crawlo 1.3.3__py3-none-any.whl → 1.3.4__py3-none-any.whl

tests/test_config_validator.py

@@ -1,183 +1,183 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-"""
-配置验证器测试
-测试 Crawlo 框架的配置验证功能
-"""
-
-import sys
-import os
-import unittest
-
-# 添加项目根目录到路径
-sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
-
-from crawlo.config_validator import ConfigValidator
-
-
-class TestConfigValidator(unittest.TestCase):
-    """配置验证器测试类"""
-
-    def setUp(self):
-        """测试初始化"""
-        self.validator = ConfigValidator()
-
-    def test_valid_standalone_config(self):
-        """测试有效的单机模式配置"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'memory',
-            'CONCURRENCY': 8,
-            'DOWNLOAD_DELAY': 1.0,
-            'LOG_LEVEL': 'INFO',
-            'MIDDLEWARES': [
-                'crawlo.middleware.request_ignore.RequestIgnoreMiddleware',
-                'crawlo.middleware.download_delay.DownloadDelayMiddleware',
-            ],
-            'PIPELINES': [
-                'crawlo.pipelines.console_pipeline.ConsolePipeline',
-            ]
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertTrue(is_valid)
-        self.assertEqual(len(errors), 0)
-
-    def test_valid_distributed_config(self):
-        """测试有效的分布式模式配置"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'redis',
-            'CONCURRENCY': 16,
-            'DOWNLOAD_DELAY': 0.5,
-            'LOG_LEVEL': 'INFO',
-            'REDIS_HOST': '127.0.0.1',
-            'REDIS_PORT': 6379,
-            'REDIS_PASSWORD': '',
-            'SCHEDULER_QUEUE_NAME': 'crawlo:test_project:queue:requests',
-            'MIDDLEWARES': [
-                'crawlo.middleware.request_ignore.RequestIgnoreMiddleware',
-                'crawlo.middleware.download_delay.DownloadDelayMiddleware',
-            ],
-            'PIPELINES': [
-                'crawlo.pipelines.console_pipeline.ConsolePipeline',
-                'crawlo.pipelines.redis_dedup_pipeline.RedisDedupPipeline',
-            ],
-            'EXTENSIONS': [
-                'crawlo.extension.memory_monitor.MemoryMonitorExtension',
-            ]
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertTrue(is_valid)
-        self.assertEqual(len(errors), 0)
-
-    def test_invalid_project_name(self):
-        """测试无效的项目名称"""
-        config = {
-            'PROJECT_NAME': '',  # 空项目名称
-            'QUEUE_TYPE': 'memory',
-            'CONCURRENCY': 8,
-            'LOG_LEVEL': 'INFO'
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertFalse(is_valid)
-        self.assertIn("PROJECT_NAME 必须是非空字符串", errors)
-
-    def test_invalid_queue_type(self):
-        """测试无效的队列类型"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'invalid_type',  # 无效队列类型
-            'CONCURRENCY': 8,
-            'LOG_LEVEL': 'INFO'
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertFalse(is_valid)
-        self.assertIn("QUEUE_TYPE 必须是以下值之一: ['memory', 'redis', 'auto']", errors)
-
-    def test_invalid_concurrency(self):
-        """测试无效的并发数"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'memory',
-            'CONCURRENCY': -1,  # 负数并发数
-            'LOG_LEVEL': 'INFO'
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertFalse(is_valid)
-        self.assertIn("CONCURRENCY 必须是正整数", errors)
-
-    def test_invalid_log_level(self):
-        """测试无效的日志级别"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'memory',
-            'CONCURRENCY': 8,
-            'LOG_LEVEL': 'INVALID_LEVEL'  # 无效日志级别
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertFalse(is_valid)
-        self.assertIn("LOG_LEVEL 必须是以下值之一: ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']", errors)
-
-    def test_invalid_middleware_config(self):
-        """测试无效的中间件配置"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'memory',
-            'CONCURRENCY': 8,
-            'LOG_LEVEL': 'INFO',
-            'MIDDLEWARES': 'not_a_list'  # 不是列表
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertFalse(is_valid)
-        self.assertIn("MIDDLEWARES 必须是列表", errors)
-
-    def test_invalid_pipeline_config(self):
-        """测试无效的管道配置"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'memory',
-            'CONCURRENCY': 8,
-            'LOG_LEVEL': 'INFO',
-            'PIPELINES': 'not_a_list'  # 不是列表
-        }
-        
-        is_valid, errors, warnings = self.validator.validate(config)
-        self.assertFalse(is_valid)
-        self.assertIn("PIPELINES 必须是列表", errors)
-
-    def test_convenience_function(self):
-        """测试便利函数"""
-        config = {
-            'PROJECT_NAME': 'test_project',
-            'QUEUE_TYPE': 'memory',
-            'CONCURRENCY': 8,
-            'LOG_LEVEL': 'INFO'
-        }
-        
-        from crawlo.config_validator import validate_config
-        is_valid, errors, warnings = validate_config(config)
-        self.assertTrue(is_valid)
-        self.assertEqual(len(errors), 0)
-
-
-def main():
-    """主测试函数"""
-    print("开始配置验证器测试...")
-    print("=" * 50)
-    
-    # 运行测试
-    unittest.main(argv=['first-arg-is-ignored'], exit=False, verbosity=2)
-    
-    print("=" * 50)
-    print("配置验证器测试完成")
-
-
-if __name__ == "__main__":
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+配置验证器测试
+测试 Crawlo 框架的配置验证功能
+"""
+
+import sys
+import os
+import unittest
+
+# 添加项目根目录到路径
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
+
+from crawlo.config_validator import ConfigValidator
+
+
+class TestConfigValidator(unittest.TestCase):
+    """配置验证器测试类"""
+
+    def setUp(self):
+        """测试初始化"""
+        self.validator = ConfigValidator()
+
+    def test_valid_standalone_config(self):
+        """测试有效的单机模式配置"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'memory',
+            'CONCURRENCY': 8,
+            'DOWNLOAD_DELAY': 1.0,
+            'LOG_LEVEL': 'INFO',
+            'MIDDLEWARES': [
+                'crawlo.middleware.request_ignore.RequestIgnoreMiddleware',
+                'crawlo.middleware.download_delay.DownloadDelayMiddleware',
+            ],
+            'PIPELINES': [
+                'crawlo.pipelines.console_pipeline.ConsolePipeline',
+            ]
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertTrue(is_valid)
+        self.assertEqual(len(errors), 0)
+
+    def test_valid_distributed_config(self):
+        """测试有效的分布式模式配置"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'redis',
+            'CONCURRENCY': 16,
+            'DOWNLOAD_DELAY': 0.5,
+            'LOG_LEVEL': 'INFO',
+            'REDIS_HOST': '127.0.0.1',
+            'REDIS_PORT': 6379,
+            'REDIS_PASSWORD': '',
+            'SCHEDULER_QUEUE_NAME': 'crawlo:test_project:queue:requests',
+            'MIDDLEWARES': [
+                'crawlo.middleware.request_ignore.RequestIgnoreMiddleware',
+                'crawlo.middleware.download_delay.DownloadDelayMiddleware',
+            ],
+            'PIPELINES': [
+                'crawlo.pipelines.console_pipeline.ConsolePipeline',
+                'crawlo.pipelines.redis_dedup_pipeline.RedisDedupPipeline',
+            ],
+            'EXTENSIONS': [
+                'crawlo.extension.memory_monitor.MemoryMonitorExtension',
+            ]
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertTrue(is_valid)
+        self.assertEqual(len(errors), 0)
+
+    def test_invalid_project_name(self):
+        """测试无效的项目名称"""
+        config = {
+            'PROJECT_NAME': '',  # 空项目名称
+            'QUEUE_TYPE': 'memory',
+            'CONCURRENCY': 8,
+            'LOG_LEVEL': 'INFO'
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertFalse(is_valid)
+        self.assertIn("PROJECT_NAME 必须是非空字符串", errors)
+
+    def test_invalid_queue_type(self):
+        """测试无效的队列类型"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'invalid_type',  # 无效队列类型
+            'CONCURRENCY': 8,
+            'LOG_LEVEL': 'INFO'
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertFalse(is_valid)
+        self.assertIn("QUEUE_TYPE 必须是以下值之一: ['memory', 'redis', 'auto']", errors)
+
+    def test_invalid_concurrency(self):
+        """测试无效的并发数"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'memory',
+            'CONCURRENCY': -1,  # 负数并发数
+            'LOG_LEVEL': 'INFO'
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertFalse(is_valid)
+        self.assertIn("CONCURRENCY 必须是正整数", errors)
+
+    def test_invalid_log_level(self):
+        """测试无效的日志级别"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'memory',
+            'CONCURRENCY': 8,
+            'LOG_LEVEL': 'INVALID_LEVEL'  # 无效日志级别
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertFalse(is_valid)
+        self.assertIn("LOG_LEVEL 必须是以下值之一: ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']", errors)
+
+    def test_invalid_middleware_config(self):
+        """测试无效的中间件配置"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'memory',
+            'CONCURRENCY': 8,
+            'LOG_LEVEL': 'INFO',
+            'MIDDLEWARES': 'not_a_list'  # 不是列表
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertFalse(is_valid)
+        self.assertIn("MIDDLEWARES 必须是列表", errors)
+
+    def test_invalid_pipeline_config(self):
+        """测试无效的管道配置"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'memory',
+            'CONCURRENCY': 8,
+            'LOG_LEVEL': 'INFO',
+            'PIPELINES': 'not_a_list'  # 不是列表
+        }
+        
+        is_valid, errors, warnings = self.validator.validate(config)
+        self.assertFalse(is_valid)
+        self.assertIn("PIPELINES 必须是列表", errors)
+
+    def test_convenience_function(self):
+        """测试便利函数"""
+        config = {
+            'PROJECT_NAME': 'test_project',
+            'QUEUE_TYPE': 'memory',
+            'CONCURRENCY': 8,
+            'LOG_LEVEL': 'INFO'
+        }
+        
+        from crawlo.config_validator import validate_config
+        is_valid, errors, warnings = validate_config(config)
+        self.assertTrue(is_valid)
+        self.assertEqual(len(errors), 0)
+
+
+def main():
+    """主测试函数"""
+    print("开始配置验证器测试...")
+    print("=" * 50)
+    
+    # 运行测试
+    unittest.main(argv=['first-arg-is-ignored'], exit=False, verbosity=2)
+    
+    print("=" * 50)
+    print("配置验证器测试完成")
+
+
+if __name__ == "__main__":
     main()

tests/test_controlled_spider_mixin.py

@@ -0,0 +1,80 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+受控爬虫混入类测试
+测试 ControlledRequestMixin, AsyncControlledRequestMixin
+"""
+import sys
+import os
+import unittest
+from unittest.mock import Mock, patch, MagicMock
+import asyncio
+
+# 添加项目根目录到 Python 路径
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
+
+from crawlo.utils.controlled_spider_mixin import ControlledRequestMixin, AsyncControlledRequestMixin
+
+
+class TestControlledRequestMixin(unittest.TestCase):
+    """受控请求混入类测试"""
+
+    def setUp(self):
+        """测试前准备"""
+        self.mixin = ControlledRequestMixin()
+        
+    def test_mixin_initialization(self):
+        """测试混入类初始化"""
+        self.assertEqual(self.mixin.max_pending_requests, 100)
+        self.assertEqual(self.mixin.batch_size, 50)
+        self.assertEqual(self.mixin.generation_interval, 0.1)
+        self.assertEqual(self.mixin.backpressure_threshold, 200)
+        
+    def test_mixin_configuration(self):
+        """测试混入类配置"""
+        # 修改配置
+        self.mixin.max_pending_requests = 200
+        self.mixin.batch_size = 100
+        self.mixin.generation_interval = 0.05
+        self.mixin.backpressure_threshold = 300
+        
+        self.assertEqual(self.mixin.max_pending_requests, 200)
+        self.assertEqual(self.mixin.batch_size, 100)
+        self.assertEqual(self.mixin.generation_interval, 0.05)
+        self.assertEqual(self.mixin.backpressure_threshold, 300)
+        
+    def test_get_generation_stats(self):
+        """测试获取生成统计信息"""
+        stats = self.mixin.get_generation_stats()
+        self.assertIsInstance(stats, dict)
+        self.assertIn('generated', stats)
+        self.assertIn('skipped', stats)
+        self.assertIn('backpressure_events', stats)
+        self.assertIn('total_generated', stats)
+        self.assertIn('last_generation_time', stats)
+
+
+class TestAsyncControlledRequestMixin(unittest.TestCase):
+    """异步受控请求混入类测试"""
+
+    def setUp(self):
+        """测试前准备"""
+        self.mixin = AsyncControlledRequestMixin()
+        
+    def test_async_mixin_initialization(self):
+        """测试异步混入类初始化"""
+        self.assertEqual(self.mixin.max_concurrent_generations, 10)
+        self.assertEqual(self.mixin.queue_monitor_interval, 1.0)
+        
+    def test_async_mixin_configuration(self):
+        """测试异步混入类配置"""
+        # 修改配置
+        self.mixin.max_concurrent_generations = 20
+        self.mixin.queue_monitor_interval = 0.5
+        
+        self.assertEqual(self.mixin.max_concurrent_generations, 20)
+        self.assertEqual(self.mixin.queue_monitor_interval, 0.5)
+
+
+if __name__ == '__main__':
+    unittest.main()

tests/test_crawlo_proxy_integration.py

@@ -1,109 +1,109 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-"""
-Crawlo框架代理集成测试
-测试代理中间件与框架的集成
-"""
-
-import asyncio
-import sys
-import os
-from unittest.mock import Mock, patch
-
-# 添加项目根目录到路径
-sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
-
-from crawlo.config import CrawloConfig
-from crawlo.crawler import CrawlerProcess
-from crawlo import Spider, Request, Item
-from crawlo.middleware.proxy import ProxyMiddleware
-
-
-class TestItem(Item):
-    """测试结果项"""
-    url = ''
-    status = 0
-    proxy = ''
-
-
-class ProxyTestSpider(Spider):
-    """代理测试爬虫"""
-    name = 'proxy_test_spider'
-    
-    async def start_requests(self):
-        """发起测试请求"""
-        yield Request('https://httpbin.org/ip', callback=self.parse)
-    
-    async def parse(self, response):
-        """解析响应"""
-        import json
-        try:
-            data = json.loads(response.text)
-            ip_info = data.get('origin', 'Unknown')
-        except:
-            ip_info = 'Parse error'
-        
-        item = TestItem(
-            url=response.url,
-            status=response.status,
-            proxy=str(response.meta.get('proxy', 'No proxy'))
-        )
-        
-        self.logger.info(f"Proxy test result: IP={ip_info}, Proxy={item.proxy}")
-        yield item
-
-
-async def test_proxy_integration():
-    """测试代理集成"""
-    print("🔍 测试代理集成...")
-    
-    # 创建配置
-    config = CrawloConfig.standalone(
-        concurrency=1,
-        download_delay=0.1,
-        PROXY_ENABLED=True,
-        PROXY_API_URL="https://api.proxyprovider.com/get",  # 模拟代理API
-        LOG_LEVEL='WARNING'  # 减少日志输出
-    )
-    
-    # 添加代理中间件
-    config.set('CUSTOM_MIDDLEWARES', [
-        'crawlo.middleware.proxy.ProxyMiddleware',
-    ])
-    
-    # 创建爬虫进程
-    process = CrawlerProcess(settings=config.to_dict())
-    
-    # 添加爬虫
-    process.crawl(ProxyTestSpider)
-    
-    # 运行测试
-    await process.start()
-    
-    print("   代理集成测试完成")
-
-
-async def main():
-    """主测试函数"""
-    print("开始Crawlo框架代理集成测试...")
-    print("=" * 50)
-    
-    try:
-        await test_proxy_integration()
-        
-        print("=" * 50)
-        print("所有代理集成测试通过！")
-        
-    except Exception as e:
-        print("=" * 50)
-        print(f"测试失败: {e}")
-        import traceback
-        traceback.print_exc()
-        return 1
-    
-    return 0
-
-
-if __name__ == "__main__":
-    exit_code = asyncio.run(main())
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+Crawlo框架代理集成测试
+测试代理中间件与框架的集成
+"""
+
+import asyncio
+import sys
+import os
+from unittest.mock import Mock, patch
+
+# 添加项目根目录到路径
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
+
+from crawlo.config import CrawloConfig
+from crawlo.crawler import CrawlerProcess
+from crawlo import Spider, Request, Item
+from crawlo.middleware.proxy import ProxyMiddleware
+
+
+class TestItem(Item):
+    """测试结果项"""
+    url = ''
+    status = 0
+    proxy = ''
+
+
+class ProxyTestSpider(Spider):
+    """代理测试爬虫"""
+    name = 'proxy_test_spider'
+    
+    async def start_requests(self):
+        """发起测试请求"""
+        yield Request('https://httpbin.org/ip', callback=self.parse)
+    
+    async def parse(self, response):
+        """解析响应"""
+        import json
+        try:
+            data = json.loads(response.text)
+            ip_info = data.get('origin', 'Unknown')
+        except:
+            ip_info = 'Parse error'
+        
+        item = TestItem(
+            url=response.url,
+            status=response.status,
+            proxy=str(response.meta.get('proxy', 'No proxy'))
+        )
+        
+        self.logger.info(f"Proxy test result: IP={ip_info}, Proxy={item.proxy}")
+        yield item
+
+
+async def test_proxy_integration():
+    """测试代理集成"""
+    print("🔍 测试代理集成...")
+    
+    # 创建配置
+    config = CrawloConfig.standalone(
+        concurrency=1,
+        download_delay=0.1,
+        PROXY_ENABLED=True,
+        PROXY_API_URL="https://api.proxyprovider.com/get",  # 模拟代理API
+        LOG_LEVEL='WARNING'  # 减少日志输出
+    )
+    
+    # 添加代理中间件
+    config.set('CUSTOM_MIDDLEWARES', [
+        'crawlo.middleware.proxy.ProxyMiddleware',
+    ])
+    
+    # 创建爬虫进程
+    process = CrawlerProcess(settings=config.to_dict())
+    
+    # 添加爬虫
+    process.crawl(ProxyTestSpider)
+    
+    # 运行测试
+    await process.start()
+    
+    print("   代理集成测试完成")
+
+
+async def main():
+    """主测试函数"""
+    print("开始Crawlo框架代理集成测试...")
+    print("=" * 50)
+    
+    try:
+        await test_proxy_integration()
+        
+        print("=" * 50)
+        print("所有代理集成测试通过！")
+        
+    except Exception as e:
+        print("=" * 50)
+        print(f"测试失败: {e}")
+        import traceback
+        traceback.print_exc()
+        return 1
+    
+    return 0
+
+
+if __name__ == "__main__":
+    exit_code = asyncio.run(main())
     sys.exit(exit_code)