app.py

import gradio as gr
import numpy as np
import pandas as pd
from sklearn.ensemble import IsolationForest
from sklearn.preprocessing import LabelEncoder
import matplotlib
matplotlib.use('Agg')  # Use 'Agg' backend for compatibility
import matplotlib.pyplot as plt
from PIL import Image
import os

# Initialize the Isolation Forest model
model = IsolationForest(contamination=0.1, random_state=42)

def preprocess_data(data):
    """
    Preprocess the input data:
    - Handle missing values
    - Encode categorical variables
    - Ensure all data is numeric
    """
    # Handle missing values by filling with the mean of the column
    data = data.fillna(data.mean(numeric_only=True))

    # Encode categorical variables
    label_encoders = {}
    for column in data.select_dtypes(include=['object']).columns:
        le = LabelEncoder()
        data[column] = le.fit_transform(data[column])
        label_encoders[column] = le

    # Ensure all data is numeric
    data = data.apply(pd.to_numeric, errors='coerce')

    return data, label_encoders

def detect_anomalies(data_input, file_input):
    try:
        # Load data from file if provided
        if file_input:
            file_extension = os.path.splitext(file_input.name)[-1].lower()
            if file_extension == ".csv":
                data = pd.read_csv(file_input.name)
            elif file_extension == ".json":
                data = pd.read_json(file_input.name)
            else:
                return {"Error": f"Unsupported file type: {file_extension}"}, None
        else:
            # Use manual input if no file provided
            data = pd.DataFrame([x.split(",") for x in data_input.split("\n")])

        # Preprocess the data
        data, label_encoders = preprocess_data(data)

        # Fit model and predict anomalies
        model.fit(data)
        predictions = model.predict(data)

        # Extract anomalies
        anomalies = data[predictions == -1]

        # Create plot
        plt.figure(figsize=(8, 6))
        plt.plot(data.index, data, 'o', label="Data Points")
        plt.plot(anomalies.index, anomalies, 'ro', label="Anomalies")
        plt.title("Anomaly Detection Results")
        plt.xlabel("Data Index")
        plt.ylabel("Data Value")
        plt.legend()
        plt.tight_layout()

        # Save plot to a temporary file
        temp_plot_path = "temp_plot.png"
        plt.savefig(temp_plot_path)
        plt.close()

        # Convert plot to PIL Image for Gradio
        plot_image = Image.open(temp_plot_path)

        # Clean up the temporary plot file
        os.remove(temp_plot_path)

        return {"Anomalies": anomalies.to_dict(orient='records'), "Total Anomalies": len(anomalies)}, plot_image

    except Exception as e:
        return {"Error": str(e)}, None

# Gradio Interface
iface = gr.Interface(
    fn=detect_anomalies,
    inputs=[
        gr.Textbox(label="Enter data (comma-separated values, one row per line)"),
        gr.File(label="Upload Data File")
    ],
    outputs=[
        "json",
        gr.Image(type="pil", label="Anomaly Plot")
    ],
    title="Anomaly Detector",
    description="Enter a series of numbers or upload a data file to detect anomalies using Isolation Forest."
)

# Launch the interface
iface.launch()