make_scaler.py

import joblib
import numpy as np
from sklearn.preprocessing import MinMaxScaler

# EXACT Feature List from your Frontend
# We define [Min, Max] based on your specific slider values
ranges = [
    # 1-2: Demographics
    [1, 100],      # Age
    [0, 1],        # Gender (0=Female, 1=Male)

    # 3-11: RBC & Iron
    [5, 20],       # HGB
    [2, 8],        # RBC Count
    [15, 60],      # HCT
    [50, 120],     # MCV
    [15, 40],      # MCH
    [28, 40],      # MCHC
    [10, 30],      # RDW-CV
    [30, 60],      # RDW-SD
    [10, 500],     # Ferritin

    # 12-18: WBC
    [2000, 20000], # WBC Count
    [0, 100],      # Neutrophils %
    [0, 100],      # Lymphocytes %
    [0, 100],      # Monocytes %
    [0, 100],      # Eosinophils %
    [0, 10],       # Basophils %
    [0, 10],       # NLR

    # 19-22: Platelets
    [50000, 600000], # Platelets
    [5, 15],       # MPV
    [0, 1],        # PCT
    [0, 30],       # PDW

    # 23-25: History
    [0, 1],        # History_Anemia (0=No, 1=Yes)
    [0, 1],        # Substance_Use
    [0, 1]         # Chronic_Illness
]

print(f"✅ Configured for {len(ranges)} features.")

# Generate mock data within these ranges to "fit" the scaler
np.random.seed(42)
mock_data = []
for r in ranges:
    # Create 1000 random patients within your specific min/max limits
    column = np.random.uniform(r[0], r[1], 1000)
    mock_data.append(column)

# Transpose to shape (1000 rows, 25 columns)
mock_data = np.array(mock_data).T 

print("⚙️  Fitting Scaler to your specific Feature Map...")
scaler = MinMaxScaler()
scaler.fit(mock_data)

# Save
joblib.dump(scaler, "scaler.pkl")
print("✅ 'scaler.pkl' created! It now understands values like 'WBC = 7000'.")