import gradio as gr
import pandas as pd
from sklearn.linear_model import Lasso
from sklearn.model_selection import train_test_split
from algo import description, article

# Load the data
data = pd.read_csv('hearing.csv')
data = data.drop("AGE",axis=1)
data = data.dropna(axis=0)

# Split the data into features and target
X = data.drop('HEARING LOSS(100)/LESS', axis=1)
y = data['HEARING LOSS(100)/LESS']


# Split the data into training and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

# Train the model
model = Lasso(alpha=0.1,random_state=42)
model.fit(X_train, y_train)

def predict_hearing_loss(NAME,AGE,GENDER, ES, GENETICAL, MENARIKAM, PROBLEM_PRENATAL, PROBLEM_POSTNATAL, DURING_BIRTH, SIBLINGS_PROBLEM, LEFT_EAR, RIGHT_EAR):
    # Create a DataFrame from the inputs
    input_data = pd.DataFrame([{
        'GENDER': 1 if GENDER == 'Male' else 0,
        'GENETICAL': 1 if GENETICAL == 'Yes' else 0,
        'MENARIKAM': 1 if MENARIKAM == 'Yes' else 0,
        'PROBLEM_PRENATAL': 1 if PROBLEM_PRENATAL == 'Yes' else 0,
        'PROBLEM_POSTNATAL': 1 if PROBLEM_POSTNATAL == 'Yes' else 0,
        'DURING_BIRTH': 1 if DURING_BIRTH == 'Yes' else 0,
        'SIBLINGS_PROBLEM': 1 if SIBLINGS_PROBLEM == 'Yes' else 0,
        'LEFT_N': LEFT_EAR,
        'RIGHT_N': RIGHT_EAR
    }])
    # Make the prediction
    prediction = model.predict(input_data)

    rec = ""
    if prediction[0] >= 77:
        rec = "Severe hearing loss detected later. Immediate attention needed. Seek professional medical advice if not taken already and consider hearing aids or cochlear implants"
    elif prediction[0] >=50:
        rec = "Moderate hearing loss detected. Schedule regular hearing assessments and consider using hearing aids if necessary."
    else:
        rec = "Minor hearing loss detected. Keep monitoring the child's hearing and consult a doctor if there are any concerns."
        
    return round(prediction[0],2),rec

interface = gr.Interface(
    fn=predict_hearing_loss,
    # inputs=[
    #     gr.Textbox(label="Patient Id"),
    #     gr.Number(label="AGE"),
    #     gr.Radio(['Male', 'Female'], label="GENDER"),
    #     gr.Radio(['Yes', 'No'], label="GENETICAL"),
    #     gr.Radio(['Yes', 'No'], label="Consanguineous Marriages"),
    #     gr.Radio(['Yes', 'No'], label="PRENATAL PROBLEM "),
    #     gr.Radio(['Yes', 'No'], label="POSTNATAL PROBLEM"),
    #     gr.Radio(['Yes', 'No'], label="DURING BIRTH"),
    #     gr.Radio(['Yes', 'No'], label="SIBLINGS PROBLEM"),
    #     gr.Number(label="LEFT EAR"),
    #     gr.Number(label="RIGHT EAR")
    # ],
    inputs=[
        gr.Textbox(label="Child Identification number"),
        gr.Number(label="Age"),
        gr.Radio(['Male', 'Female'], label="Gender"),
        gr.Radio(['Poor', 'BPL', 'Middle Class', 'Upper Class'], label="Economic Status"),
        gr.Radio(['Yes', 'No'], label="Any Known Genetic Hearing Loss in Family?"),
        gr.Radio(['Yes', 'No'], label="Is the parent marriage a Consanguineous Marriage?"),
        gr.Radio(['Yes', 'No'], label="Any Prenatal Problems?"),
        gr.Radio(['Yes', 'No'], label="Any Postnatal Problems?"),
        gr.Radio(['Yes', 'No'], label="Any Complications During Birth?"),
        gr.Radio(['Yes', 'No'], label="Any Siblings with Hearing Problems?"),
        gr.Number(label="Left Ear Hearing Loss Severity (in dB)"),
        gr.Number(label="Right Ear Hearing Loss Severity (in dB)")
    ],
    outputs=[gr.Textbox(label="HEARING LOSS %", type="text"),gr.Textbox(label="Recommendation", type="text")],
    title="A Predictive Model to Find Hearing Loss in Children Using Machine Learning Algorithms",
    theme="gradio/monochrome",
    allow_flagging=False,
    css="footer {display: none !important}",
    article=article,
    description=description
)

interface.launch()