Analyse la réutilisation des têtes de minifigs

lib/plots/head_reuse.py

@@ -0,0 +1,53 @@
+"""Visualisation de la réutilisation des têtes de minifigs."""
+
+import csv
+from pathlib import Path
+from typing import List
+
+import matplotlib.pyplot as plt
+
+from lib.filesystem import ensure_parent_dir
+
+
+def load_head_reuse(path: Path) -> List[dict]:
+    """Charge le CSV head_reuse."""
+    rows: List[dict] = []
+    with path.open() as csv_file:
+        reader = csv.DictReader(csv_file)
+        for row in reader:
+            rows.append(row)
+    return rows
+
+
+def format_label(row: dict) -> str:
+    """Formate le label affiché sur l'axe vertical."""
+    character = row["known_character"]
+    if character != "":
+        return f"{row['part_num']} — {character}"
+    return row["part_num"]
+
+
+def plot_head_reuse(path: Path, destination_path: Path, top: int = 30) -> None:
+    """Trace un bar chart horizontal mettant en avant les têtes exclusives ou rares."""
+    rows = load_head_reuse(path)
+    rows.sort(key=lambda r: (int(r["other_sets"]), -int(r["filtered_sets"]), r["part_num"]))
+    selected = rows[:top]
+    labels = [format_label(r) for r in selected]
+    filtered_counts = [int(r["filtered_sets"]) for r in selected]
+    other_counts = [int(r["other_sets"]) for r in selected]
+    positions = list(reversed(range(len(selected))))
+
+    fig, ax = plt.subplots(figsize=(12, 0.5 * len(selected) + 1.5))
+    ax.barh(positions, filtered_counts, color="#1f78b4", label="Sets filtrés")
+    ax.barh(positions, other_counts, left=filtered_counts, color="#b2df8a", label="Autres sets")
+    ax.set_yticks(positions)
+    ax.set_yticklabels(reversed(labels))
+    ax.set_xlabel("Nombre de sets contenant la tête")
+    ax.invert_yaxis()
+    ax.grid(axis="x", linestyle="--", alpha=0.4)
+    ax.legend()
+    fig.tight_layout()
+
+    ensure_parent_dir(destination_path)
+    fig.savefig(destination_path, dpi=150)
+    plt.close(fig)