月度归档： 2021 年 12 月

from dateutil import parser

a=parser.parse('20100101123145')

 

>>a=datetime(2010,1,1,12,31,45)

from dateutil import parser 

a=parser.parse('20100101')

如果想获得这个时间的字符串表达，则可用

a.strftime('%Y%m%d')

https://s tackoverflow.com

/questions/6618515/sorting-list-based-on-values-from-another-list

比如有如下两个list：

people =['Jim','Pam','Micheal','Dwight']

ages =[27,25,4,9]

按ages排序

方法一：

[x for _,x in sorted(zip(ages,people))]

方法二：

import numpy as np

people=np.array(people)

ages=np.array(ages)

indx=ages.argsort()

sortedpeople=people[indx]

方法三：

sorted_index=sorted(range(len(age)),key=lambda x:people[x])

[people[i] for i in sorted_index]



顺便鄙视一下百度和墙。

import matplotlib as mpl

mpl.rcParams['font.family']='serif'

mpl.rcParams['font.serif']='Times New Roman'

使用这个函数的时候要保证dataframe的列名有year，month，day，hour，minute，second这样的，缩写成yr，mm，dy这样的都不行。

将一个hypodd格式的目录读入进了dataframe，想弄成年、月、日这样的顺序，涉及到重新排列columns的位置，网上建议这么操作newcata=cata[[‘year’,’month’,’day’]]

In order to generate the earth model les IASP91.HED and IASP91.TBL, first run program REMODL,then program SETBRN. The program REMODL has the earth model hardwired. Note: These binary lesCANNOT be moved between platforms. They are included with SEISAN for each respective distribution.If lost, they must be regenerated on the same platform

之前用到了subprocess模块，但有人提出可能会导致阻塞，最近在网上看到了pexpect模块，操作起来就已经跟perl很像了。

import pexpect

child=pexpect.spawn("sac")
child.sendline("r test.sac")
child.sendline("div 100")
child.sendline("w over")
child.sendline("q")

25个线程要做完1000个任务

import threading,time,random,os,threadpool
from shutil import copyfile

def dealing(cata):
    print threading.current_thread().name
    print 'entering ',cata
    os.popen('sh DORFTN '+cata).read()
#    os.chdir('..')

if __name__ == "__main__":
    cata=[c for c in os.listdir('.') if os.path.isdir(c)]
   pool=threadpool.ThreadPool(25)
   requests=threadpool.makeRequests(dealing,cata)
    [pool.putRequest(req) for req in requests]
    pool.wait()


如果函数内有多个参数，可如下处理：
import threadpool, os, glob

out=[]

def processing(sample,trace):
   tmp=sample.split('/')
   tracefile=trace+'/'+tmp[-1]
   ret=os.popen('./doublet_search %s %s' % (sample,tracefile)).read()
    print(ret)

if __name__=='__main__':
   traces=glob.glob('Trace/2*')
   samples=glob.glob('Template/2*')
    arglist=[]
    for sample in samples:
       Zs=glob.glob(sample+'/*.?HZ')
       for Z in Zs:
          sam=Z[0:-1]
           for trace in traces:
             arglist.append(([sam,trace],None))
#    print(arglist)
   pool_t=threadpool.ThreadPool(8)
   requests=threadpool.makeRequests(processing,arglist)
    [pool_t.putRequest(req) for req in requests]
    pool_t.wait()

****注：一般在函数中不要用到os.chdir，因为chdir是全局的，如果要用，则要用绝对路径。

a[:,2:4]=10

a[a>5]=10

np.where(a>5,a,10) # if a>5, values equal to a themselves, else equal to 10

a.clip(3,10) # if a<3, =3, if a<10, =10