crawlo 1.2.6__py3-none-any.whl → 1.2.8__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.


This version of crawlo might be problematic. Click here for more details.

Files changed (209) hide show
  1. crawlo/__init__.py +61 -61
  2. crawlo/__version__.py +1 -1
  3. crawlo/cleaners/__init__.py +60 -60
  4. crawlo/cleaners/data_formatter.py +225 -225
  5. crawlo/cleaners/encoding_converter.py +125 -125
  6. crawlo/cleaners/text_cleaner.py +232 -232
  7. crawlo/cli.py +75 -88
  8. crawlo/commands/__init__.py +14 -14
  9. crawlo/commands/check.py +594 -594
  10. crawlo/commands/genspider.py +151 -151
  11. crawlo/commands/help.py +138 -144
  12. crawlo/commands/list.py +155 -155
  13. crawlo/commands/run.py +323 -323
  14. crawlo/commands/startproject.py +436 -436
  15. crawlo/commands/stats.py +187 -187
  16. crawlo/commands/utils.py +186 -186
  17. crawlo/config.py +312 -312
  18. crawlo/config_validator.py +251 -251
  19. crawlo/core/__init__.py +2 -2
  20. crawlo/core/engine.py +365 -356
  21. crawlo/core/processor.py +40 -40
  22. crawlo/core/scheduler.py +251 -239
  23. crawlo/crawler.py +1099 -1110
  24. crawlo/data/__init__.py +5 -5
  25. crawlo/data/user_agents.py +107 -107
  26. crawlo/downloader/__init__.py +266 -266
  27. crawlo/downloader/aiohttp_downloader.py +228 -221
  28. crawlo/downloader/cffi_downloader.py +256 -256
  29. crawlo/downloader/httpx_downloader.py +259 -259
  30. crawlo/downloader/hybrid_downloader.py +212 -212
  31. crawlo/downloader/playwright_downloader.py +402 -402
  32. crawlo/downloader/selenium_downloader.py +472 -472
  33. crawlo/event.py +11 -11
  34. crawlo/exceptions.py +81 -81
  35. crawlo/extension/__init__.py +39 -38
  36. crawlo/extension/health_check.py +141 -141
  37. crawlo/extension/log_interval.py +57 -57
  38. crawlo/extension/log_stats.py +81 -81
  39. crawlo/extension/logging_extension.py +43 -43
  40. crawlo/extension/memory_monitor.py +104 -104
  41. crawlo/extension/performance_profiler.py +133 -133
  42. crawlo/extension/request_recorder.py +107 -107
  43. crawlo/filters/__init__.py +154 -154
  44. crawlo/filters/aioredis_filter.py +234 -234
  45. crawlo/filters/memory_filter.py +269 -269
  46. crawlo/items/__init__.py +23 -23
  47. crawlo/items/base.py +21 -21
  48. crawlo/items/fields.py +52 -52
  49. crawlo/items/items.py +104 -104
  50. crawlo/middleware/__init__.py +21 -21
  51. crawlo/middleware/default_header.py +131 -131
  52. crawlo/middleware/download_delay.py +104 -104
  53. crawlo/middleware/middleware_manager.py +136 -135
  54. crawlo/middleware/offsite.py +114 -114
  55. crawlo/middleware/proxy.py +367 -367
  56. crawlo/middleware/request_ignore.py +86 -86
  57. crawlo/middleware/response_code.py +163 -163
  58. crawlo/middleware/response_filter.py +136 -136
  59. crawlo/middleware/retry.py +124 -124
  60. crawlo/mode_manager.py +211 -211
  61. crawlo/network/__init__.py +21 -21
  62. crawlo/network/request.py +338 -338
  63. crawlo/network/response.py +359 -359
  64. crawlo/pipelines/__init__.py +21 -21
  65. crawlo/pipelines/bloom_dedup_pipeline.py +156 -156
  66. crawlo/pipelines/console_pipeline.py +39 -39
  67. crawlo/pipelines/csv_pipeline.py +316 -316
  68. crawlo/pipelines/database_dedup_pipeline.py +222 -222
  69. crawlo/pipelines/json_pipeline.py +218 -218
  70. crawlo/pipelines/memory_dedup_pipeline.py +115 -115
  71. crawlo/pipelines/mongo_pipeline.py +131 -131
  72. crawlo/pipelines/mysql_pipeline.py +317 -317
  73. crawlo/pipelines/pipeline_manager.py +62 -61
  74. crawlo/pipelines/redis_dedup_pipeline.py +166 -165
  75. crawlo/project.py +314 -279
  76. crawlo/queue/pqueue.py +37 -37
  77. crawlo/queue/queue_manager.py +377 -376
  78. crawlo/queue/redis_priority_queue.py +306 -306
  79. crawlo/settings/__init__.py +7 -7
  80. crawlo/settings/default_settings.py +219 -215
  81. crawlo/settings/setting_manager.py +122 -122
  82. crawlo/spider/__init__.py +639 -639
  83. crawlo/stats_collector.py +59 -59
  84. crawlo/subscriber.py +129 -129
  85. crawlo/task_manager.py +30 -30
  86. crawlo/templates/crawlo.cfg.tmpl +10 -10
  87. crawlo/templates/project/__init__.py.tmpl +3 -3
  88. crawlo/templates/project/items.py.tmpl +17 -17
  89. crawlo/templates/project/middlewares.py.tmpl +118 -118
  90. crawlo/templates/project/pipelines.py.tmpl +96 -96
  91. crawlo/templates/project/settings.py.tmpl +288 -288
  92. crawlo/templates/project/settings_distributed.py.tmpl +157 -157
  93. crawlo/templates/project/settings_gentle.py.tmpl +100 -100
  94. crawlo/templates/project/settings_high_performance.py.tmpl +134 -134
  95. crawlo/templates/project/settings_simple.py.tmpl +98 -98
  96. crawlo/templates/project/spiders/__init__.py.tmpl +5 -5
  97. crawlo/templates/run.py.tmpl +47 -45
  98. crawlo/templates/spider/spider.py.tmpl +143 -143
  99. crawlo/tools/__init__.py +182 -182
  100. crawlo/tools/anti_crawler.py +268 -268
  101. crawlo/tools/authenticated_proxy.py +240 -240
  102. crawlo/tools/data_validator.py +180 -180
  103. crawlo/tools/date_tools.py +35 -35
  104. crawlo/tools/distributed_coordinator.py +386 -386
  105. crawlo/tools/retry_mechanism.py +220 -220
  106. crawlo/tools/scenario_adapter.py +262 -262
  107. crawlo/utils/__init__.py +35 -35
  108. crawlo/utils/batch_processor.py +259 -259
  109. crawlo/utils/controlled_spider_mixin.py +439 -439
  110. crawlo/utils/date_tools.py +290 -290
  111. crawlo/utils/db_helper.py +343 -343
  112. crawlo/utils/enhanced_error_handler.py +356 -356
  113. crawlo/utils/env_config.py +143 -106
  114. crawlo/utils/error_handler.py +123 -123
  115. crawlo/utils/func_tools.py +82 -82
  116. crawlo/utils/large_scale_config.py +286 -286
  117. crawlo/utils/large_scale_helper.py +344 -344
  118. crawlo/utils/log.py +128 -128
  119. crawlo/utils/performance_monitor.py +285 -285
  120. crawlo/utils/queue_helper.py +175 -175
  121. crawlo/utils/redis_connection_pool.py +351 -351
  122. crawlo/utils/redis_key_validator.py +198 -198
  123. crawlo/utils/request.py +267 -267
  124. crawlo/utils/request_serializer.py +218 -218
  125. crawlo/utils/spider_loader.py +61 -61
  126. crawlo/utils/system.py +11 -11
  127. crawlo/utils/tools.py +4 -4
  128. crawlo/utils/url.py +39 -39
  129. {crawlo-1.2.6.dist-info → crawlo-1.2.8.dist-info}/METADATA +764 -764
  130. crawlo-1.2.8.dist-info/RECORD +209 -0
  131. examples/__init__.py +7 -7
  132. tests/DOUBLE_CRAWLO_PREFIX_FIX_REPORT.md +81 -81
  133. tests/__init__.py +7 -7
  134. tests/advanced_tools_example.py +275 -275
  135. tests/authenticated_proxy_example.py +236 -236
  136. tests/cleaners_example.py +160 -160
  137. tests/config_validation_demo.py +102 -102
  138. tests/controlled_spider_example.py +205 -205
  139. tests/date_tools_example.py +180 -180
  140. tests/dynamic_loading_example.py +523 -523
  141. tests/dynamic_loading_test.py +104 -104
  142. tests/env_config_example.py +133 -133
  143. tests/error_handling_example.py +171 -171
  144. tests/redis_key_validation_demo.py +130 -130
  145. tests/response_improvements_example.py +144 -144
  146. tests/test_advanced_tools.py +148 -148
  147. tests/test_all_redis_key_configs.py +145 -145
  148. tests/test_authenticated_proxy.py +141 -141
  149. tests/test_cleaners.py +54 -54
  150. tests/test_comprehensive.py +146 -146
  151. tests/test_config_consistency.py +81 -0
  152. tests/test_config_validator.py +193 -193
  153. tests/test_crawlo_proxy_integration.py +172 -172
  154. tests/test_date_tools.py +123 -123
  155. tests/test_default_header_middleware.py +158 -158
  156. tests/test_double_crawlo_fix.py +207 -207
  157. tests/test_double_crawlo_fix_simple.py +124 -124
  158. tests/test_download_delay_middleware.py +221 -221
  159. tests/test_downloader_proxy_compatibility.py +268 -268
  160. tests/test_dynamic_downloaders_proxy.py +124 -124
  161. tests/test_dynamic_proxy.py +92 -92
  162. tests/test_dynamic_proxy_config.py +146 -146
  163. tests/test_dynamic_proxy_real.py +109 -109
  164. tests/test_edge_cases.py +303 -303
  165. tests/test_enhanced_error_handler.py +270 -270
  166. tests/test_env_config.py +121 -121
  167. tests/test_error_handler_compatibility.py +112 -112
  168. tests/test_final_validation.py +153 -153
  169. tests/test_framework_env_usage.py +103 -103
  170. tests/test_integration.py +356 -356
  171. tests/test_item_dedup_redis_key.py +122 -122
  172. tests/test_mode_consistency.py +52 -0
  173. tests/test_offsite_middleware.py +221 -221
  174. tests/test_parsel.py +29 -29
  175. tests/test_performance.py +327 -327
  176. tests/test_proxy_api.py +264 -264
  177. tests/test_proxy_health_check.py +32 -32
  178. tests/test_proxy_middleware.py +121 -121
  179. tests/test_proxy_middleware_enhanced.py +216 -216
  180. tests/test_proxy_middleware_integration.py +136 -136
  181. tests/test_proxy_providers.py +56 -56
  182. tests/test_proxy_stats.py +19 -19
  183. tests/test_proxy_strategies.py +59 -59
  184. tests/test_queue_manager_double_crawlo.py +173 -173
  185. tests/test_queue_manager_redis_key.py +176 -176
  186. tests/test_real_scenario_proxy.py +195 -195
  187. tests/test_redis_config.py +28 -28
  188. tests/test_redis_connection_pool.py +294 -294
  189. tests/test_redis_key_naming.py +181 -181
  190. tests/test_redis_key_validator.py +123 -123
  191. tests/test_redis_queue.py +224 -224
  192. tests/test_request_ignore_middleware.py +182 -182
  193. tests/test_request_serialization.py +70 -70
  194. tests/test_response_code_middleware.py +349 -349
  195. tests/test_response_filter_middleware.py +427 -427
  196. tests/test_response_improvements.py +152 -152
  197. tests/test_retry_middleware.py +241 -241
  198. tests/test_scheduler.py +252 -241
  199. tests/test_scheduler_config_update.py +134 -0
  200. tests/test_simple_response.py +61 -61
  201. tests/test_telecom_spider_redis_key.py +205 -205
  202. tests/test_template_content.py +87 -87
  203. tests/test_template_redis_key.py +134 -134
  204. tests/test_tools.py +153 -153
  205. tests/tools_example.py +257 -257
  206. crawlo-1.2.6.dist-info/RECORD +0 -206
  207. {crawlo-1.2.6.dist-info → crawlo-1.2.8.dist-info}/WHEEL +0 -0
  208. {crawlo-1.2.6.dist-info → crawlo-1.2.8.dist-info}/entry_points.txt +0 -0
  209. {crawlo-1.2.6.dist-info → crawlo-1.2.8.dist-info}/top_level.txt +0 -0
crawlo/data/__init__.py CHANGED
@@ -1,6 +1,6 @@
1
- #!/usr/bin/env python
2
- # -*- coding: utf-8 -*-
3
- """
4
- Crawlo 数据模块
5
- 包含用户代理、字典等数据文件
1
+ #!/usr/bin/env python
2
+ # -*- coding: utf-8 -*-
3
+ """
4
+ Crawlo 数据模块
5
+ 包含用户代理、字典等数据文件
6
6
  """
@@ -1,108 +1,108 @@
1
- #!/usr/bin/env python
2
- # -*- coding: utf-8 -*-
3
- """
4
- User-Agent列表
5
- 包含各种设备和浏览器的User-Agent字符串,用于爬虫伪装
6
- """
7
-
8
- # 桌面浏览器User-Agent
9
- DESKTOP_USER_AGENTS = [
10
- # Chrome
11
- "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
12
- "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
13
- "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
14
- "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
15
-
16
- # Firefox
17
- "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:125.0) Gecko/20100101 Firefox/125.0",
18
- "Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:125.0) Gecko/20100101 Firefox/125.0",
19
- "Mozilla/5.0 (X11; Linux i686; rv:125.0) Gecko/20100101 Firefox/125.0",
20
- "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:125.0) Gecko/20100101 Firefox/125.0",
21
-
22
- # Safari
23
- "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.4 Safari/605.1.15",
24
- "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Safari/605.1.15",
25
-
26
- # Edge
27
- "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36 Edg/124.0.2478.80",
28
- "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36 Edg/123.0.2420.40",
29
-
30
- # Internet Explorer
31
- "Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko",
32
- ]
33
-
34
- # 移动设备User-Agent
35
- MOBILE_USER_AGENTS = [
36
- # iPhone
37
- "Mozilla/5.0 (iPhone; CPU iPhone OS 17_4 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.4 Mobile/15E148 Safari/604.1",
38
- "Mozilla/5.0 (iPhone; CPU iPhone OS 16_6 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1",
39
-
40
- # iPad
41
- "Mozilla/5.0 (iPad; CPU OS 17_4 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.4 Mobile/15E148 Safari/604.1",
42
- "Mozilla/5.0 (iPad; CPU OS 16_6 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1",
43
-
44
- # Android
45
- "Mozilla/5.0 (Linux; Android 14; SM-S911B) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
46
- "Mozilla/5.0 (Linux; Android 13; SM-S908B) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
47
- "Mozilla/5.0 (Linux; Android 12; Pixel 6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
48
- "Mozilla/5.0 (Linux; Android 11; Pixel 5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
49
-
50
- # Android平板
51
- "Mozilla/5.0 (Linux; Android 13; SM-X906C Build/TP1A.220624.014; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/124.0.6367.118 Safari/537.36",
52
- ]
53
-
54
- # 爬虫/机器人User-Agent (用于测试)
55
- BOT_USER_AGENTS = [
56
- "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)",
57
- "Mozilla/5.0 (compatible; Bingbot/2.0; +http://www.bing.com/bingbot.htm)",
58
- "Mozilla/5.0 (compatible; Baiduspider/2.0; +http://www.baidu.com/search/spider.html)",
59
- "Mozilla/5.0 (compatible; Yahoo! Slurp; http://help.yahoo.com/help/us/ysearch/slurp)",
60
- ]
61
-
62
- # 所有User-Agent的组合列表
63
- ALL_USER_AGENTS = DESKTOP_USER_AGENTS + MOBILE_USER_AGENTS
64
-
65
- # 按设备类型分类的User-Agent字典
66
- USER_AGENTS_BY_TYPE = {
67
- "desktop": DESKTOP_USER_AGENTS,
68
- "mobile": MOBILE_USER_AGENTS,
69
- "bot": BOT_USER_AGENTS,
70
- "all": ALL_USER_AGENTS,
71
- }
72
-
73
- def get_user_agents(device_type="all"):
74
- """
75
- 获取指定类型的User-Agent列表
76
-
77
- Args:
78
- device_type (str): 设备类型,可选值: "desktop", "mobile", "bot", "all"
79
-
80
- Returns:
81
- list: User-Agent字符串列表
82
- """
83
- return USER_AGENTS_BY_TYPE.get(device_type, ALL_USER_AGENTS)
84
-
85
- def get_random_user_agent(device_type="all"):
86
- """
87
- 获取随机User-Agent
88
-
89
- Args:
90
- device_type (str): 设备类型,可选值: "desktop", "mobile", "bot", "all"
91
-
92
- Returns:
93
- str: 随机User-Agent字符串
94
- """
95
- import random
96
- user_agents = get_user_agents(device_type)
97
- return random.choice(user_agents) if user_agents else ""
98
-
99
- # 导出常用的User-Agent列表
100
- __all__ = [
101
- "DESKTOP_USER_AGENTS",
102
- "MOBILE_USER_AGENTS",
103
- "BOT_USER_AGENTS",
104
- "ALL_USER_AGENTS",
105
- "USER_AGENTS_BY_TYPE",
106
- "get_user_agents",
107
- "get_random_user_agent"
1
+ #!/usr/bin/env python
2
+ # -*- coding: utf-8 -*-
3
+ """
4
+ User-Agent列表
5
+ 包含各种设备和浏览器的User-Agent字符串,用于爬虫伪装
6
+ """
7
+
8
+ # 桌面浏览器User-Agent
9
+ DESKTOP_USER_AGENTS = [
10
+ # Chrome
11
+ "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
12
+ "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
13
+ "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
14
+ "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
15
+
16
+ # Firefox
17
+ "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:125.0) Gecko/20100101 Firefox/125.0",
18
+ "Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:125.0) Gecko/20100101 Firefox/125.0",
19
+ "Mozilla/5.0 (X11; Linux i686; rv:125.0) Gecko/20100101 Firefox/125.0",
20
+ "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:125.0) Gecko/20100101 Firefox/125.0",
21
+
22
+ # Safari
23
+ "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.4 Safari/605.1.15",
24
+ "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Safari/605.1.15",
25
+
26
+ # Edge
27
+ "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36 Edg/124.0.2478.80",
28
+ "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36 Edg/123.0.2420.40",
29
+
30
+ # Internet Explorer
31
+ "Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko",
32
+ ]
33
+
34
+ # 移动设备User-Agent
35
+ MOBILE_USER_AGENTS = [
36
+ # iPhone
37
+ "Mozilla/5.0 (iPhone; CPU iPhone OS 17_4 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.4 Mobile/15E148 Safari/604.1",
38
+ "Mozilla/5.0 (iPhone; CPU iPhone OS 16_6 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1",
39
+
40
+ # iPad
41
+ "Mozilla/5.0 (iPad; CPU OS 17_4 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.4 Mobile/15E148 Safari/604.1",
42
+ "Mozilla/5.0 (iPad; CPU OS 16_6 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1",
43
+
44
+ # Android
45
+ "Mozilla/5.0 (Linux; Android 14; SM-S911B) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
46
+ "Mozilla/5.0 (Linux; Android 13; SM-S908B) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
47
+ "Mozilla/5.0 (Linux; Android 12; Pixel 6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
48
+ "Mozilla/5.0 (Linux; Android 11; Pixel 5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.6367.118 Mobile Safari/537.36",
49
+
50
+ # Android平板
51
+ "Mozilla/5.0 (Linux; Android 13; SM-X906C Build/TP1A.220624.014; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/124.0.6367.118 Safari/537.36",
52
+ ]
53
+
54
+ # 爬虫/机器人User-Agent (用于测试)
55
+ BOT_USER_AGENTS = [
56
+ "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)",
57
+ "Mozilla/5.0 (compatible; Bingbot/2.0; +http://www.bing.com/bingbot.htm)",
58
+ "Mozilla/5.0 (compatible; Baiduspider/2.0; +http://www.baidu.com/search/spider.html)",
59
+ "Mozilla/5.0 (compatible; Yahoo! Slurp; http://help.yahoo.com/help/us/ysearch/slurp)",
60
+ ]
61
+
62
+ # 所有User-Agent的组合列表
63
+ ALL_USER_AGENTS = DESKTOP_USER_AGENTS + MOBILE_USER_AGENTS
64
+
65
+ # 按设备类型分类的User-Agent字典
66
+ USER_AGENTS_BY_TYPE = {
67
+ "desktop": DESKTOP_USER_AGENTS,
68
+ "mobile": MOBILE_USER_AGENTS,
69
+ "bot": BOT_USER_AGENTS,
70
+ "all": ALL_USER_AGENTS,
71
+ }
72
+
73
+ def get_user_agents(device_type="all"):
74
+ """
75
+ 获取指定类型的User-Agent列表
76
+
77
+ Args:
78
+ device_type (str): 设备类型,可选值: "desktop", "mobile", "bot", "all"
79
+
80
+ Returns:
81
+ list: User-Agent字符串列表
82
+ """
83
+ return USER_AGENTS_BY_TYPE.get(device_type, ALL_USER_AGENTS)
84
+
85
+ def get_random_user_agent(device_type="all"):
86
+ """
87
+ 获取随机User-Agent
88
+
89
+ Args:
90
+ device_type (str): 设备类型,可选值: "desktop", "mobile", "bot", "all"
91
+
92
+ Returns:
93
+ str: 随机User-Agent字符串
94
+ """
95
+ import random
96
+ user_agents = get_user_agents(device_type)
97
+ return random.choice(user_agents) if user_agents else ""
98
+
99
+ # 导出常用的User-Agent列表
100
+ __all__ = [
101
+ "DESKTOP_USER_AGENTS",
102
+ "MOBILE_USER_AGENTS",
103
+ "BOT_USER_AGENTS",
104
+ "ALL_USER_AGENTS",
105
+ "USER_AGENTS_BY_TYPE",
106
+ "get_user_agents",
107
+ "get_random_user_agent"
108
108
  ]