Python多线程分片下载远端大文件 - threading paramiko

Python多xian程分片下载远端大文件 ，可以按照以下流程设计代码框架：

1.  导入需要的模块：首先，导入所需的模块，包括paramiko、os、time和threading。
2.  创建下载函数：创建一个用于分片下载文件的函数。该函数将使用SSH连接到远程服务器，并使用SFTP协议下载文件的指定分片到本地路径。使用sftp_file.seek() 和 file.seek() 确保正确的块被下载
3.  主函数：在主函数中，您需要设置远程服务器的主机名、用户名、密码、远程文件路径和本地存储路径。还需要确定要划分的分片数量。我们使用os.cpu_count()获取当前系统的CPU数量，并将文件分割成块，即多线程的数量为，2倍CPU个数以获取最大性能。我们创建一个Thread实例，每个实例负责一个文件块的下载。我们再次使用thread.join()等待所有进程完成，在完成后输出一条完成信息。
4.  运行主函数：在主函数中，我们创建了一个进程池，并使用Threading方法来并发地调用分片下载函数。通过确定每个xian程所负责的起始字节和结束字节，从而实现对服务器端文件的分片下载。然后，我们关闭并等待进程池中的所有任务完成。

Python多线程分片下载远端大文件源码（多线程单独封装函数）：

import paramiko
import threading
import time
import os

def get_remote_file_size(ssh_info, remote_path):
    ssh = paramiko.SSHClient()
    ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
    ssh.connect(**ssh_info)
    sftp = ssh.open_sftp()
    # get remote file size
    remote_file_size = sftp.stat(remote_path).st_size
    print ("remote_file_size:{}".format(remote_file_size))
    sftp.close()
    ssh.close()
    return remote_file_size

def download_chunk_file(ssh_info, remote_path, local_path, start_pos, end_pos):
    print("download_chunk_file start")
    client = paramiko.SSHClient()
    client.set_missing_host_key_policy(paramiko.AutoAddPolicy())
    client.connect(**ssh_info)

    sftp = client.open_sftp()
    # open both local and remote file
    local_file = open(local_path, "r+b")
    remote_file = sftp.open(remote_path, "rb")

    # right shift to same start position
    local_file.seek(start_pos)
    remote_file.seek(start_pos)

    # print("start_pos ~ end_pos: {} ~ {}".format(start_pos, end_pos))
    while True:
        # read chunk file from remote file
        read_start = time.time()
        buffer = remote_file.read(end_pos - start_pos)
        if not buffer:
            break
        else:
            print("read  cost time {:.2f}s".format(time.time() - read_start))
        # write chunk file to local file
        write_start = time.time()
        local_file.write(buffer)
        print("write cost time {:.2f}s".format(time.time() - write_start))
    print("chunk file with start_pos ~ end_pos: {} ~ {}, Download successfully!".format(start_pos, end_pos))
    remote_file.close()
    local_file.close()
    client.close()
    print("download_chunk_file end")

def download_multithreading(ssh_info, remote_path, local_path):
    # get number of CPU with high efficiency
    # double the number of CPU cores for optimal performance
    num_threads = os.cpu_count() * 2
    #num_threads = 1
    print("number of CPU is {}, number of thread is {}".format(os.cpu_count(),num_threads))
    # get remote file size
    file_size = get_remote_file_size(ssh_info, remote_path)
    # create new empty local file, same size with remote file
    with open(local_path, "wb") as f:
        f.truncate(file_size)

    # calculate the chunk size, n-1 threading handle each chunk size sub-file, and last threading handle last remaining sub-file
    chunk_size = file_size // num_threads
    print("chunk_size is {}".format(chunk_size))
    # create number of thread
    threads = []
    # create a thread for each chunk
    for index in range(num_threads):
        #print("thread: {}".format(index))
        start_pos = index * chunk_size
        end_pos = start_pos + chunk_size
        # last thread will download the remaining bytes
        if index == num_threads - 1:
            end_pos = file_size - 1

        args = (ssh_info, remote_path, local_path, start_pos, end_pos)
        thread = threading.Thread(target=download_chunk_file, args=args)

        thread.start()
        print(thread)
        threads.append(thread)

    # wait for all the threads to finish
    for thread in threads:
        thread.join()
        print(thread)

def main():
    host = "host"
    port = 22
    username = "username"
    password = "password"
 
    remote_path = '/remote_dir/remote_file'
    local_path = '/local_dir/local_file'

    ssh_info = {
        "hostname": host,
        "port": port,
        "username": username,
        "password": password,
    }



    multithread_download_start = time.time()
    download_multithreading(ssh_info, remote_path, local_path)
    multithread_download_end = time.time()
    multithread_download_cost = multithread_download_end - multithread_download_start
    print("Full file Download successfully! Cost time: {:.2f}s".format(multithread_download_cost))


if __name__ == "__main__":
    main()

运行结果输出：

$ python multi_thread_download_single_bigfile_def.py
number of CPU is 4, number of thread is 8
remote_file_size:63376366
chunk_size is 7922045
download_chunk_file start

download_chunk_file start

download_chunk_file start

download_chunk_file start

download_chunk_file start

download_chunk_file start

download_chunk_file start

download_chunk_file start

read  cost time 3.84s
write cost time 0.01s
read  cost time 3.94s
read  cost time 4.00s
write cost time 0.02s
write cost time 0.02s
read  cost time 3.88s
write cost time 0.01s
read  cost time 3.96s
write cost time 0.01s
read  cost time 4.01s
read  cost time 4.00s
write cost time 0.01s
write cost time 0.02s
read  cost time 4.29s
write cost time 0.01s
read  cost time 0.01s
write cost time 0.00s
chunk file with start_pos ~ end_pos: 55454315 ~ 63376365, Download successfully!
download_chunk_file end
read  cost time 2.95s
write cost time 0.01s
read  cost time 2.94s
write cost time 0.00s
read  cost time 3.11s
write cost time 0.01s
read  cost time 3.06s
write cost time 0.01s
read  cost time 3.10s
write cost time 0.01s
read  cost time 3.07s
write cost time 0.00s
read  cost time 3.07s
write cost time 0.01s
read  cost time 0.01s
write cost time 0.00s
chunk file with start_pos ~ end_pos: 47532270 ~ 55454315, Download successfully!
download_chunk_file end
read  cost time 2.50s
write cost time 0.00s
read  cost time 2.67s
read  cost time 2.61s
write cost time 0.01s
write cost time 0.00s
read  cost time 0.04s
write cost time 0.00s
chunk file with start_pos ~ end_pos: 39610225 ~ 47532270, Download successfully!
download_chunk_file end
read  cost time 2.58s
write cost time 0.00s
read  cost time 2.64s
write cost time 0.00s
read  cost time 2.57s
write cost time 0.00s
read  cost time 2.30s
write cost time 0.01s
read  cost time 2.39s
write cost time 0.00s
read  cost time 0.00s
write cost time 0.00s
chunk file with start_pos ~ end_pos: 31688180 ~ 39610225, Download successfully!
download_chunk_file end
read  cost time 2.28s
write cost time 0.00s
read  cost time 2.29s
write cost time 0.00s
read  cost time 2.37s
write cost time 0.00s
read  cost time 2.20s
write cost time 0.00s
read  cost time 0.00s
write cost time 0.00s
chunk file with start_pos ~ end_pos: 23766135 ~ 31688180, Download successfully!
download_chunk_file end
read  cost time 2.18s
write cost time 0.00s
read  cost time 2.18s
write cost time 0.00s
read  cost time 2.30s
write cost time 0.00s
read  cost time 2.18s
write cost time 0.02s
read  cost time 2.13s
write cost time 0.00s
read  cost time 0.04s
write cost time 0.00s
chunk file with start_pos ~ end_pos: 15844090 ~ 23766135, Download successfully!
download_chunk_file end
read  cost time 2.24s
write cost time 0.00s
read  cost time 2.05s
write cost time 0.00s
read  cost time 2.06s
read  cost time 2.07s
write cost time 6.32s
read  cost time 0.00s
write cost time 0.00s
write cost time 4.46s
chunk file with start_pos ~ end_pos: 7922045 ~ 15844090, Download successfully!
read  cost time 0.00s
write cost time 0.00s
download_chunk_file end
chunk file with start_pos ~ end_pos: 0 ~ 7922045, Download successfully!
download_chunk_file end








Full file Download successfully! Cost: 25.37s