Update app.py
Browse files
app.py
CHANGED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import pandas as pd
|
| 3 |
+
import numpy as np
|
| 4 |
+
from sklearn.ensemble import IsolationForest
|
| 5 |
+
|
| 6 |
+
def generate_sample_data():
|
| 7 |
+
# Simulate 200 time points of energy consumption
|
| 8 |
+
rng = np.random.default_rng(seed=42)
|
| 9 |
+
normal_consumption = rng.normal(loc=100, scale=10, size=200)
|
| 10 |
+
|
| 11 |
+
# Inject anomalies (potential theft) - unusually low or high consumption spikes
|
| 12 |
+
anomalies = rng.choice([50, 200], size=10) # Very low or very high
|
| 13 |
+
anomaly_indices = rng.choice(range(200), size=10, replace=False)
|
| 14 |
+
normal_consumption[anomaly_indices] = anomalies
|
| 15 |
+
|
| 16 |
+
df = pd.DataFrame({
|
| 17 |
+
'timestamp': pd.date_range(start='2025-01-01', periods=200, freq='H'),
|
| 18 |
+
'consumption': normal_consumption
|
| 19 |
+
})
|
| 20 |
+
return df
|
| 21 |
+
|
| 22 |
+
def detect_energy_theft(file=None, use_sample=False):
|
| 23 |
+
if use_sample:
|
| 24 |
+
df = generate_sample_data()
|
| 25 |
+
else:
|
| 26 |
+
try:
|
| 27 |
+
df = pd.read_csv(file.name)
|
| 28 |
+
except Exception as e:
|
| 29 |
+
return f"Error reading file: {e}"
|
| 30 |
+
|
| 31 |
+
if 'consumption' not in df.columns:
|
| 32 |
+
return "CSV must have a 'consumption' column."
|
| 33 |
+
|
| 34 |
+
X = df[['consumption']].values
|
| 35 |
+
|
| 36 |
+
model = IsolationForest(contamination=0.05, random_state=42)
|
| 37 |
+
model.fit(X)
|
| 38 |
+
|
| 39 |
+
df['anomaly_score'] = model.decision_function(X)
|
| 40 |
+
df['anomaly'] = model.predict(X)
|
| 41 |
+
df['theft_detected'] = df['anomaly'] == -1
|
| 42 |
+
|
| 43 |
+
suspicious = df[df['theft_detected'] == True][['timestamp', 'consumption', 'anomaly_score']]
|
| 44 |
+
if suspicious.empty:
|
| 45 |
+
return "No energy theft anomalies detected."
|
| 46 |
+
else:
|
| 47 |
+
return suspicious
|
| 48 |
+
|
| 49 |
+
# Gradio UI
|
| 50 |
+
with gr.Blocks() as demo:
|
| 51 |
+
gr.Markdown("# Energy Theft Detection")
|
| 52 |
+
gr.Markdown("Upload CSV with 'consumption' column or use sample data to detect anomalies.")
|
| 53 |
+
|
| 54 |
+
with gr.Row():
|
| 55 |
+
file_input = gr.File(label="Upload your CSV file", file_types=['.csv'])
|
| 56 |
+
sample_btn = gr.Button("Use Sample Data")
|
| 57 |
+
|
| 58 |
+
output = gr.Dataframe(headers=["timestamp", "consumption", "anomaly_score"], label="Suspicious Anomalies Detected")
|
| 59 |
+
|
| 60 |
+
def on_file_submit(file):
|
| 61 |
+
return detect_energy_theft(file=file, use_sample=False)
|
| 62 |
+
|
| 63 |
+
def on_sample_click():
|
| 64 |
+
return detect_energy_theft(use_sample=True)
|
| 65 |
+
|
| 66 |
+
file_input.change(on_file_submit, inputs=file_input, outputs=output)
|
| 67 |
+
sample_btn.click(on_sample_click, outputs=output)
|
| 68 |
+
|
| 69 |
+
if __name__ == "__main__":
|
| 70 |
+
demo.launch()
|