Skip to content

Commit 80cb13c

Browse files
Copilotlurenss
Copilot
and
lurenss
committed
Add interactive quickstart script for easy demo
Co-authored-by: lurenss <[email protected]>
1 parent e86fc1a commit 80cb13c

File tree

1 file changed

+244
-0
lines changed

1 file changed

+244
-0
lines changed

‎quickstart.py‎

Lines changed: 244 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,244 @@
1+
#!/usr/bin/env python
2+
"""
3+
Quick Start Script for ScrapeGraphAI Elasticsearch Demo
4+
5+
This script provides an interactive demo of the key features.
6+
Run this after setting up Elasticsearch to see the integration in action.
7+
"""
8+
9+
import sys
10+
import os
11+
12+
# Add parent directory to path
13+
sys.path.insert(0, os.path.abspath(os.path.dirname(__file__)))
14+
15+
from src.scrapegraph_demo import Config, ElasticsearchClient, MarketplaceScraper
16+
from src.scrapegraph_demo.models import ProductComparison
17+
18+
19+
def print_header(text):
20+
"""Print a formatted header"""
21+
print("\n" + "=" * 70)
22+
print(f" {text}")
23+
print("=" * 70 + "\n")
24+
25+
26+
def print_step(number, text):
27+
"""Print a formatted step"""
28+
print(f"\n[{number}] {text}")
29+
print("-" * 70)
30+
31+
32+
def wait_for_user():
33+
"""Wait for user to press Enter"""
34+
input("\nPress Enter to continue...")
35+
36+
37+
def main():
38+
"""Run the interactive quickstart demo"""
39+
40+
print_header("ScrapeGraphAI Elasticsearch Demo - Quick Start")
41+
42+
print("This interactive demo will showcase:")
43+
print(" • Configuration loading")
44+
print(" • Elasticsearch connection")
45+
print(" • Product scraping (using mock data)")
46+
print(" • Data indexing")
47+
print(" • Product search")
48+
print(" • Product comparison")
49+
print()
50+
51+
input("Press Enter to begin...")
52+
53+
# Step 1: Load Configuration
54+
print_step(1, "Loading Configuration")
55+
config = Config.from_env()
56+
print(f"✓ Configuration loaded")
57+
print(f" Elasticsearch URL: {config.elasticsearch_url}")
58+
wait_for_user()
59+
60+
# Step 2: Connect to Elasticsearch
61+
print_step(2, "Connecting to Elasticsearch")
62+
try:
63+
es_client = ElasticsearchClient(config)
64+
print(f"✓ Connected to Elasticsearch")
65+
print(f" Index name: {es_client.INDEX_NAME}")
66+
es_connected = True
67+
except Exception as e:
68+
print(f"✗ Could not connect to Elasticsearch: {e}")
69+
print("\nNote: Elasticsearch is not running. Continuing with mock data only.")
70+
print("To use Elasticsearch, run: docker-compose up -d")
71+
es_connected = False
72+
wait_for_user()
73+
74+
# Step 3: Initialize Scraper
75+
print_step(3, "Initializing Marketplace Scraper")
76+
scraper = MarketplaceScraper(config)
77+
print("✓ Scraper initialized")
78+
print(" Using mock data for demonstration")
79+
wait_for_user()
80+
81+
# Step 4: Scrape Products
82+
print_step(4, "Scraping Products from Multiple Marketplaces")
83+
84+
search_query = "wireless headphones"
85+
marketplaces = ["Amazon", "eBay", "BestBuy"]
86+
all_products = []
87+
88+
print(f"Search query: '{search_query}'")
89+
print()
90+
91+
for marketplace in marketplaces:
92+
print(f" Scraping {marketplace}...", end=" ")
93+
products = scraper.scrape_search_results(search_query, marketplace, max_results=2)
94+
all_products.extend(products)
95+
print(f"✓ Found {len(products)} products")
96+
97+
print(f"\n✓ Total products scraped: {len(all_products)}")
98+
wait_for_user()
99+
100+
# Step 5: Display Sample Products
101+
print_step(5, "Sample Product Data")
102+
103+
for i, product in enumerate(all_products[:3], 1):
104+
print(f"\nProduct {i}:")
105+
print(f" Name: {product.name}")
106+
print(f" Price: ${product.price:.2f} {product.currency}")
107+
print(f" Marketplace: {product.marketplace}")
108+
print(f" Brand: {product.brand}")
109+
print(f" Rating: {product.rating}/5.0")
110+
print(f" Reviews: {product.review_count}")
111+
112+
wait_for_user()
113+
114+
# Step 6: Index Products (if Elasticsearch is available)
115+
if es_connected:
116+
print_step(6, "Indexing Products in Elasticsearch")
117+
118+
try:
119+
success, failed = es_client.index_products(all_products)
120+
print(f"✓ Successfully indexed {success} products")
121+
if failed:
122+
print(f"✗ Failed to index {len(failed)} products")
123+
except Exception as e:
124+
print(f"✗ Error indexing products: {e}")
125+
es_connected = False
126+
127+
wait_for_user()
128+
129+
# Step 7: Product Comparison
130+
print_step(7, "Product Comparison Analysis")
131+
132+
comparison = ProductComparison(
133+
query=search_query,
134+
products=all_products
135+
)
136+
137+
print(f"Query: {comparison.query}")
138+
print(f"Total products: {len(comparison.products)}")
139+
print()
140+
141+
# Price analysis
142+
min_price, max_price = comparison.get_price_range()
143+
print(f"Price Range: ${min_price:.2f} - ${max_price:.2f}")
144+
print()
145+
146+
# Cheapest product
147+
cheapest = comparison.get_cheapest()
148+
print("Cheapest Product:")
149+
print(f" {cheapest.name}")
150+
print(f" ${cheapest.price:.2f} on {cheapest.marketplace}")
151+
print()
152+
153+
# Best rated
154+
best_rated = comparison.get_best_rated()
155+
if best_rated:
156+
print("Best Rated Product:")
157+
print(f" {best_rated.name}")
158+
print(f" {best_rated.rating}/5.0 ({best_rated.review_count} reviews)")
159+
print(f" ${best_rated.price:.2f} on {best_rated.marketplace}")
160+
161+
wait_for_user()
162+
163+
# Step 8: Products by Marketplace
164+
print_step(8, "Products Grouped by Marketplace")
165+
166+
grouped = comparison.group_by_marketplace()
167+
for marketplace, products in grouped.items():
168+
print(f"\n{marketplace} ({len(products)} products):")
169+
for product in products:
170+
print(f" • {product.name}")
171+
print(f" ${product.price:.2f} | {product.rating}/5.0")
172+
173+
wait_for_user()
174+
175+
# Step 9: Search (if Elasticsearch is available)
176+
if es_connected:
177+
print_step(9, "Searching Products in Elasticsearch")
178+
179+
try:
180+
# Search with price filter
181+
print(f"Searching for '{search_query}' under $50...")
182+
results = es_client.search_products(search_query, max_price=50.0, size=5)
183+
184+
print(f"\n✓ Found {len(results)} results:")
185+
for i, product in enumerate(results, 1):
186+
print(f"\n{i}. {product.name}")
187+
print(f" ${product.price:.2f} | {product.marketplace}")
188+
print(f" Rating: {product.rating}/5.0")
189+
190+
# Statistics
191+
print("\n" + "-" * 70)
192+
stats = es_client.aggregate_by_marketplace()
193+
print("\nProducts by marketplace:")
194+
for marketplace, count in stats.items():
195+
print(f" {marketplace}: {count}")
196+
197+
price_stats = es_client.get_price_statistics()
198+
print(f"\nPrice Statistics:")
199+
print(f" Average: ${price_stats['avg']:.2f}")
200+
print(f" Min: ${price_stats['min']:.2f}")
201+
print(f" Max: ${price_stats['max']:.2f}")
202+
203+
except Exception as e:
204+
print(f"Error searching: {e}")
205+
206+
wait_for_user()
207+
208+
# Final message
209+
print_header("Quick Start Complete!")
210+
211+
print("✓ You've successfully explored the ScrapeGraphAI Elasticsearch Demo!")
212+
print()
213+
print("Next Steps:")
214+
print(" 1. Review the README.md for detailed documentation")
215+
print(" 2. Check out examples/ directory for more use cases")
216+
print(" 3. Run the test suite: python run_tests.py")
217+
print(" 4. Try the example scripts:")
218+
print(" - python examples/basic_usage.py")
219+
print(" - python examples/product_comparison.py")
220+
print(" - python examples/advanced_search.py")
221+
print()
222+
223+
if es_connected:
224+
print(" 5. Access Kibana at http://localhost:5601 for data visualization")
225+
print()
226+
es_client.close()
227+
else:
228+
print(" 5. Start Elasticsearch: docker-compose up -d")
229+
print(" Then re-run this quickstart script")
230+
print()
231+
232+
print("Thank you for trying ScrapeGraphAI Elasticsearch Demo! 🎉")
233+
print()
234+
235+
236+
if __name__ == "__main__":
237+
try:
238+
main()
239+
except KeyboardInterrupt:
240+
print("\n\nQuickstart interrupted by user.")
241+
sys.exit(0)
242+
except Exception as e:
243+
print(f"\n\nError: {e}")
244+
sys.exit(1)

0 commit comments

Comments
 (0)