import pickle
import pandas as pd
import math
import gradio as gr
#from sklearn.ensemble import RandomForestRegressor
import pandas as pd




#C:\Users\prebe\OneDrive\Projects\Doc_Summarizer\web\Gradio Apps\hospitalStay\model\hospitalStay.pkl

with open('hospitalStay.pkl', 'rb') as f:
    model = pickle.load(f)



default_values = {
    'rcount': 0.0,
    'gender': 0.0,
    'dialysisrenalendstage': False,
    'asthma': False,
    'irondef': False,
    'pneum': False,
    'substancedependence': False,
    'psychologicaldisordermajor': False,
    'depress': False,
    'psychother': False,
    'fibrosisandother': False,
    'malnutrition': False,
    'hemo': 0.0,
    'hematocrit': 11.9,
    'neutrophils': 9.4,
    'sodium': 135.885126,
    'glucose': 23.765383,
    'bloodureanitro': 12.0,
    'creatinine': 0.268453,
    'bmi': 29.798116,
    'pulse': 74.0,
    'respiration': 6.5,
    'secondarydiagnosisnonicd9': 1.0,
    'facid': 4.0
}


def predict_length_of_stay(rcount, gender, asthma, hematocrit, bmi):
    
    # Konverterer input verdier fra Gradio input
    if rcount < 0 or rcount > 5:
            return "Ugylding verdig for Antall innleggelser. Bruk et tall mellom 0 og 5"
    gender = 1.0 if gender == "Male" else 0.0
    
    input_values = default_values.copy()
    input_values.update({
        'rcount': rcount,
        'gender': gender,
        'asthma': asthma,
        'hematocrit': hematocrit,
        'bmi': bmi
    })
    
    df = pd.DataFrame([input_values])
    
    prediction = model.predict(df) 
    rounded_prediction = math.ceil(prediction[0])  
    return rounded_prediction


iface = gr.Interface(
    fn=predict_length_of_stay,
    title="Beregn lengden på et sykehusopphold til en pasient.",
    description="Modellen tar inn 25 variabler for å predikere lengden på sykehusoppholdet, men jeg har begreset denne demoen til å bare inneholde 5 input verdier.",
    inputs=[
        gr.components.Number(label="Antall innleggelser de siste 180 dager", default=0.0),
        gr.components.Dropdown(choices=["Male", "Female"], label="Gender", default="Male"),
        gr.components.Checkbox(label="Asthma", default=False),
        gr.components.Number(label="Hematocrit", default=11.9),
        gr.components.Number(label="BMI", default=29.798116),
    ],
    allow_flagging="never",
    outputs=gr.components.Textbox(label="Predicted Length of Stay (days)")
)

iface.launch()