object]] = None,
+    actual: Optional[Sequence[float]] = None,
+    weight: Optional[Sequence[float]] = None,
+    *,
+    actual_col: str = "w_act",
+    weight_col: str = "weight",
+    pred_labels: Optional[Sequence[str]] = None,
+    n_bins: int = 10,
+    pred_weighted: bool = False,
+    actual_weighted: bool = True,
+    show: bool = False,
+    save_dir: Optional[str] = None,
+    filename_prefix: str = "lift",
+    style: Optional[PlotStyle] = None,
+) -> List[plt.Figure]:
+    """Compatibility helper: plot multiple lift curves in one call."""
+    if isinstance(data_or_preds, pd.DataFrame):
+        df = data_or_preds
+        pairs = _resolve_pred_pairs_from_df(df, pred_cols, pred_labels)
+        actual_arr = _to_array(df[actual_col], actual_col) if actual is None else _to_array(actual, "actual")
+        weight_arr = None
+        if weight is not None:
+            weight_arr = _to_array(weight, "weight")
+        elif weight_col in df.columns:
+            weight_arr = _to_array(df[weight_col], weight_col)
+    else:
+        pairs = _resolve_pred_pairs_from_arrays(data_or_preds, pred_labels)
+        actual_arr = _to_array(actual, "actual")
+        weight_arr = _to_array(weight, "weight") if weight is not None else None
+
+    figs: List[plt.Figure] = []
+    for label, pred in pairs:
+        save_path = None
+        if save_dir:
+            safe_dir = save_dir.rstrip("/\\")
+            save_path = f"{safe_dir}/{filename_prefix}_{_safe_tag(label)}.png"
+        fig = plot_lift_curve(
+            pred,
+            actual_arr,
+            weight_arr,
+            n_bins=n_bins,
+            title=f"{label} Lift Chart",
+            pred_label="Predicted",
+            act_label="Actual",
+            weight_label="Earned Exposure",
+            pred_weighted=pred_weighted,
+            actual_weighted=actual_weighted,
+            show=show,
+            save_path=save_path,
+            style=style,
+        )
+        figs.append(fig)
+    return figs
+
+
+def plot_dlift_list(
+    data_or_preds: pd.DataFrame | Mapping[str, object] | Sequence[object],
+    pred_cols: Optional[Sequence[str] | Mapping[str, object]] = None,
+    actual: Optional[Sequence[float]] = None,
+    weight: Optional[Sequence[float]] = None,
+    *,
+    actual_col: str = "w_act",
+    weight_col: str = "weight",
+    pred_labels: Optional[Sequence[str]] = None,
+    pairs: Optional[Iterable[Tuple[object, object]]] = None,
+    n_bins: int = 10,
+    pred_weighted: bool = False,
+    actual_weighted: bool = True,
+    show: bool = False,
+    save_dir: Optional[str] = None,
+    filename_prefix: str = "double_lift",
+    style: Optional[PlotStyle] = None,
+) -> List[plt.Figure]:
+    """Compatibility helper: plot double-lift curves for multiple model pairs."""
+    if isinstance(data_or_preds, pd.DataFrame):
+        df = data_or_preds
+        pairs_list = _resolve_pred_pairs_from_df(df, pred_cols, pred_labels)
+        actual_arr = _to_array(df[actual_col], actual_col) if actual is None else _to_array(actual, "actual")
+        weight_arr = None
+        if weight is not None:
+            weight_arr = _to_array(weight, "weight")
+        elif weight_col in df.columns:
+            weight_arr = _to_array(df[weight_col], weight_col)
+    else:
+        pairs_list = _resolve_pred_pairs_from_arrays(data_or_preds, pred_labels)
+        actual_arr = _to_array(actual, "actual")
+        weight_arr = _to_array(weight, "weight") if weight is not None else None
+
+    pred_map = {label: arr for label, arr in pairs_list}
+    labels = [label for label, _ in pairs_list]
+
+    pair_labels: List[Tuple[str, str]] = []
+    if pairs is None:
+        for idx, first in enumerate(labels):
+            for second in labels[idx + 1 :]:
+                pair_labels.append((first, second))
+    else:
+        for left, right in pairs:
+            if isinstance(left, int):
+                left_label = labels[left]
+            else:
+                left_label = str(left)
+            if isinstance(right, int):
+                right_label = labels[right]
+            else:
+                right_label = str(right)
+            pair_labels.append((left_label, right_label))
+
+    figs: List[plt.Figure] = []
+    for label1, label2 in pair_labels:
+        pred1 = pred_map.get(label1)
+        pred2 = pred_map.get(label2)
+        if pred1 is None or pred2 is None:
+            continue
+        save_path = None
+        if save_dir:
+            safe_dir = save_dir.rstrip("/\\")
+            tag = f"{_safe_tag(label1)}_vs_{_safe_tag(label2)}"
+            save_path = f"{safe_dir}/{filename_prefix}_{tag}.png"
+        fig = plot_double_lift_curve(
+            pred1,
+            pred2,
+            actual_arr,
+            weight_arr,
+            n_bins=n_bins,
+            title=f"Double Lift: {label1} vs {label2}",
+            label1=label1,
+            label2=label2,
+            pred1_weighted=pred_weighted,
+            pred2_weighted=pred_weighted,
+            actual_weighted=actual_weighted,
+            show=show,
+            save_path=save_path,
+            style=style,
+        )
+        figs.append(fig)
+    return figs
+
+
+class PlotUtils:
+    """Compatibility wrapper for legacy plotting helpers."""
+
+    plot_lift_curve = staticmethod(plot_lift_curve)
+    plot_double_lift_curve = staticmethod(plot_double_lift_curve)
+    plot_lift_list = staticmethod(plot_lift_list)
+    plot_dlift_list = staticmethod(plot_dlift_list)
+
+
+__all__ = [
+    "PlotUtils",
+    "plot_lift_list",
+    "plot_dlift_list",
+]