app.py

import gradio as gr
import pandas as pd
import numpy as np
import requests

WEBHOOK_URL = "https://eliabigdata.app.n8n.cloud/webhook/restaurant-analyzer"

restaurants = ['PastaPlace', 'BurgerHub', 'SushiBar', 'TacoLoco', 'CafeBleu']

np.random.seed(42)
df_sales = pd.DataFrame({
    'restaurant': np.repeat(restaurants, 20),
    'week': list(range(1, 21)) * 5,
    'revenue': np.random.randint(3000, 15000, 100),
    'avg_price': np.random.uniform(12, 45, 100).round(2)
})

def analyze_restaurant(restaurant, review):
    try:
        # Send to n8n webhook
        response = requests.post(WEBHOOK_URL, json={
            "restaurant": restaurant,
            "review": review
        }, timeout=10)
        
        if response.status_code == 200:
            data = response.json()
            sentiment = data.get('sentiment', 'Neutral')
            recommendation = data.get('recommendation', 'Keep current strategy')
        else:
            raise Exception("Webhook error")
            
    except:
        # Fallback local analysis
        positive_words = ['amazing', 'great', 'fantastic', 'loved', 'best', 'delicious']
        negative_words = ['terrible', 'bad', 'disappointing', 'never', 'worst', 'expensive']
        review_lower = review.lower()
        pos = sum(w in review_lower for w in positive_words)
        neg = sum(w in review_lower for w in negative_words)
        if pos > neg:
            sentiment = "Positive"
            recommendation = "Increase prices by 10%"
        elif neg > pos:
            sentiment = "Negative"
            recommendation = "Improve service first"
        else:
            sentiment = "Neutral"
            recommendation = "Keep current strategy"

    stats = df_sales[df_sales['restaurant'] == restaurant]['revenue']
    avg_rev = stats.mean().round(2)

    result = f"""
🍽️ Restaurant: {restaurant}
📊 Sentiment: {sentiment}
💰 Average Weekly Revenue: €{avg_rev}
💡 Recommendation: {recommendation}
    """
    return result

iface = gr.Interface(
    fn=analyze_restaurant,
    inputs=[
        gr.Dropdown(choices=restaurants, label="Select Restaurant"),
        gr.Textbox(label="Enter a customer review",
                   placeholder="e.g. Amazing food, loved it!")
    ],
    outputs=gr.Textbox(label="Analysis & Recommendation"),
    title="🍽️ Restaurant Performance Analyzer",
    description="Analyze customer sentiment via n8n webhook and get pricing recommendations"
)

iface.launch()