###### 不要用代理,代理来了不能用,草他妈的!
import pandas as pd
import xarray as xr
from multiprocessing import Pool
import os
from pyesgf.logon import LogonManager
lm = LogonManager()
lm.is_logged_on()
def mkdir(path):
def CMIP6_download(file, IPCC_folder, try_download_num):
def CMIP6_creatMutil_download(file_list, IPCC_folder, pool_pre_num, try_download_num):
if __name__ == '__main__':
folder = os.path.join("K:\IPCC6\\nc2")
for csv in [x for x in os.listdir(folder) if x.endswith(".csv")][-16:]:
print(csv)
path = os.path.join(folder, csv)
df = pd.read_csv(path)
# print(df.columns)
file_list = df["url"].tolist()
file_list2 = []
for i in file_list:
# print(i)
file_name = str(i).split("/")[-1]
print(file_name)
if os.path.exists(os.path.join(folder, csv.split("_")[0], file_name)):
pass
else:
file_list2.append(i)
IPCC_folder = os.path.join("K:\\IPCC6\\nc2\\", csv.split("_")[0])
#note 开启多进程进行下载
file_list = file_list2
IPCC_folder = IPCC_folder
try_download_num = 1
pool_pre_num = 40
CMIP6_creatMutil_download(file_list, IPCC_folder, pool_pre_num, try_download_num)
中间有一些机构的下不到,这都是正常现象
比如:
下载下来基本上就这样:
私信获取代码或数据
一直以这个速度在狂飙