crawlo 1.4.6__py3-none-any.whl → 1.4.7__py3-none-any.whl

tests/test_cli_arguments.py

@@ -0,0 +1,119 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+测试CLI参数解析
+验证crawlo run命令是否正确解析--log-level、--config和--concurrency参数
+"""
+
+import sys
+import os
+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.commands.run import main as run_main
+
+
+class TestCLIArguments(unittest.TestCase):
+    
+    def setUp(self):
+        """测试前准备"""
+        # 保存原始的sys.path
+        self.original_path = sys.path[:]
+        
+    def tearDown(self):
+        """测试后清理"""
+        # 恢复原始的sys.path
+        sys.path = self.original_path[:]
+    
+    @patch('crawlo.commands.run._find_project_root')
+    @patch('crawlo.commands.run.initialize_framework')
+    @patch('crawlo.commands.run.CrawlerProcess')
+    def test_log_level_argument(self, mock_crawler_process, mock_initialize, mock_find_project):
+        """测试--log-level参数解析"""
+        # 模拟项目环境
+        mock_find_project.return_value = os.path.join(os.path.dirname(__file__), '..')
+        mock_initialize.return_value = {}
+        mock_process_instance = MagicMock()
+        mock_crawler_process.return_value = mock_process_instance
+        mock_process_instance.get_spider_names.return_value = ['test_spider']
+        mock_process_instance.is_spider_registered.return_value = True
+        mock_process_instance.get_spider_class.return_value = MagicMock(__name__='TestSpider')
+        
+        # 测试参数解析
+        args = ['test_spider', '--log-level', 'DEBUG']
+        result = run_main(args)
+        
+        # 验证initialize_framework被调用时传入了正确的日志级别
+        mock_initialize.assert_called_once()
+        call_args = mock_initialize.call_args
+        if call_args and call_args[0]:  # 检查位置参数
+            settings = call_args[0][0]
+            self.assertEqual(settings.get('LOG_LEVEL'), 'DEBUG')
+        elif call_args and call_args[1]:  # 检查关键字参数
+            settings = call_args[1].get('custom_settings', {})
+            self.assertEqual(settings.get('LOG_LEVEL'), 'DEBUG')
+    
+    @patch('crawlo.commands.run._find_project_root')
+    @patch('crawlo.commands.run.initialize_framework')
+    @patch('crawlo.commands.run.CrawlerProcess')
+    def test_concurrency_argument(self, mock_crawler_process, mock_initialize, mock_find_project):
+        """测试--concurrency参数解析"""
+        # 模拟项目环境
+        mock_find_project.return_value = os.path.join(os.path.dirname(__file__), '..')
+        mock_initialize.return_value = {}
+        mock_process_instance = MagicMock()
+        mock_crawler_process.return_value = mock_process_instance
+        mock_process_instance.get_spider_names.return_value = ['test_spider']
+        mock_process_instance.is_spider_registered.return_value = True
+        mock_process_instance.get_spider_class.return_value = MagicMock(__name__='TestSpider')
+        
+        # 测试参数解析
+        args = ['test_spider', '--concurrency', '32']
+        result = run_main(args)
+        
+        # 验证initialize_framework被调用时传入了正确的并发数
+        mock_initialize.assert_called_once()
+        call_args = mock_initialize.call_args
+        if call_args and call_args[0]:  # 检查位置参数
+            settings = call_args[0][0]
+            self.assertEqual(settings.get('CONCURRENCY'), 32)
+        elif call_args and call_args[1]:  # 检查关键字参数
+            settings = call_args[1].get('custom_settings', {})
+            self.assertEqual(settings.get('CONCURRENCY'), 32)
+    
+    @patch('crawlo.commands.run._find_project_root')
+    @patch('crawlo.commands.run.initialize_framework')
+    @patch('crawlo.commands.run.CrawlerProcess')
+    def test_combined_arguments(self, mock_crawler_process, mock_initialize, mock_find_project):
+        """测试组合参数解析"""
+        # 模拟项目环境
+        mock_find_project.return_value = os.path.join(os.path.dirname(__file__), '..')
+        mock_initialize.return_value = {}
+        mock_process_instance = MagicMock()
+        mock_crawler_process.return_value = mock_process_instance
+        mock_process_instance.get_spider_names.return_value = ['test_spider']
+        mock_process_instance.is_spider_registered.return_value = True
+        mock_process_instance.get_spider_class.return_value = MagicMock(__name__='TestSpider')
+        
+        # 测试参数解析
+        args = ['test_spider', '--log-level', 'DEBUG', '--concurrency', '16']
+        result = run_main(args)
+        
+        # 验证initialize_framework被调用时传入了正确的参数
+        mock_initialize.assert_called_once()
+        call_args = mock_initialize.call_args
+        if call_args and call_args[0]:  # 检查位置参数
+            settings = call_args[0][0]
+            self.assertEqual(settings.get('LOG_LEVEL'), 'DEBUG')
+            self.assertEqual(settings.get('CONCURRENCY'), 16)
+        elif call_args and call_args[1]:  # 检查关键字参数
+            settings = call_args[1].get('custom_settings', {})
+            self.assertEqual(settings.get('LOG_LEVEL'), 'DEBUG')
+            self.assertEqual(settings.get('CONCURRENCY'), 16)
+
+
+if __name__ == '__main__':
+    unittest.main()

tests/test_component_factory.py

@@ -1,175 +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__':
+#!/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_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())