ai-agent-book 精选快照（<2MB 代码与文档，来自 github.com/bojieli/ai-agent-book）

chapter9/browser-use-rpa/browser-use/examples/cloud/05_search_api.py

@@ -0,0 +1,361 @@
+"""
+Cloud Example 5: Search API (Beta) 🔍
+=====================================
+
+This example demonstrates the Browser Use Search API (BETA):
+- Simple search: Search Google and extract from multiple results
+- URL search: Extract specific content from a target URL
+- Deep navigation through websites (depth parameter)
+- Real-time content extraction vs cached results
+
+Perfect for: Content extraction, research, competitive analysis
+"""
+
+import argparse
+import asyncio
+import json
+import os
+import time
+from typing import Any
+
+import aiohttp
+
+try:
+	from dotenv import load_dotenv
+	load_dotenv()
+except ImportError:
+	pass
+
+# Configuration
+API_KEY = os.getenv('BROWSER_USE_API_KEY')
+if not API_KEY:
+	raise ValueError(
+		'Please set BROWSER_USE_API_KEY environment variable. You can also create an API key at https://cloud.browser-use.com'
+	)
+
+BASE_URL = os.getenv('BROWSER_USE_BASE_URL', 'https://api.browser-use.com/api/v1')
+TIMEOUT = int(os.getenv('BROWSER_USE_TIMEOUT', '30'))
+HEADERS = {'Authorization': f'Bearer {API_KEY}', 'Content-Type': 'application/json'}
+
+
+async def simple_search(query: str, max_websites: int = 5, depth: int = 2) -> dict[str, Any]:
+	"""
+	Search Google and extract content from multiple top results.
+
+	Args:
+	    query: Search query to process
+	    max_websites: Number of websites to process (1-10)
+	    depth: How deep to navigate (2-5)
+
+	Returns:
+	    Dictionary with results from multiple websites
+	"""
+	# Validate input parameters
+	max_websites = max(1, min(max_websites, 10))  # Clamp to 1-10
+	depth = max(2, min(depth, 5))  # Clamp to 2-5
+
+	start_time = time.time()
+
+	print(f"🔍 Simple Search: '{query}'")
+	print(f'📊 Processing {max_websites} websites at depth {depth}')
+	print(f'💰 Estimated cost: {depth * max_websites}¢')
+
+	payload = {'query': query, 'max_websites': max_websites, 'depth': depth}
+
+	timeout = aiohttp.ClientTimeout(total=TIMEOUT)
+	connector = aiohttp.TCPConnector(limit=10)  # Limit concurrent connections
+
+	async with aiohttp.ClientSession(timeout=timeout, connector=connector) as session:
+		async with session.post(f'{BASE_URL}/simple-search', json=payload, headers=HEADERS) as response:
+			elapsed = time.time() - start_time
+			if response.status == 200:
+				try:
+					result = await response.json()
+					print(f'✅ Found results from {len(result.get("results", []))} websites in {elapsed:.1f}s')
+					return result
+				except (aiohttp.ContentTypeError, json.JSONDecodeError) as e:
+					error_text = await response.text()
+					print(f'❌ Invalid JSON response: {e} (after {elapsed:.1f}s)')
+					return {'error': 'Invalid JSON', 'details': error_text}
+			else:
+				error_text = await response.text()
+				print(f'❌ Search failed: {response.status} - {error_text} (after {elapsed:.1f}s)')
+				return {'error': f'HTTP {response.status}', 'details': error_text}
+
+
+async def search_url(url: str, query: str, depth: int = 2) -> dict[str, Any]:
+	"""
+	Extract specific content from a target URL.
+
+	Args:
+	    url: Target URL to extract from
+	    query: What specific content to look for
+	    depth: How deep to navigate (2-5)
+
+	Returns:
+	    Dictionary with extracted content
+	"""
+	# Validate input parameters
+	depth = max(2, min(depth, 5))  # Clamp to 2-5
+
+	start_time = time.time()
+
+	print(f'🎯 URL Search: {url}')
+	print(f"🔍 Looking for: '{query}'")
+	print(f'📊 Navigation depth: {depth}')
+	print(f'💰 Estimated cost: {depth}¢')
+
+	payload = {'url': url, 'query': query, 'depth': depth}
+
+	timeout = aiohttp.ClientTimeout(total=TIMEOUT)
+	connector = aiohttp.TCPConnector(limit=10)  # Limit concurrent connections
+
+	async with aiohttp.ClientSession(timeout=timeout, connector=connector) as session:
+		async with session.post(f'{BASE_URL}/search-url', json=payload, headers=HEADERS) as response:
+			elapsed = time.time() - start_time
+			if response.status == 200:
+				try:
+					result = await response.json()
+					print(f'✅ Extracted content from {result.get("url", "website")} in {elapsed:.1f}s')
+					return result
+				except (aiohttp.ContentTypeError, json.JSONDecodeError) as e:
+					error_text = await response.text()
+					print(f'❌ Invalid JSON response: {e} (after {elapsed:.1f}s)')
+					return {'error': 'Invalid JSON', 'details': error_text}
+			else:
+				error_text = await response.text()
+				print(f'❌ URL search failed: {response.status} - {error_text} (after {elapsed:.1f}s)')
+				return {'error': f'HTTP {response.status}', 'details': error_text}
+
+
+def display_simple_search_results(results: dict[str, Any]):
+	"""Display simple search results in a readable format."""
+	if 'error' in results:
+		print(f'❌ Error: {results["error"]}')
+		return
+
+	websites = results.get('results', [])
+
+	print(f'\n📋 Search Results ({len(websites)} websites)')
+	print('=' * 50)
+
+	for i, site in enumerate(websites, 1):
+		url = site.get('url', 'Unknown URL')
+		content = site.get('content', 'No content')
+
+		print(f'\n{i}. 🌐 {url}')
+		print('-' * 40)
+
+		# Show first 300 chars of content
+		if len(content) > 300:
+			print(f'{content[:300]}...')
+			print(f'[Content truncated - {len(content)} total characters]')
+		else:
+			print(content)
+
+	# Show execution URLs if available
+	if results.get('live_url'):
+		print(f'\n🔗 Live Preview: {results["live_url"]}')
+	if results.get('public_share_url'):
+		print(f'🌐 Share URL: {results["public_share_url"]}')
+	elif results.get('share_url'):
+		print(f'🌐 Share URL: {results["share_url"]}')
+
+
+def display_url_search_results(results: dict[str, Any]):
+	"""Display URL search results in a readable format."""
+	if 'error' in results:
+		print(f'❌ Error: {results["error"]}')
+		return
+
+	url = results.get('url', 'Unknown URL')
+	content = results.get('content', 'No content')
+
+	print(f'\n📄 Extracted Content from: {url}')
+	print('=' * 60)
+	print(content)
+
+	# Show execution URLs if available
+	if results.get('live_url'):
+		print(f'\n🔗 Live Preview: {results["live_url"]}')
+	if results.get('public_share_url'):
+		print(f'🌐 Share URL: {results["public_share_url"]}')
+	elif results.get('share_url'):
+		print(f'🌐 Share URL: {results["share_url"]}')
+
+
+async def demo_news_search():
+	"""Demo: Search for latest news across multiple sources."""
+	print('\n📰 Demo 1: Latest News Search')
+	print('-' * 35)
+
+	demo_start = time.time()
+	query = 'latest developments in artificial intelligence 2024'
+	results = await simple_search(query, max_websites=4, depth=2)
+	demo_elapsed = time.time() - demo_start
+
+	display_simple_search_results(results)
+	print(f'\n⏱️  Total demo time: {demo_elapsed:.1f}s')
+
+	return results
+
+
+async def demo_competitive_analysis():
+	"""Demo: Analyze competitor websites."""
+	print('\n🏢 Demo 2: Competitive Analysis')
+	print('-' * 35)
+
+	query = 'browser automation tools comparison features pricing'
+	results = await simple_search(query, max_websites=3, depth=3)
+	display_simple_search_results(results)
+
+	return results
+
+
+async def demo_deep_website_analysis():
+	"""Demo: Deep analysis of a specific website."""
+	print('\n🎯 Demo 3: Deep Website Analysis')
+	print('-' * 35)
+
+	demo_start = time.time()
+	url = 'https://docs.browser-use.com'
+	query = 'Browser Use features, pricing, and API capabilities'
+	results = await search_url(url, query, depth=3)
+	demo_elapsed = time.time() - demo_start
+
+	display_url_search_results(results)
+	print(f'\n⏱️  Total demo time: {demo_elapsed:.1f}s')
+
+	return results
+
+
+async def demo_product_research():
+	"""Demo: Product research and comparison."""
+	print('\n🛍️  Demo 4: Product Research')
+	print('-' * 30)
+
+	query = 'best wireless headphones 2024 reviews comparison'
+	results = await simple_search(query, max_websites=5, depth=2)
+	display_simple_search_results(results)
+
+	return results
+
+
+async def demo_real_time_vs_cached():
+	"""Demo: Show difference between real-time and cached results."""
+	print('\n⚡ Demo 5: Real-time vs Cached Data')
+	print('-' * 40)
+
+	print('🔄 Browser Use Search API benefits:')
+	print('• Actually browses websites like a human')
+	print('• Gets live, current data (not cached)')
+	print('• Navigates deep into sites via clicks')
+	print('• Handles JavaScript and dynamic content')
+	print('• Accesses pages requiring navigation')
+
+	# Example with live data
+	query = 'current Bitcoin price USD live'
+	results = await simple_search(query, max_websites=3, depth=2)
+
+	print('\n💰 Live Bitcoin Price Search Results:')
+	display_simple_search_results(results)
+
+	return results
+
+
+async def demo_search_depth_comparison():
+	"""Demo: Compare different search depths."""
+	print('\n📊 Demo 6: Search Depth Comparison')
+	print('-' * 40)
+
+	url = 'https://news.ycombinator.com'
+	query = 'trending technology discussions'
+
+	depths = [2, 3, 4]
+	results = {}
+
+	for depth in depths:
+		print(f'\n🔍 Testing depth {depth}:')
+		result = await search_url(url, query, depth)
+		results[depth] = result
+
+		if 'content' in result:
+			content_length = len(result['content'])
+			print(f'📏 Content length: {content_length} characters')
+
+		# Brief pause between requests
+		await asyncio.sleep(1)
+
+	# Summary
+	print('\n📊 Depth Comparison Summary:')
+	print('-' * 30)
+	for depth, result in results.items():
+		if 'content' in result:
+			length = len(result['content'])
+			print(f'Depth {depth}: {length} characters')
+		else:
+			print(f'Depth {depth}: Error or no content')
+
+	return results
+
+
+async def main():
+	"""Demonstrate comprehensive Search API usage."""
+	print('🔍 Browser Use Cloud - Search API (BETA)')
+	print('=' * 45)
+
+	print('⚠️  Note: This API is in BETA and may change')
+	print()
+	print('🎯 Search API Features:')
+	print('• Real-time website browsing (not cached)')
+	print('• Deep navigation through multiple pages')
+	print('• Dynamic content and JavaScript handling')
+	print('• Multiple result aggregation')
+	print('• Cost-effective content extraction')
+
+	print('\n💰 Pricing:')
+	print('• Simple Search: 1¢ × depth × websites')
+	print('• URL Search: 1¢ × depth')
+	print('• Example: depth=2, 5 websites = 10¢')
+
+	try:
+		# Parse command line arguments
+		parser = argparse.ArgumentParser(description='Search API (BETA) examples')
+		parser.add_argument(
+			'--demo',
+			choices=['news', 'competitive', 'deep', 'product', 'realtime', 'depth', 'all'],
+			default='news',
+			help='Which demo to run',
+		)
+		args = parser.parse_args()
+
+		print(f'\n🔍 Running {args.demo} demo(s)...')
+
+		if args.demo == 'news':
+			await demo_news_search()
+		elif args.demo == 'competitive':
+			await demo_competitive_analysis()
+		elif args.demo == 'deep':
+			await demo_deep_website_analysis()
+		elif args.demo == 'product':
+			await demo_product_research()
+		elif args.demo == 'realtime':
+			await demo_real_time_vs_cached()
+		elif args.demo == 'depth':
+			await demo_search_depth_comparison()
+		elif args.demo == 'all':
+			await demo_news_search()
+			await demo_competitive_analysis()
+			await demo_deep_website_analysis()
+			await demo_product_research()
+			await demo_real_time_vs_cached()
+			await demo_search_depth_comparison()
+
+	except aiohttp.ClientError as e:
+		print(f'❌ Network Error: {e}')
+	except Exception as e:
+		print(f'❌ Error: {e}')
+
+
+if __name__ == '__main__':
+	asyncio.run(main())