crawlo 1.3.6__py3-none-any.whl → 1.3.7__py3-none-any.whl

tests/test_item_dedup_redis_key.py

@@ -1,123 +1,123 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-"""
-数据项去重Redis Key测试脚本
-用于验证RedisDedupPipeline和示例项目中的Redis去重管道是否使用统一的Redis key命名规范
-"""
-import sys
-import os
-import traceback
-
-# 添加项目根目录到路径
-sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
-
-from crawlo.pipelines.redis_dedup_pipeline import RedisDedupPipeline
-
-
-class MockSettings:
-    """模拟设置类"""
-    def __init__(self, project_name="test_project"):
-        self.project_name = project_name
-    
-    def get(self, key, default=None):
-        if key == 'PROJECT_NAME':
-            return self.project_name
-        elif key == 'REDIS_HOST':
-            return 'localhost'
-        elif key == 'REDIS_PORT':
-            return 6379
-        elif key == 'REDIS_DB':
-            return 2
-        elif key == 'REDIS_PASSWORD':
-            return None
-        elif key == 'LOG_LEVEL':
-            return 'INFO'
-        return default
-    
-    def getint(self, key, default=0):
-        if key == 'REDIS_PORT':
-            return 6379
-        elif key == 'REDIS_DB':
-            return 2
-        return default
-
-
-class MockCrawler:
-    """模拟爬虫类"""
-    def __init__(self, project_name="test_project"):
-        self.settings = MockSettings(project_name)
-
-
-async def test_item_dedup_redis_key():
-    """测试数据项去重Redis key命名规范"""
-    print("测试数据项去重Redis key命名规范...")
-    
-    try:
-        # 测试不同的项目名称
-        test_cases = [
-            {
-                "project_name": "books_distributed",
-                "expected_key": "crawlo:books_distributed:item:fingerprint",
-                "description": "书籍分布式项目"
-            },
-            {
-                "project_name": "api_data_collection",
-                "expected_key": "crawlo:api_data_collection:item:fingerprint",
-                "description": "API数据采集项目"
-            },
-            {
-                "project_name": "test_project",
-                "expected_key": "crawlo:test_project:item:fingerprint",
-                "description": "测试项目"
-            }
-        ]
-        
-        for i, test_case in enumerate(test_cases, 1):
-            print(f"   {i}. 测试 {test_case['description']}...")
-            
-            # 测试RedisDedupPipeline
-            mock_crawler = MockCrawler(test_case["project_name"])
-            pipeline = RedisDedupPipeline.from_crawler(mock_crawler)
-            
-            # 验证Redis key是否符合规范
-            assert pipeline.redis_key == test_case["expected_key"], \
-                f"Redis key不匹配: {pipeline.redis_key} != {test_case['expected_key']}"
-            
-            print(f"      Redis key: {pipeline.redis_key}")
-        
-        print("数据项去重Redis key命名规范测试通过！")
-        return True
-        
-    except Exception as e:
-        print(f"数据项去重Redis key命名规范测试失败: {e}")
-        traceback.print_exc()
-        return False
-
-
-def main():
-    """主测试函数"""
-    print("开始数据项去重Redis key命名规范测试...")
-    print("=" * 50)
-    
-    try:
-        success = test_item_dedup_redis_key()
-        
-        print("=" * 50)
-        if success:
-            print("所有测试通过！数据项去重使用统一的Redis key命名规范")
-        else:
-            print("测试失败，请检查实现")
-            return 1
-            
-    except Exception as e:
-        print("=" * 50)
-        print(f"测试过程中发生异常: {e}")
-        traceback.print_exc()
-        return 1
-    
-    return 0
-
-
-if __name__ == "__main__":
-    exit_code = main()
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+数据项去重Redis Key测试脚本
+用于验证RedisDedupPipeline和示例项目中的Redis去重管道是否使用统一的Redis key命名规范
+"""
+import sys
+import os
+import traceback
+
+# 添加项目根目录到路径
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
+
+from crawlo.pipelines.redis_dedup_pipeline import RedisDedupPipeline
+
+
+class MockSettings:
+    """模拟设置类"""
+    def __init__(self, project_name="test_project"):
+        self.project_name = project_name
+    
+    def get(self, key, default=None):
+        if key == 'PROJECT_NAME':
+            return self.project_name
+        elif key == 'REDIS_HOST':
+            return 'localhost'
+        elif key == 'REDIS_PORT':
+            return 6379
+        elif key == 'REDIS_DB':
+            return 2
+        elif key == 'REDIS_PASSWORD':
+            return None
+        elif key == 'LOG_LEVEL':
+            return 'INFO'
+        return default
+    
+    def getint(self, key, default=0):
+        if key == 'REDIS_PORT':
+            return 6379
+        elif key == 'REDIS_DB':
+            return 2
+        return default
+
+
+class MockCrawler:
+    """模拟爬虫类"""
+    def __init__(self, project_name="test_project"):
+        self.settings = MockSettings(project_name)
+
+
+async def test_item_dedup_redis_key():
+    """测试数据项去重Redis key命名规范"""
+    print("测试数据项去重Redis key命名规范...")
+    
+    try:
+        # 测试不同的项目名称
+        test_cases = [
+            {
+                "project_name": "books_distributed",
+                "expected_key": "crawlo:books_distributed:item:fingerprint",
+                "description": "书籍分布式项目"
+            },
+            {
+                "project_name": "api_data_collection",
+                "expected_key": "crawlo:api_data_collection:item:fingerprint",
+                "description": "API数据采集项目"
+            },
+            {
+                "project_name": "test_project",
+                "expected_key": "crawlo:test_project:item:fingerprint",
+                "description": "测试项目"
+            }
+        ]
+        
+        for i, test_case in enumerate(test_cases, 1):
+            print(f"   {i}. 测试 {test_case['description']}...")
+            
+            # 测试RedisDedupPipeline
+            mock_crawler = MockCrawler(test_case["project_name"])
+            pipeline = RedisDedupPipeline.from_crawler(mock_crawler)
+            
+            # 验证Redis key是否符合规范
+            assert pipeline.redis_key == test_case["expected_key"], \
+                f"Redis key不匹配: {pipeline.redis_key} != {test_case['expected_key']}"
+            
+            print(f"      Redis key: {pipeline.redis_key}")
+        
+        print("数据项去重Redis key命名规范测试通过！")
+        return True
+        
+    except Exception as e:
+        print(f"数据项去重Redis key命名规范测试失败: {e}")
+        traceback.print_exc()
+        return False
+
+
+def main():
+    """主测试函数"""
+    print("开始数据项去重Redis key命名规范测试...")
+    print("=" * 50)
+    
+    try:
+        success = test_item_dedup_redis_key()
+        
+        print("=" * 50)
+        if success:
+            print("所有测试通过！数据项去重使用统一的Redis key命名规范")
+        else:
+            print("测试失败，请检查实现")
+            return 1
+            
+    except Exception as e:
+        print("=" * 50)
+        print(f"测试过程中发生异常: {e}")
+        traceback.print_exc()
+        return 1
+    
+    return 0
+
+
+if __name__ == "__main__":
+    exit_code = main()
     sys.exit(exit_code)

tests/test_large_scale_config.py

@@ -1,113 +1,113 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-"""
-大规模配置工具测试
-测试 LargeScaleConfig, apply_large_scale_config
-"""
-import sys
-import os
-import unittest
-from unittest.mock import Mock, 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.large_scale_config import LargeScaleConfig, apply_large_scale_config
-
-
-class TestLargeScaleConfig(unittest.TestCase):
-    """大规模配置工具测试"""
-
-    def test_conservative_config(self):
-        """测试保守配置"""
-        config = LargeScaleConfig.conservative_config(concurrency=8)
-        
-        # 验证基本配置
-        self.assertEqual(config['CONCURRENCY'], 8)
-        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 80)  # 8 * 10
-        self.assertEqual(config['DOWNLOAD_DELAY'], 0.2)
-        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 1)
-        
-        # 验证连接池配置
-        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 16)  # 8 * 2
-        
-        # 验证重试配置
-        self.assertEqual(config['MAX_RETRY_TIMES'], 2)
-        
-    def test_balanced_config(self):
-        """测试平衡配置"""
-        config = LargeScaleConfig.balanced_config(concurrency=16)
-        
-        # 验证基本配置
-        self.assertEqual(config['CONCURRENCY'], 16)
-        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 240)  # 16 * 15
-        self.assertEqual(config['DOWNLOAD_DELAY'], 0.1)
-        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 2)
-        
-        # 验证连接池配置
-        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 48)  # 16 * 3
-        
-        # 验证重试配置
-        self.assertEqual(config['MAX_RETRY_TIMES'], 3)
-        
-    def test_aggressive_config(self):
-        """测试激进配置"""
-        config = LargeScaleConfig.aggressive_config(concurrency=32)
-        
-        # 验证基本配置
-        self.assertEqual(config['CONCURRENCY'], 32)
-        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 640)  # 32 * 20
-        self.assertEqual(config['DOWNLOAD_DELAY'], 0.05)
-        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 3)
-        
-        # 验证连接池配置
-        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 128)  # 32 * 4
-        
-        # 验证重试配置
-        self.assertEqual(config['MAX_RETRY_TIMES'], 5)
-        
-    def test_memory_optimized_config(self):
-        """测试内存优化配置"""
-        config = LargeScaleConfig.memory_optimized_config(concurrency=12)
-        
-        # 验证基本配置
-        self.assertEqual(config['CONCURRENCY'], 12)
-        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 60)  # 12 * 5
-        self.assertEqual(config['DOWNLOAD_DELAY'], 0.1)
-        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 1)
-        
-        # 验证连接池配置
-        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 12)  # 12 * 1
-        
-        # 验证内存限制配置
-        self.assertEqual(config['DOWNLOAD_MAXSIZE'], 2 * 1024 * 1024)  # 2MB
-        self.assertEqual(config['DOWNLOAD_WARN_SIZE'], 512 * 1024)  # 512KB
-        
-        # 验证重试配置
-        self.assertEqual(config['MAX_RETRY_TIMES'], 2)
-        
-    def test_apply_large_scale_config(self):
-        """测试应用大规模配置"""
-        settings_dict = {}
-        
-        # 应用平衡配置
-        apply_large_scale_config(settings_dict, "balanced", 16)
-        
-        # 验证配置已应用
-        self.assertEqual(settings_dict['CONCURRENCY'], 16)
-        self.assertEqual(settings_dict['SCHEDULER_MAX_QUEUE_SIZE'], 240)
-        self.assertEqual(settings_dict['DOWNLOAD_DELAY'], 0.1)
-        
-    def test_apply_large_scale_config_invalid_type(self):
-        """测试应用无效的大规模配置类型"""
-        settings_dict = {}
-        
-        # 应用无效配置类型应该抛出异常
-        with self.assertRaises(ValueError) as context:
-            apply_large_scale_config(settings_dict, "invalid_type", 16)
-            
-        self.assertIn("不支持的配置类型", str(context.exception))
-
-
-if __name__ == '__main__':
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+大规模配置工具测试
+测试 LargeScaleConfig, apply_large_scale_config
+"""
+import sys
+import os
+import unittest
+from unittest.mock import Mock, 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.large_scale_config import LargeScaleConfig, apply_large_scale_config
+
+
+class TestLargeScaleConfig(unittest.TestCase):
+    """大规模配置工具测试"""
+
+    def test_conservative_config(self):
+        """测试保守配置"""
+        config = LargeScaleConfig.conservative_config(concurrency=8)
+        
+        # 验证基本配置
+        self.assertEqual(config['CONCURRENCY'], 8)
+        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 80)  # 8 * 10
+        self.assertEqual(config['DOWNLOAD_DELAY'], 0.2)
+        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 1)
+        
+        # 验证连接池配置
+        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 16)  # 8 * 2
+        
+        # 验证重试配置
+        self.assertEqual(config['MAX_RETRY_TIMES'], 2)
+        
+    def test_balanced_config(self):
+        """测试平衡配置"""
+        config = LargeScaleConfig.balanced_config(concurrency=16)
+        
+        # 验证基本配置
+        self.assertEqual(config['CONCURRENCY'], 16)
+        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 240)  # 16 * 15
+        self.assertEqual(config['DOWNLOAD_DELAY'], 0.1)
+        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 2)
+        
+        # 验证连接池配置
+        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 48)  # 16 * 3
+        
+        # 验证重试配置
+        self.assertEqual(config['MAX_RETRY_TIMES'], 3)
+        
+    def test_aggressive_config(self):
+        """测试激进配置"""
+        config = LargeScaleConfig.aggressive_config(concurrency=32)
+        
+        # 验证基本配置
+        self.assertEqual(config['CONCURRENCY'], 32)
+        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 640)  # 32 * 20
+        self.assertEqual(config['DOWNLOAD_DELAY'], 0.05)
+        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 3)
+        
+        # 验证连接池配置
+        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 128)  # 32 * 4
+        
+        # 验证重试配置
+        self.assertEqual(config['MAX_RETRY_TIMES'], 5)
+        
+    def test_memory_optimized_config(self):
+        """测试内存优化配置"""
+        config = LargeScaleConfig.memory_optimized_config(concurrency=12)
+        
+        # 验证基本配置
+        self.assertEqual(config['CONCURRENCY'], 12)
+        self.assertEqual(config['SCHEDULER_MAX_QUEUE_SIZE'], 60)  # 12 * 5
+        self.assertEqual(config['DOWNLOAD_DELAY'], 0.1)
+        self.assertEqual(config['MAX_RUNNING_SPIDERS'], 1)
+        
+        # 验证连接池配置
+        self.assertEqual(config['CONNECTION_POOL_LIMIT'], 12)  # 12 * 1
+        
+        # 验证内存限制配置
+        self.assertEqual(config['DOWNLOAD_MAXSIZE'], 2 * 1024 * 1024)  # 2MB
+        self.assertEqual(config['DOWNLOAD_WARN_SIZE'], 512 * 1024)  # 512KB
+        
+        # 验证重试配置
+        self.assertEqual(config['MAX_RETRY_TIMES'], 2)
+        
+    def test_apply_large_scale_config(self):
+        """测试应用大规模配置"""
+        settings_dict = {}
+        
+        # 应用平衡配置
+        apply_large_scale_config(settings_dict, "balanced", 16)
+        
+        # 验证配置已应用
+        self.assertEqual(settings_dict['CONCURRENCY'], 16)
+        self.assertEqual(settings_dict['SCHEDULER_MAX_QUEUE_SIZE'], 240)
+        self.assertEqual(settings_dict['DOWNLOAD_DELAY'], 0.1)
+        
+    def test_apply_large_scale_config_invalid_type(self):
+        """测试应用无效的大规模配置类型"""
+        settings_dict = {}
+        
+        # 应用无效配置类型应该抛出异常
+        with self.assertRaises(ValueError) as context:
+            apply_large_scale_config(settings_dict, "invalid_type", 16)
+            
+        self.assertIn("不支持的配置类型", str(context.exception))
+
+
+if __name__ == '__main__':
     unittest.main()