import gradio as gr
# import plotly.graph_objects as go
import pandas as pd
import numpy as np
from scipy.stats import multivariate_normal as mvn
import random


AIRTABLE_TOKEN = os.environ.get("AIRTABLE_TOKEN")
SUMMARY_TABLE_URL = os.environ.get("SUMMARY_TABLE_URL")

def send_to_airtable(data, url):
    headers = {
        "Authorization": f"Bearer {AIRTABLE_TOKEN}",
        "Content-Type": "application/json"
    }
    data = json.dumps(data)
    response = requests.post(url, headers=headers, data=data)
    print(response.text)
    return response.text

def log_final_decision(group_name, total_yield, injection, production, url):
    data = {"fields": {"group_name": group_name, "total_yield": total_yield, "injection": injection, "production": production}}
    return send_to_airtable(data, url)


df = pd.read_csv('volcs.csv')
cities = ["Serang", "Bogor", "Jogja", "Bandung", "Tasikmalaya", "Purwokerto", "Semarang", "Malang", "Banyuwangi", "Tarogong", "Temanggung"]

gammas = [0.9, 0.9, 0.93, 0.89, 0.81, 0.90, 0.98, 0.99, 0.89, 0.85, 0.80]
cities_lats = [-6.12028, -6.59444, -7.7956, -6.9175, -7.3268, -7.4214, -6.96667, -7.9839, -8.2186, -7.3178564, -6.3929]
cities_lons = [106.15028, 106.78917, 110.3644, 107.6191, 108.2142, 109.2341, 110.4208, 112.6315, 114.3695, 110.1779144, 105.8391]


sigma = 0.05
sigma_diag1 = 0.01
sigma_diag2 = 0.005
offset = 1
num_grid = 800
Sigma = [[sigma, sigma_diag1], [sigma_diag2, sigma]]

# min_lat, max_lat = df['Latitude'].min() - offset, df['Latitude'].max() + offset
# min_lon, max_lon = df['Longitude'].min() - offset, df['Longitude'].max() + offset

# lats = np.linspace(min_lat, max_lat, num_grid)
# lons = np.linspace(min_lon, max_lon, num_grid)
# LON, LAT = np.meshgrid(lons, lats)

# density = np.zeros(LAT.shape)
# for idx, row in df.iterrows():
#     lat, lon = row['Latitude'], row['Longitude']
#     gaussian = mvn(mean=[lat, lon], cov=Sigma)
#     density += gaussian.pdf(np.dstack((LAT, LON)))
    
# lon_points = LON.ravel().tolist()
# lat_points = LAT.ravel().tolist()
# density_points = density.ravel().tolist()

def get_density(city):
    idx = cities.index(city)
    lat, lon = cities_lats[idx], cities_lons[idx]
    density = 0
    for idx, row in df.iterrows():
        lat2, lon2 = row['Latitude'], row['Longitude']
        gaussian = mvn(mean=[lat2, lon2], cov=Sigma)
        density += gaussian.pdf([lat, lon])
    return density


def get_energy_yield(production, injection, group, length=30):
    density = get_density(production)
    gamma = gammas[cities.index(production)]
    dist = random.uniform(0.98, 1.02)
    
    yield_dict = {"Year": [], "yield": []}
    for i in range(1, length):
        yield_dict["Year"].append(i)
        yield_dict["yield"].append((10*density/dist) * gamma**i)
    df = pd.DataFrame(yield_dict)    
    
    plt = gr.ScatterPlot(
        value = df,
        x = "Year",
        y = "yield",
        title = "Energy Yield Over Time",
        x_title = "Year",
        y_title = "Energy Yield (MWh)"
    )
    total_yield = sum(yield_dict['yield'])
    total_text = f"Total Energy Yield for 30 years: {total_yield:.2f} MWh"
    
    #if group is empty, return warning as text label
    if group == "":
        total_text = "Please enter your group number"
        plt = gr.ScatterPlot()
        return total_text, plt
    else:
        log_final_decision(group, total_yield, injection, production, SUMMARY_TABLE_URL)
        return total_text, plt
    
    


# def draw_map():
#     fig = go.Figure()

#     fig.add_trace(go.Densitymapbox(
#         lat=lat_points,
#         lon=lon_points,
#         z=density_points,
#         radius=2,
#         colorscale='Jet',
#         opacity=0.5, 
#         hoverinfo="skip",
#     ))

#     fig.add_trace(go.Scattermapbox(
#         lat=cities_lats,
#         lon=cities_lons,
#         mode='markers',
#         marker=go.scattermapbox.Marker(color="black"),
#         text=cities,  
#         hoverinfo="text"  
#     ))


#     fig.update_layout(
#         title='Geothermal Potentials in Java, Indonesia (Old Data)',
#         mapbox_style="open-street-map",
#         mapbox=dict(
#             center=dict(lat=(min_lat + max_lat) / 2, lon=(min_lon + max_lon) / 2),
#             zoom=6
#         ),
#         margin=dict(l=0, r=0, t=0, b=0)
        
#     )


#     return fig

with gr.Blocks() as demo:   
    
    
    gr.Label("Geothermal Potential in Java, Indonesia")
    gr.HTML("<center><img src='https://analytics-project-simt-its.github.io/assets/img/geothermal_map.png'></center>")
    gr.HTML("Interactive map: <a href='https://analytics-project-simt-its.github.io/geothermal_map.html'>here</a>")
#     map = gr.Plot()     
    
    with gr.Row():        
        injection = gr.Radio(cities, label="Location", info="Choose injection well location"),
        production = gr.Radio(cities, label="Location", info="Choose production well location")
            
            
    with gr.Row():
        group = gr.Textbox(label="Group number", info="Enter your group number")
        btn = gr.Button("Build Geothermal Wells")
        
    with gr.Row():
        energy = gr.Label()
        energy_30 = gr.ScatterPlot()
            
        
    # demo.load(draw_map, [], map)
    demo.load()
    btn.click(
        get_energy_yield, 
        inputs = [production, injection, group],
        outputs = [energy, energy_30]
        )    
    
    
    

demo.launch()