线程池

rieuse · rieuse · commit 1074a29cd468 · 2018-09-19T10:37:41.000+08:00
diff --git a/Python基础代码/线程池.py b/Python基础代码/线程池.py
@@ -0,0 +1,67 @@
+from concurrent.futures import ThreadPoolExecutor, as_completed
+import threading
+import requests
+import time
+
+
+def download_html(i):
+    url = f'https://www.baidu.com/s?ie=UTF-8&wd={i}'
+    response = requests.get(url).text
+    return response
+
+ids = list(range(100))
+
+# with ThreadPoolExecutor(max_workers=8) as exe:
+#     exe.map(download_html,ids)
+
+
+# 其他接口使用：
+from concurrent.futures import ThreadPoolExecutor, as_completed,wait
+
+
+executor = ThreadPoolExecutor(max_workers=8)
+
+# 通过 submit 提交执行的函数到线程中
+task1 = executor.submit(download_html, (1))
+task2 = executor.submit(download_html, (3))
+
+# done() 判断 task 是否完成
+print(task1.done())
+time.sleep(4)
+print(task1.done())
+
+# result() 获取 task 的执行结果 阻塞
+print(task1.result())
+
+# cancel() 取消任务，如果任务在执行中或者执行完了是不能取消的
+# 现在线程池是8 两个任务都会被提交任务去执行，如果 max_workers = 1，执行task2.cancel()就会成功取消
+print(task2.cancel())
+
+
+# as_completed() 获取已经成功的task的返回数据，阻塞
+# as_completed实际上是一个生成器，里面有 yield 会把已经完成的 future (task) 返回结果
+ids = list(range(10))
+all_task = [executor.submit(download_html,(i)) for i in ids]
+time.sleep(8)
+# 这是异步的，谁完成就处理谁
+for future in as_completed(all_task):
+    data = future.result()
+    print(f'html response {data}')
+
+
+# 通过 executor 获取已经完成的task
+for data in executor.map(download_html,ids):
+    print(f'html response {data}')
+
+
+# wait() 等待task完成
+ids = list(range(10))
+all_task = [executor.submit(download_html,(i)) for i in ids]
+
+#  wait 的 return_when 可选项
+FIRST_COMPLETED = 'FIRST_COMPLETED'
+FIRST_EXCEPTION = 'FIRST_EXCEPTION'
+ALL_COMPLETED = 'ALL_COMPLETED'
+_AS_COMPLETED = '_AS_COMPLETED'
+
+wait(all_task, return_when=ALL_COMPLETED)
