import joblib
import pandas as pd

# Load models and pipeline
model_rf = joblib.load("rf_model.joblib")
model_gb = joblib.load("gb_model.joblib")
model_lr = joblib.load("lr_model.joblib")
pipeline = joblib.load("pipeline.joblib")

# Create sample input
sample = pd.DataFrame([{
    'MOTHER_AGE_GRP': '25–34',
    'MOTHER_MARITALSTATUS_AT_BIRTH': 'Married',
    'MOTHER_RESID_COUNTY_TYPE': 'Urban',
    'SMOKING_DURING_PREG_IND': 'No',
    'SMOKING_BEFORE_PREG_IND': 'No',
    'NUM_BIRTHS_BY_MOTHER': 1,
    'VISITS_IN_1ST_TRIMESTER_IND': 'Yes'
}])

# Preprocess
X = pipeline.transform(sample)

# Predict
print("Logistic Regression Probability:", model_lr.predict_proba(X)[0][1])
print("Random Forest Probability:", model_rf.predict_proba(X)[0][1])
print("Gradient Boosting Probability:", model_gb.predict_proba(X)[0][1])