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

tests/test_component_factory.py

@@ -0,0 +1,175 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+组件工厂测试
+"""
+import sys
+import os
+import unittest
+
+# 添加项目根目录到 Python 路径
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
+
+from crawlo.factories import (
+    ComponentRegistry, 
+    ComponentFactory, 
+    ComponentSpec, 
+    DefaultComponentFactory,
+    CrawlerComponentFactory,
+    get_component_registry
+)
+from crawlo.factories.base import ComponentSpec
+
+
+class TestComponent:
+    """测试组件类"""
+    def __init__(self, name="test_component", value=42):
+        self.name = name
+        self.value = value
+
+
+class TestComponentFactory(unittest.TestCase):
+    """组件工厂测试类"""
+
+    def setUp(self):
+        """测试前准备"""
+        self.registry = ComponentRegistry()
+        
+    def test_component_spec_creation(self):
+        """测试组件规范创建"""
+        def factory_func(**kwargs):
+            return TestComponent(**kwargs)
+            
+        spec = ComponentSpec(
+            name="test_component",
+            component_type=TestComponent,
+            factory_func=factory_func,
+            dependencies=[],
+            singleton=False
+        )
+        
+        self.assertEqual(spec.name, "test_component")
+        self.assertEqual(spec.component_type, TestComponent)
+        self.assertEqual(spec.dependencies, [])
+        self.assertFalse(spec.singleton)
+        
+    def test_default_component_factory(self):
+        """测试默认组件工厂"""
+        factory = DefaultComponentFactory()
+        
+        def factory_func(**kwargs):
+            return TestComponent(**kwargs)
+            
+        spec = ComponentSpec(
+            name="test_component",
+            component_type=TestComponent,
+            factory_func=factory_func
+        )
+        
+        # 测试创建组件
+        component = factory.create(spec, name="created_component", value=100)
+        self.assertIsInstance(component, TestComponent)
+        self.assertEqual(component.name, "created_component")
+        self.assertEqual(component.value, 100)
+        
+        # 测试单例模式
+        spec_singleton = ComponentSpec(
+            name="singleton_component",
+            component_type=TestComponent,
+            factory_func=factory_func,
+            singleton=True
+        )
+        
+        component1 = factory.create(spec_singleton, name="singleton_1", value=200)
+        component2 = factory.create(spec_singleton, name="singleton_2", value=300)
+        
+        # 单例应该返回相同的实例
+        self.assertIs(component1, component2)
+        self.assertEqual(component1.value, 200)  # 应该是第一次创建时的值
+        
+    def test_component_registry_registration(self):
+        """测试组件注册表注册功能"""
+        def factory_func(**kwargs):
+            return TestComponent(**kwargs)
+            
+        spec = ComponentSpec(
+            name="registered_component",
+            component_type=TestComponent,
+            factory_func=factory_func
+        )
+        
+        # 注册组件规范
+        self.registry.register(spec)
+        
+        # 验证注册
+        retrieved_spec = self.registry.get_spec("registered_component")
+        self.assertEqual(retrieved_spec, spec)
+        
+        # 测试列出组件
+        components = self.registry.list_components()
+        self.assertIn("registered_component", components)
+        
+    def test_component_registry_creation(self):
+        """测试组件注册表创建功能"""
+        def factory_func(**kwargs):
+            return TestComponent(**kwargs)
+            
+        spec = ComponentSpec(
+            name="creatable_component",
+            component_type=TestComponent,
+            factory_func=factory_func
+        )
+        
+        # 注册组件规范
+        self.registry.register(spec)
+        
+        # 创建组件
+        component = self.registry.create("creatable_component", name="created", value=500)
+        self.assertIsInstance(component, TestComponent)
+        self.assertEqual(component.name, "created")
+        self.assertEqual(component.value, 500)
+        
+    def test_global_component_registry(self):
+        """测试全局组件注册表"""
+        registry = get_component_registry()
+        self.assertIsInstance(registry, ComponentRegistry)
+        
+        # 测试注册表是否包含预注册的组件
+        components = registry.list_components()
+        # 应该至少包含crawler组件
+        self.assertGreater(len(components), 0)
+        
+    def test_crawler_component_factory(self):
+        """测试Crawler组件工厂"""
+        factory = CrawlerComponentFactory()
+        
+        # 测试支持检查
+        class MockEngine:
+            pass
+            
+        self.assertTrue(factory.supports(MockEngine))  # 默认支持所有类型
+        
+        # 测试创建功能（需要crawler依赖）
+        def mock_engine_factory(crawler=None, **kwargs):
+            if crawler is None:
+                raise ValueError("需要crawler实例")
+            return "mock_engine"
+            
+        spec = ComponentSpec(
+            name="mock_engine",
+            component_type=type('MockEngine', (), {}),
+            factory_func=mock_engine_factory,
+            dependencies=['crawler']
+        )
+        
+        # 测试缺少依赖时的错误处理
+        with self.assertRaises(ValueError):
+            factory.create(spec)
+            
+        # 测试带依赖的创建
+        result = factory.create(spec, crawler="mock_crawler")
+        self.assertEqual(result, "mock_engine")
+
+
+if __name__ == '__main__':
+    unittest.main()

tests/test_comprehensive.py

@@ -1,147 +1,147 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-"""
-综合测试
-验证所有改进的集成效果
-"""
-import sys
-import os
-import asyncio
-import unittest
-from unittest.mock import patch, MagicMock
-
-# 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.env_config import get_env_var, get_redis_config, get_runtime_config
-from crawlo.utils.error_handler import ErrorHandler, handle_exception
-from crawlo.core.engine import Engine
-from crawlo.settings.setting_manager import SettingManager
-from crawlo.settings import default_settings
-from crawlo.queue.queue_manager import QueueManager, QueueConfig, QueueType
-
-
-class TestComprehensiveIntegration(unittest.TestCase):
-    """综合集成测试"""
-
-    def setUp(self):
-        """测试前准备"""
-        # 设置测试环境变量
-        self.test_env = {
-            'PROJECT_NAME': 'test_project',
-            'CONCURRENCY': '4',
-            'REDIS_HOST': 'localhost',
-            'REDIS_PORT': '6379'
-        }
-        self.original_env = {}
-        for key, value in self.test_env.items():
-            self.original_env[key] = os.environ.get(key)
-            os.environ[key] = value
-
-    def tearDown(self):
-        """测试后清理"""
-        # 恢复原始环境变量
-        for key, value in self.original_env.items():
-            if value is None:
-                os.environ.pop(key, None)
-            else:
-                os.environ[key] = value
-
-    def test_env_config_integration(self):
-        """测试环境变量配置集成"""
-        # 验证环境变量工具正常工作
-        project_name = get_env_var('PROJECT_NAME', 'default', str)
-        self.assertEqual(project_name, 'test_project')
-        
-        concurrency = get_env_var('CONCURRENCY', 1, int)
-        self.assertEqual(concurrency, 4)
-        
-        # 验证Redis配置工具
-        redis_config = get_redis_config()
-        self.assertEqual(redis_config['REDIS_HOST'], 'localhost')
-        self.assertEqual(redis_config['REDIS_PORT'], 6379)
-
-    def test_error_handler_integration(self):
-        """测试错误处理集成"""
-        # 验证错误处理模块正常工作
-        error_handler = ErrorHandler("test")
-        
-        # 测试错误处理
-        try:
-            error_handler.handle_error(ValueError("Test error"), raise_error=False)
-        except Exception:
-            self.fail("Error handler should not raise exception when raise_error=False")
-        
-        # 测试安全调用
-        result = error_handler.safe_call(lambda x: x*2, 5, default_return=0)
-        self.assertEqual(result, 10)
-        
-        # 测试装饰器
-        @handle_exception(raise_error=False)
-        def failing_function():
-            raise RuntimeError("Test")
-        
-        try:
-            failing_function()
-        except Exception:
-            self.fail("Decorated function should not raise exception")
-
-    def test_settings_integration(self):
-        """测试设置管理器集成"""
-        # 重新加载默认设置以获取环境变量
-        import importlib
-        import crawlo.settings.default_settings
-        importlib.reload(crawlo.settings.default_settings)
-        
-        # 创建设置管理器
-        settings = SettingManager()
-        settings.set_settings(crawlo.settings.default_settings)
-        
-        # 验证设置正确加载
-        self.assertEqual(settings.get('PROJECT_NAME'), 'test_project')
-        self.assertEqual(settings.get_int('CONCURRENCY'), 4)
-        self.assertEqual(settings.get('REDIS_HOST'), 'localhost')
-
-    def test_queue_manager_config(self):
-        """测试队列管理器配置"""
-        # 重新加载默认设置
-        import importlib
-        import crawlo.settings.default_settings
-        importlib.reload(crawlo.settings.default_settings)
-        
-        # 创建设置管理器
-        settings = SettingManager()
-        settings.set_settings(crawlo.settings.default_settings)
-        
-        # 从设置创建队列配置
-        queue_config = QueueConfig.from_settings(settings)
-        
-        # 验证配置正确
-        self.assertEqual(queue_config.queue_type, QueueType.AUTO)
-        self.assertIn('test_project', queue_config.queue_name)
-
-    async def test_async_components(self):
-        """测试异步组件"""
-        # 测试异步错误处理装饰器
-        @handle_exception(raise_error=False)
-        async def async_failing_function():
-            raise RuntimeError("Async test")
-        
-        try:
-            await async_failing_function()
-        except Exception:
-            self.fail("Async decorated function should not raise exception")
-
-
-if __name__ == '__main__':
-    # 运行同步测试
-    unittest.main(exit=False)
-    
-    # 运行异步测试
-    async def run_async_tests():
-        test_instance = TestComprehensiveIntegration()
-        test_instance.setUp()
-        await test_instance.test_async_components()
-        test_instance.tearDown()
-    
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+综合测试
+验证所有改进的集成效果
+"""
+import sys
+import os
+import asyncio
+import unittest
+from unittest.mock import patch, MagicMock
+
+# 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.env_config import get_env_var, get_redis_config, get_runtime_config
+from crawlo.utils.error_handler import ErrorHandler, handle_exception
+from crawlo.core.engine import Engine
+from crawlo.settings.setting_manager import SettingManager
+from crawlo.settings import default_settings
+from crawlo.queue.queue_manager import QueueManager, QueueConfig, QueueType
+
+
+class TestComprehensiveIntegration(unittest.TestCase):
+    """综合集成测试"""
+
+    def setUp(self):
+        """测试前准备"""
+        # 设置测试环境变量
+        self.test_env = {
+            'PROJECT_NAME': 'test_project',
+            'CONCURRENCY': '4',
+            'REDIS_HOST': 'localhost',
+            'REDIS_PORT': '6379'
+        }
+        self.original_env = {}
+        for key, value in self.test_env.items():
+            self.original_env[key] = os.environ.get(key)
+            os.environ[key] = value
+
+    def tearDown(self):
+        """测试后清理"""
+        # 恢复原始环境变量
+        for key, value in self.original_env.items():
+            if value is None:
+                os.environ.pop(key, None)
+            else:
+                os.environ[key] = value
+
+    def test_env_config_integration(self):
+        """测试环境变量配置集成"""
+        # 验证环境变量工具正常工作
+        project_name = get_env_var('PROJECT_NAME', 'default', str)
+        self.assertEqual(project_name, 'test_project')
+        
+        concurrency = get_env_var('CONCURRENCY', 1, int)
+        self.assertEqual(concurrency, 4)
+        
+        # 验证Redis配置工具
+        redis_config = get_redis_config()
+        self.assertEqual(redis_config['REDIS_HOST'], 'localhost')
+        self.assertEqual(redis_config['REDIS_PORT'], 6379)
+
+    def test_error_handler_integration(self):
+        """测试错误处理集成"""
+        # 验证错误处理模块正常工作
+        error_handler = ErrorHandler("test")
+        
+        # 测试错误处理
+        try:
+            error_handler.handle_error(ValueError("Test error"), raise_error=False)
+        except Exception:
+            self.fail("Error handler should not raise exception when raise_error=False")
+        
+        # 测试安全调用
+        result = error_handler.safe_call(lambda x: x*2, 5, default_return=0)
+        self.assertEqual(result, 10)
+        
+        # 测试装饰器
+        @handle_exception(raise_error=False)
+        def failing_function():
+            raise RuntimeError("Test")
+        
+        try:
+            failing_function()
+        except Exception:
+            self.fail("Decorated function should not raise exception")
+
+    def test_settings_integration(self):
+        """测试设置管理器集成"""
+        # 重新加载默认设置以获取环境变量
+        import importlib
+        import crawlo.settings.default_settings
+        importlib.reload(crawlo.settings.default_settings)
+        
+        # 创建设置管理器
+        settings = SettingManager()
+        settings.set_settings(crawlo.settings.default_settings)
+        
+        # 验证设置正确加载
+        self.assertEqual(settings.get('PROJECT_NAME'), 'test_project')
+        self.assertEqual(settings.get_int('CONCURRENCY'), 4)
+        self.assertEqual(settings.get('REDIS_HOST'), 'localhost')
+
+    def test_queue_manager_config(self):
+        """测试队列管理器配置"""
+        # 重新加载默认设置
+        import importlib
+        import crawlo.settings.default_settings
+        importlib.reload(crawlo.settings.default_settings)
+        
+        # 创建设置管理器
+        settings = SettingManager()
+        settings.set_settings(crawlo.settings.default_settings)
+        
+        # 从设置创建队列配置
+        queue_config = QueueConfig.from_settings(settings)
+        
+        # 验证配置正确
+        self.assertEqual(queue_config.queue_type, QueueType.AUTO)
+        self.assertIn('test_project', queue_config.queue_name)
+
+    async def test_async_components(self):
+        """测试异步组件"""
+        # 测试异步错误处理装饰器
+        @handle_exception(raise_error=False)
+        async def async_failing_function():
+            raise RuntimeError("Async test")
+        
+        try:
+            await async_failing_function()
+        except Exception:
+            self.fail("Async decorated function should not raise exception")
+
+
+if __name__ == '__main__':
+    # 运行同步测试
+    unittest.main(exit=False)
+    
+    # 运行异步测试
+    async def run_async_tests():
+        test_instance = TestComprehensiveIntegration()
+        test_instance.setUp()
+        await test_instance.test_async_components()
+        test_instance.tearDown()
+    
     asyncio.run(run_async_tests())

tests/test_config_consistency.py

@@ -1,81 +1,81 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-"""
-测试配置一致性优化
-"""
-import asyncio
-import sys
-import os
-
-# 添加项目根目录到路径
-sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
-
-from crawlo.project import get_settings
-from crawlo.crawler import Crawler
-from crawlo.spider import Spider
-from crawlo.utils.log import get_logger
-from crawlo import Request
-
-
-class TestSpider(Spider):
-    name = "test_spider"
-    
-    def start_requests(self):
-        yield Request("https://example.com")
-
-
-async def test_config_consistency():
-    """测试配置一致性优化"""
-    print("测试配置一致性优化...")
-    
-    # 模拟单机模式配置但Redis可用的情况
-    custom_settings = {
-        'QUEUE_TYPE': 'auto',  # 自动检测模式
-        'CONCURRENCY': 4,
-        'DOWNLOAD_DELAY': 1.0,
-        'LOG_LEVEL': 'INFO'
-    }
-    
-    try:
-        # 获取配置
-        settings = get_settings(custom_settings)
-        
-        # 创建爬虫实例
-        crawler = Crawler(TestSpider, settings)
-        
-        # 启动爬虫（这会触发调度器初始化）
-        print("开始初始化爬虫...")
-        await crawler.crawl()
-        
-        print("配置一致性测试完成")
-        
-    except Exception as e:
-        print(f"测试失败: {e}")
-        import traceback
-        traceback.print_exc()
-
-
-async def main():
-    """主测试函数"""
-    print("开始测试配置一致性优化...")
-    print("=" * 50)
-    
-    try:
-        await test_config_consistency()
-        
-        print("=" * 50)
-        print("配置一致性优化测试完成！")
-        
-    except Exception as e:
-        print("=" * 50)
-        print(f"测试失败: {e}")
-        import traceback
-        traceback.print_exc()
-
-
-if __name__ == "__main__":
-    # 设置日志级别
-    import logging
-    logging.basicConfig(level=logging.INFO)
-    
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+测试配置一致性优化
+"""
+import asyncio
+import sys
+import os
+
+# 添加项目根目录到路径
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
+
+from crawlo.project import get_settings
+from crawlo.crawler import Crawler
+from crawlo.spider import Spider
+from crawlo.utils.log import get_logger
+from crawlo import Request
+
+
+class TestSpider(Spider):
+    name = "test_spider"
+    
+    def start_requests(self):
+        yield Request("https://example.com")
+
+
+async def test_config_consistency():
+    """测试配置一致性优化"""
+    print("测试配置一致性优化...")
+    
+    # 模拟单机模式配置但Redis可用的情况
+    custom_settings = {
+        'QUEUE_TYPE': 'auto',  # 自动检测模式
+        'CONCURRENCY': 4,
+        'DOWNLOAD_DELAY': 1.0,
+        'LOG_LEVEL': 'INFO'
+    }
+    
+    try:
+        # 获取配置
+        settings = get_settings(custom_settings)
+        
+        # 创建爬虫实例
+        crawler = Crawler(TestSpider, settings)
+        
+        # 启动爬虫（这会触发调度器初始化）
+        print("开始初始化爬虫...")
+        await crawler.crawl()
+        
+        print("配置一致性测试完成")
+        
+    except Exception as e:
+        print(f"测试失败: {e}")
+        import traceback
+        traceback.print_exc()
+
+
+async def main():
+    """主测试函数"""
+    print("开始测试配置一致性优化...")
+    print("=" * 50)
+    
+    try:
+        await test_config_consistency()
+        
+        print("=" * 50)
+        print("配置一致性优化测试完成！")
+        
+    except Exception as e:
+        print("=" * 50)
+        print(f"测试失败: {e}")
+        import traceback
+        traceback.print_exc()
+
+
+if __name__ == "__main__":
+    # 设置日志级别
+    import logging
+    logging.basicConfig(level=logging.INFO)
+    
     asyncio.run(main())