test3.py

import pandas as pd
import xgboost as xgb
import numpy as np

# Load data and model
df = pd.read_csv("data/processed/app_features.csv")

mv_rename_map = {col: 'market_value_in_eur' for col in df.columns if 'market' in col.lower() and 'value' in col.lower()}
if mv_rename_map:
    df.rename(columns=mv_rename_map, inplace=True)
df = df.loc[:, ~df.columns.duplicated()].copy()

model = xgb.XGBRegressor()
model.load_model("fairvalue_xgboost.json")
expected_cols = model.feature_names_in_

player_data = df.median(numeric_only=True).to_frame().T

# Simulate what api/main.py does
player_data['Contract_Years_Left'] = 2.5
player_data['Age'] = 28
player_data['market_value_in_eur'] = (120 * 1_000_000) / 0.85

X_infer = player_data.reindex(columns=expected_cols, fill_value=0)

preds = model.predict(X_infer)
log_pv = preds[0]
baseline_pv = np.expm1(log_pv)

print(f"Log PV: {log_pv}")
print(f"Baseline PV (Euros): {baseline_pv}")
print(f"Baseline PV_m (Millions): {baseline_pv / 1_000_000}")