Spaces:
Runtime error
Runtime error
File size: 14,002 Bytes
14f4b1a 49cf140 14f4b1a 9db4f31 14f4b1a 9db4f31 14f4b1a 9db4f31 14f4b1a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 |
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
import pickle
from datetime import datetime, timedelta
import tensorflow as tf
from sklearn.preprocessing import MinMaxScaler
import json
import requests
import gradio as gr
import os.path
import matplotlib.pyplot as plt
import tempfile
data=pd.read_csv("weatherdatafinal.csv")
def add_daytime_column(data):
data['sunrise'] = pd.to_datetime(data['sunrise'])
data['sunset'] = pd.to_datetime(data['sunset'])
data['daytime'] = (data['sunset'] - data['sunrise']).dt.total_seconds() / 3600.0
return data
data=add_daytime_column(data)
data = data.drop(columns=['name','datetime', 'severerisk', 'conditions', 'description', 'icon', 'stations','snow','snowdepth','sunrise','sunset','precip'])
data['preciptype'] = data['preciptype'].fillna(0)
data['preciptype'] = data['preciptype'].replace({'rain': 1, 'rain,snow': 2, 'snow': 3, 'rain,freezingrain,snow':3})
data['windgust'] = data['windgust'].fillna(data['windgust'].median())
data['sealevelpressure'] = data['sealevelpressure'].fillna(data['sealevelpressure'].median())
data['pressure'] = data['sealevelpressure']
data=data.drop("sealevelpressure", axis=1)
feature_names = list(data.columns)
def train_model(ideal_max_temp, ideal_min_temp, ideal_humidity):
ideal_weights = {
'tempmax': 6,
'tempmin': 6,
'temp': 6,
'humidity': 2,
'windspeed': 3,
'windgust': 1.5,
'cloudcover': 3,
'daytime': 1,
'precipprob': 1.5,
'visibility': 1,
'stability': 1
}
def normalize(value, min_value, max_value):
return (value - min_value) / (max_value - min_value)
for idx, row in data.iterrows():
tempmax_score = 1-normalize(abs(ideal_max_temp - row['tempmax']), min(data['tempmax']), max(data['tempmax']))
tempmin_score = 1-normalize(abs(ideal_min_temp - row['tempmin']), min(data['tempmin']), max(data['tempmin']))
temp_score = 1-normalize(abs(((ideal_max_temp + ideal_min_temp) / 2) - row['temp']), min(data['temp']), max(data['temp']))
stability_score = 1-normalize(abs(row['tempmax'] - row['tempmin']), min(data['tempmin']), max(data['tempmax']))
humidity_score = 1-normalize(abs(ideal_humidity - row['humidity']), min(data['humidity']), max(data['humidity']))
windspeed_score = 1-normalize(row['windspeed'], min(data['windspeed']), max(data['windspeed']))
windgust_score = 1-normalize(row['windgust'], min(data['windgust']), max(data['windgust']))
cloudcover_score = 1-normalize(row['cloudcover'], min(data['cloudcover']), max(data['cloudcover']))
daytime_score = normalize(row['daytime'], min(data['daytime']), max(data['daytime']))
precipprob_score = 1-normalize(row['precipprob'], min(data['precipprob']), max(data['precipprob']))
visibility_score = normalize(row['visibility'], min(data['visibility']), max(data['visibility']))
scores = [
tempmax_score * ideal_weights['tempmax'],
tempmin_score * ideal_weights['tempmin'],
temp_score * ideal_weights['temp'],
humidity_score * ideal_weights['humidity'],
windspeed_score * ideal_weights['windspeed'],
windgust_score * ideal_weights['windgust'],
cloudcover_score * ideal_weights['cloudcover'],
daytime_score * ideal_weights['daytime'],
precipprob_score * ideal_weights['precipprob'],
visibility_score * ideal_weights['visibility'],
stability_score * ideal_weights['stability']
]
daily_score = np.mean(scores)
data.loc[idx, 'daily_score'] = daily_score
scaler = MinMaxScaler(feature_range=(0, 95))
scaled_scores = scaler.fit_transform(data[['daily_score']])
data['daily_score'] = scaled_scores
X_train, X_test, y_train, y_test = train_test_split(data.drop('daily_score', axis=1), data['daily_score'], test_size=0.3, random_state=42)
scaler = MinMaxScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)
model = tf.keras.Sequential([
tf.keras.layers.Dense(128, activation='relu', input_shape=(X_train.shape[1],)),
tf.keras.layers.BatchNormalization(),
tf.keras.layers.Dense(64, activation='relu'),
tf.keras.layers.BatchNormalization(),
tf.keras.layers.Dropout(0.1),
tf.keras.layers.Dense(32, activation='relu'),
tf.keras.layers.BatchNormalization(),
tf.keras.layers.Dropout(0.1),
tf.keras.layers.Dense(16, activation='relu'),
tf.keras.layers.BatchNormalization(),
tf.keras.layers.Dropout(0.1),
tf.keras.layers.Dense(1)
])
model.compile(loss='mse', optimizer=tf.keras.optimizers.Adam(learning_rate=0.001), metrics=['mse', 'mae', 'msle'])
early_stop = tf.keras.callbacks.EarlyStopping(monitor='val_mse', patience=20, restore_best_weights=True)
history = model.fit(X_train, y_train, epochs=200, batch_size=32, validation_split=0.2, callbacks=[early_stop])
model.save('trainedmodel.h5')
with open('scaler.pkl', 'wb') as f:
pickle.dump(scaler, f)
return "Model trained based on your preferences."
def predict_weather(Location, Day):
model = tf.keras.models.load_model('trainedmodel.h5')
with open('scaler.pkl', 'rb') as f:
scaler = pickle.load(f)
prediction_day = Day.strip().lower()
if prediction_day == "yesterday":
day = (datetime.now() - timedelta(days=1)).strftime("%Y-%m-%d")
elif prediction_day == "today":
day = datetime.now().strftime("%Y-%m-%d")
elif prediction_day == "tomorrow":
day = (datetime.now() + timedelta(days=1)).strftime("%Y-%m-%d")
else:
print("Invalid prediction day. Defaulting to today.")
day = datetime.now().strftime("%Y-%m-%d")
url = f"https://weather.visualcrossing.com/VisualCrossingWebServices/rest/services/timeline/{Location}/{day}/{day}?unitGroup=metric&include=hours&key=TDAK3FZB5KTLU64J25LPTQ38Q&contentType=json"
response = requests.get(url)
urldata= response.json()
def add_daytime_column(urldata):
urldata['days'][0]['sunrise'] = pd.to_datetime(urldata['days'][0]['sunrise'])
urldata['days'][0]['sunset'] = pd.to_datetime(urldata['days'][0]['sunset'])
urldata['days'][0]['daytime'] = (urldata['days'][0]['sunset'] - urldata['days'][0]['sunrise']).total_seconds() / 3600.0
return urldata
urldata=add_daytime_column(urldata)
def preprocess_preciptype(urldata):
preciptype_dict = {'rain': 1, 'rain,snow': 2, 'snow': 3, 'rain,freezingrain,snow': 4,'None':0}
for day in urldata['days']:
if day.get('preciptype') is not None:
preciptype_str = day['preciptype'][0]
preciptype_code = preciptype_dict.get(preciptype_str, 0)
day['preciptype'] = preciptype_code
else:
day['preciptype'] = 0
return urldata
def replace_nan_with_median(urldata, data):
for col in ['solarradiation', 'solarenergy', 'uvindex']:
urldata['days'][0][col] = urldata['days'][0][col] or np.nan
for col in ['solarradiation', 'solarenergy', 'uvindex']:
if np.isnan(urldata['days'][0][col]):
urldata['days'][0][col] = data[col].median()
return urldata
urldata=replace_nan_with_median(urldata,data)
urldata=preprocess_preciptype(urldata)
def mean(data, key):
values = [hour[key] for hour in data]
return sum(values) / len(values)
hours_data = urldata["days"][0]["hours"][6:24]
day_data = urldata['days'][0]
new_data = {
'tempmax': [day_data['tempmax']],
'tempmin': [day_data['tempmin']],
'temp': [mean(hours_data, "temp")],
'feelslikemax': [day_data['feelslikemax']],
'feelslikemin': [day_data['feelslikemin']],
'feelslike': [mean(hours_data, "feelslike")],
'dew': [mean(hours_data, "dew")],
'humidity': [mean(hours_data, "humidity")],
'precipprob': [mean(hours_data, "precipprob")],
'precipcover': [day_data['precipcover']],
'preciptype': [day_data['preciptype']],
'windgust': [mean(hours_data, "windgust")],
'windspeed': [mean(hours_data, "windspeed")],
'winddir': [day_data['winddir']],
'pressure': [mean(hours_data, "pressure")],
'cloudcover': [mean(hours_data, "cloudcover")],
'visibility': [mean(hours_data, "visibility")],
'solarradiation': [day_data['solarradiation']],
'solarenergy': [day_data['solarenergy']],
'uvindex': [day_data['uvindex']],
'moonphase': [day_data['moonphase']],
'daytime': [day_data['daytime']]}
input_data = pd.DataFrame(new_data)
input_data = input_data[feature_names]
input_data = scaler.transform(input_data)
input_data = input_data.reshape(1, -1)
predictions = model.predict(input_data)
hourly_scores = []
new_data_hour = {
'tempmin': day_data['tempmin'],
'feelslikemin': day_data['feelslikemin'],
'precipcover': day_data['precipcover'],
'moonphase': day_data['moonphase'],
'daytime': day_data['daytime']
}
for hour_data in hours_data:
new_data_hour.update({
'tempmax': hour_data['temp'],
'feelslikemax': hour_data['feelslike'],
'temp': hour_data['temp'],
'feelslike': hour_data['feelslike'],
'dew': hour_data['dew'],
'humidity': hour_data['humidity'],
'precipprob': hour_data['precipprob'],
'preciptype': day_data['preciptype'],
'windgust': hour_data['windgust'],
'windspeed': hour_data['windspeed'],
'winddir': hour_data['winddir'],
'pressure': hour_data['pressure'],
'cloudcover': hour_data['cloudcover'],
'visibility': hour_data['visibility'],
'solarradiation': hour_data['solarradiation'],
'solarenergy': hour_data['solarenergy'],
'uvindex': hour_data['uvindex']
})
input_data_hour = pd.DataFrame([new_data_hour])
input_data_hour = input_data_hour[feature_names]
input_data_hour = scaler.transform(input_data_hour)
input_data_hour = input_data_hour.reshape(1, -1)
predictions_hour = model.predict(input_data_hour)
hourly_scores.append(predictions_hour[0][0])
score = predictions[0][0]
if score >= 80:
message = "The weather is expected to be great based on your preferences!"
elif score >= 60:
message = "The weather is expected to be good based on your preferences."
else:
message = "The weather might not be ideal based on your preferences."
return score, message, hourly_scores
def main():
mode = gr.inputs.Radio(["Train Model", "Predict Weather"], label="Mode", default="Predict Weather")
ideal_max_temp = gr.inputs.Slider(minimum=0, maximum=40, step=1, default=25, label="Ideal max temperature (°C)")
ideal_min_temp = gr.inputs.Slider(minimum=0, maximum=40, step=1, default=18, label="Ideal min temperature (°C)")
ideal_humidity = gr.inputs.Slider(minimum=40, maximum=100, step=1, default=75, label="Ideal humidity level (%)")
Location = gr.inputs.Textbox(placeholder="Enter your location (city name)", default="bangalore")
Day = gr.inputs.Radio(choices=["yesterday", "today", "tomorrow"], label="Select day:", default="today")
outputs = [
gr.outputs.Textbox(label="Training Result"),
gr.outputs.Textbox(label="Predicted Daily Score"),
gr.outputs.Textbox(label="Message"),
gr.outputs.Image(type="filepath", label="Hourly Rating Plot")
]
def wrapper(mode, ideal_max_temp, ideal_min_temp, ideal_humidity, Location, Day):
if mode == "Train Model":
result = train_model(ideal_max_temp, ideal_min_temp, ideal_humidity)
return result, None, None, None, None
else:
score, message, hourly_scores = predict_weather(Location, Day)
hours = range(6, 24)
plt.plot(hours,hourly_scores)
plt.xlabel('Hour of the Day')
plt.ylabel('Hourly Rating')
plt.xticks(range(6, 25, 1))
plt.xlim(6, 24)
plt.yticks(range(0, 101, 10))
plt.ylim(0, 100)
plt.title('Hourly Ratings Based On Your Preferences')
with tempfile.NamedTemporaryFile(suffix=".png", delete=False) as temp_file:
plt.savefig(temp_file.name, format='png')
img_filepath = temp_file.name
plt.clf()
return None, score, message, img_filepath
interface = gr.Interface(
fn=wrapper,
inputs=[mode, ideal_max_temp, ideal_min_temp, ideal_humidity, Location, Day],
outputs=outputs,
title="Weather Score",
description=(
"<b>WeatherPrediction: Personalized Weather Predictions</b><br>"
"Designed to provide you with tailored weather forecasts, taking into account your preferences for maximum and minimum temperature, humidity, and other key factors. Our advanced algorithms calculate weather features using historical and real-time data, delivering a personalized weather score to help you plan your day with confidence.<br><br>"
"<b>How to use:</b><br>"
"1. Input your preferred maximum temperature, minimum temperature, and humidity.<br>"
"2. Train the model to adapt to your preferences.<br>"
"3. Receive personalized weather scores to better plan your day.<br><br>"
"Whether you're planning outdoor activities or just want to know how the day will feel, WeatherPrediction gives you a user-focused forecast for a more enjoyable experience."
),
allow_flagging=False,
allow_screenshot=False
)
interface.launch()
if __name__ == "__main__":
main() |