import streamlit as st
import joblib
import pandas as pd
from scipy.stats import yeojohnson
import inv_transform

# Title of the app
st.title("Event Budget Estimate")

# Add custom CSS for spacing between columns
st.markdown(
    """
    <style>
    .stColumn > div {
        padding: 10px; /* Adjust the value as needed to increase spacing */
    }
    h2 {
        font-size: 20px;
    }
    </style>
    """, unsafe_allow_html=True)

# Create columns with adjusted spacing
col1, col2 = st.columns([1, 1], gap="large")  # Adjust gap as needed

with col1:
    # Smaller column title
    #st.subheader("Event Data")
    st.markdown("### Event Data")
    
    # Input fields
    atnd_num = st.number_input("Estimate Number of Attendees", min_value=0, value=180)
    ppr_num = st.number_input("Estimate Paper Number", min_value=0, value=100)
    exh_num = st.number_input("Estimate Number of Exhibits", min_value=0, value=0)
    long = st.number_input("Conference Duration, Days", min_value=0, value=3)
    
    # Dropdown menus
    event_type = st.selectbox("Event Type", ["Conference", "Workshop", "Forum", "Other"])
    cntry = st.selectbox("Conference Location Country", ["USA", "China", "India", "Romania", "Other"])
    if cntry =="USA":
        loc_state = st.selectbox("Conference Location State", [ "CA", "NJ", "NY", "FL", "Other"], index=4)
    else:
        loc_state = "Other"

    # Checkboxes
    st.markdown("<h4>Keywords</h4>", unsafe_allow_html=True)
    kw_comp = st.checkbox("Computing")
    kw_sys = st.checkbox("Systems")
    kw_app = st.checkbox("Applications")
    kw_computational = st.checkbox("Computational")
    kw_wless = st.checkbox("Wireless")
    kw_mdl = st.checkbox("Modeling")
    kw_ntwk = st.checkbox("Networking")
    kw_des = st.checkbox("Design")
    kw_adv = st.checkbox("Advanced")
    kw_dist = st.checkbox("Distributed")
    

submit = st.button("Submit")

# If the submit button is clicked, show output in the second column (col2)
with col2:
    if submit:
        if cntry != "USA" and loc_state != "Other":
            st.error("Correct Event Location: Country and State!")
        else:
            
            try:
                p = True
                regressor = joblib.load("budget_prediction_model.joblib")
            except:
                p = False
              
            if not p:
                st.write('Check the model path')
                st.error("Model doesn't Exist!")
            else:
                atnd_num = atnd_num / 1.1
                ppr_num = ppr_num / 1.08
                
                data = pd.DataFrame([{'act_atnd_tot_atnd_num': atnd_num,
                                        'long_atnd_ratio':long/atnd_num, 
                                        'act_paprs_num': ppr_num, 
                                        'longevity': long,
                                        'papr_atnd_ratio':ppr_num/atnd_num, 
                                        'exh_num': exh_num,
                                        'exh_atnd_ratio': exh_num/atnd_num,
                                        'computing': int(kw_comp),
                                        'conf_loc_cntry_nm_USA': int(cntry == "USA"),
                                        'conf_loc_cntry_nm_India': int(cntry == "India"),
                                        'conf_evnt_typ_nm_Workshop': int(event_type == "Workshop"),
                                        'conf_loc_state_nm_CA': int(loc_state == "CA"),
                                        'conf_loc_state_nm_NJ': int(loc_state == "NJ"),
                                        'systems': int(kw_sys),
                                        'applications': int(kw_app),
                                        'computational': int(kw_computational),
                                        'conf_loc_cntry_nm_China': int(cntry == "China"),
                                        'wireless': int(kw_wless),
                                        'conf_loc_cntry_nm_Romania': int(cntry == "Romania"),
                                        'conf_loc_state_nm_FL': int(loc_state == "FL"),
                                        'modeling': int(kw_mdl), 
                                        'networking': int(kw_ntwk), 
                                        'design': int(kw_des),
                                        'conf_loc_state_nm_NY': int(loc_state == "NY"),
                                        'conf_evnt_typ_nm_Forum': int(event_type == "Forum"),
                                        'advanced': int(kw_adv),
                                        'distributed': int(kw_dist)
                                    
                }])
                
                
                lambdas = pd.read_csv('lambdas_yeojohnson.csv', header=None, index_col=0)
                lambdas = lambdas.to_dict()[1]
                
                for n in ['act_atnd_tot_atnd_num', 'exh_num', 'act_paprs_num']:
                    data[n] = yeojohnson(data[n], lambdas[n])
                data.longevity -= 1
        
                # Predict income
                income = regressor.predict(data)[0]
                income = inv_transform.inv_yeojohnson(income, lambdas['fin_inc_tot_amt'])
                
                reg_fees_inc = income * 0.74
                exh_inc = min(exh_num * 3000, income - reg_fees_inc)
                
                expenses = income * 0.833
                local_arr_exp = expenses * 0.331
                socl_funcs_exp = expenses * 0.383
                admintn_exp = expenses * 0.091
                audit_exp = max(expenses*0.007, 6000)
                
                # Display results
                st.markdown("### Predicted Budget*")
                st.markdown("<h4>Income Structure</h4>", unsafe_allow_html=True)
                st.write(f"**Total Income: ${round(round(income, -3))}**")
                st.write(f"Registration Fees Income: ${round(round(reg_fees_inc, -3))}")
                st.write(f"Exhibit Income: ${round(round(exh_inc, -3))}")
                st.markdown("<h4>Expenses Structure</h4>", unsafe_allow_html=True)
                st.write(f"**Total Expenses: ${round(round(expenses, -3))}**")
                st.write(f"Local Arrangement Amount: ${round(round(local_arr_exp, -3))}")
                st.write(f"Social Functions Amount: ${round(round(socl_funcs_exp, -3))}")
                st.write(f"Administration Amount: ${round(round(admintn_exp, -3))}")
                st.write(f"Audit Fees Amount: ${round(round(audit_exp, -3))}")
                st.write("")
                st.write("***The numbers are approximate and should be adjusted according to event needs**")