crawlo 1.2.8__py3-none-any.whl → 1.2.9__py3-none-any.whl

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())

tests/test_config_merge.py

@@ -0,0 +1,153 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+配置合并测试
+测试 Crawlo 框架的配置合并功能
+"""
+
+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.settings.setting_manager import SettingManager
+
+
+class TestConfigMerge(unittest.TestCase):
+    """配置合并测试类"""
+
+    def test_middleware_merge(self):
+        """测试中间件配置合并"""
+        # 用户自定义配置
+        user_config = {
+            'MIDDLEWARES': [
+                'myproject.middlewares.CustomMiddleware',
+            ]
+        }
+        
+        settings = SettingManager(user_config)
+        
+        # 获取合并后的中间件列表
+        middlewares = settings.get('MIDDLEWARES')
+        
+        # 检查默认中间件是否存在
+        self.assertIn('crawlo.middleware.request_ignore.RequestIgnoreMiddleware', middlewares)
+        self.assertIn('crawlo.middleware.download_delay.DownloadDelayMiddleware', middlewares)
+        
+        # 检查自定义中间件是否存在
+        self.assertIn('myproject.middlewares.CustomMiddleware', middlewares)
+        
+        # 检查合并后的顺序是否正确
+        default_index = middlewares.index('crawlo.middleware.request_ignore.RequestIgnoreMiddleware')
+        custom_index = middlewares.index('myproject.middlewares.CustomMiddleware')
+        self.assertLess(default_index, custom_index)
+
+    def test_pipeline_merge(self):
+        """测试管道配置合并"""
+        # 用户自定义配置
+        user_config = {
+            'PIPELINES': [
+                'myproject.pipelines.CustomPipeline',
+            ]
+        }
+        
+        settings = SettingManager(user_config)
+        
+        # 获取合并后的管道列表
+        pipelines = settings.get('PIPELINES')
+        
+        # 检查默认管道是否存在
+        self.assertIn('crawlo.pipelines.console_pipeline.ConsolePipeline', pipelines)
+        
+        # 检查自定义管道是否存在
+        self.assertIn('myproject.pipelines.CustomPipeline', pipelines)
+        
+        # 检查去重管道是否在开头
+        dedup_pipeline = settings.get('DEFAULT_DEDUP_PIPELINE')
+        self.assertEqual(pipelines[0], dedup_pipeline)
+        self.assertEqual(dedup_pipeline, 'crawlo.pipelines.memory_dedup_pipeline.MemoryDedupPipeline')
+
+    def test_extension_merge(self):
+        """测试扩展配置合并"""
+        # 用户自定义配置
+        user_config = {
+            'EXTENSIONS': [
+                'myproject.extensions.CustomExtension',
+            ]
+        }
+        
+        settings = SettingManager(user_config)
+        
+        # 获取合并后的扩展列表
+        extensions = settings.get('EXTENSIONS')
+        
+        # 检查默认扩展是否存在
+        self.assertIn('crawlo.extension.log_interval.LogIntervalExtension', extensions)
+        self.assertIn('crawlo.extension.log_stats.LogStats', extensions)
+        self.assertIn('crawlo.extension.logging_extension.CustomLoggerExtension', extensions)
+        
+        # 检查自定义扩展是否存在
+        self.assertIn('myproject.extensions.CustomExtension', extensions)
+        
+        # 检查合并后的顺序是否正确
+        default_index = extensions.index('crawlo.extension.log_interval.LogIntervalExtension')
+        custom_index = extensions.index('myproject.extensions.CustomExtension')
+        self.assertLess(default_index, custom_index)
+
+    def test_empty_custom_config(self):
+        """测试空自定义配置"""
+        # 空用户配置
+        user_config = {}
+        
+        settings = SettingManager(user_config)
+        
+        # 获取合并后的中间件列表
+        middlewares = settings.get('MIDDLEWARES')
+        
+        # 检查默认中间件是否存在
+        self.assertIn('crawlo.middleware.request_ignore.RequestIgnoreMiddleware', middlewares)
+        self.assertIn('crawlo.middleware.download_delay.DownloadDelayMiddleware', middlewares)
+        
+        # 检查管道和扩展
+        pipelines = settings.get('PIPELINES')
+        self.assertIn('crawlo.pipelines.console_pipeline.ConsolePipeline', pipelines)
+        
+        extensions = settings.get('EXTENSIONS')
+        self.assertIn('crawlo.extension.log_interval.LogIntervalExtension', extensions)
+
+    def test_no_custom_config(self):
+        """测试无自定义配置"""
+        # 无用户配置
+        settings = SettingManager()
+        
+        # 获取合并后的中间件列表
+        middlewares = settings.get('MIDDLEWARES')
+        
+        # 检查默认中间件是否存在
+        self.assertIn('crawlo.middleware.request_ignore.RequestIgnoreMiddleware', middlewares)
+        self.assertIn('crawlo.middleware.download_delay.DownloadDelayMiddleware', middlewares)
+        
+        # 检查管道和扩展
+        pipelines = settings.get('PIPELINES')
+        self.assertIn('crawlo.pipelines.console_pipeline.ConsolePipeline', pipelines)
+        
+        extensions = settings.get('EXTENSIONS')
+        self.assertIn('crawlo.extension.log_interval.LogIntervalExtension', extensions)
+
+
+def main():
+    """主测试函数"""
+    print("开始配置合并测试...")
+    print("=" * 50)
+    
+    # 运行测试
+    unittest.main(argv=['first-arg-is-ignored'], exit=False, verbosity=2)
+    
+    print("=" * 50)
+    print("配置合并测试完成")
+
+
+if __name__ == "__main__":
+    main()