plot_mmse.py

import pandas as pd
import matplotlib.pyplot as plt
from pathlib import Path

# ----------------------------
# Paths
# ----------------------------
EXCEL_FILE = Path("/auto/home/users/v/l/vlahy/excel/TAU_Dg_neuro_complet_edit.xlsx")
SHEET_NAME = "DATA"   # adapte si besoin
OUT_DIR = Path("/auto/home/users/v/l/vlahy/test/mmse_longitudinal")
OUT_DIR.mkdir(parents=True, exist_ok=True)

# ----------------------------
# 1) Read & clean Excel
# ----------------------------
df = pd.read_excel(EXCEL_FILE, sheet_name=SHEET_NAME)

# garder uniquement les lignes patients
df = df[df["ID"].astype(str).str.isdigit()].copy()

df["ID"] = df["ID"].astype(int)
df["Visit_number"] = pd.to_numeric(df["Visit_number"], errors="coerce")
df["MMSE"] = pd.to_numeric(df["MMSE"], errors="coerce")

# garder les lignes valides
df = df.dropna(subset=["Visit_number", "MMSE"]).copy()

# option : limiter les visits si nécessaire
df = df[df["Visit_number"].isin([0, 1, 2, 3, 4, 5])].copy()

df = df.sort_values(["ID", "Visit_number"])

print("N patients =", df["ID"].nunique())

# ----------------------------
# 2) Spaghetti plot (all patients)
# ----------------------------
plt.figure(figsize=(8, 6))

for pid, g in df.groupby("ID"):
    if len(g) < 2:
        continue  # pas de ligne si une seule visite

    plt.plot(
        g["Visit_number"],
        g["MMSE"],
        color="gray",
        alpha=0.25,
        linewidth=1
    )

plt.xlabel("Visit number")
plt.ylabel("MMSE")
plt.title("MMSE longitudinal trajectories (all patients)")
plt.grid(alpha=0.3)

plt.savefig(
    OUT_DIR / "mmse_spaghetti_all_patients.png",
    dpi=200,
    bbox_inches="tight"
)
plt.close()

print("Saved spaghetti plot for all patients")