| 1 |
import os, sys, re
|
| 2 |
|
| 3 |
###############################################################
|
| 4 |
|
| 5 |
#Usage:
|
| 6 |
# python onlyLast.py <path to input directory on CASTOR>
|
| 7 |
|
| 8 |
#now updated to handle output from the latest CRAB with
|
| 9 |
#file names like this:
|
| 10 |
# myfilename_1_1_aBC.root
|
| 11 |
|
| 12 |
###############################################################
|
| 13 |
|
| 14 |
#handle input arguments
|
| 15 |
#this includes the name of the script
|
| 16 |
##for arg in sys.argv:
|
| 17 |
## print arg
|
| 18 |
|
| 19 |
#sys.argv[1] will have the first argument
|
| 20 |
tmpfile = '/tmp/joshmt/onlyLast_py_'
|
| 21 |
mypid = os.getpid()
|
| 22 |
tmpfile += `mypid`
|
| 23 |
|
| 24 |
inputdir = sys.argv[1]
|
| 25 |
if inputdir[len(inputdir)-1] != '/':
|
| 26 |
inputdir+='/'
|
| 27 |
|
| 28 |
outputdir = inputdir
|
| 29 |
outputdir += 'EXTRAS'
|
| 30 |
|
| 31 |
mkdircommand = "nsmkdir "
|
| 32 |
mkdircommand += outputdir
|
| 33 |
|
| 34 |
#get a file listing into tmpfile
|
| 35 |
cmd = 'nsls -l '
|
| 36 |
cmd += inputdir
|
| 37 |
cmd += ' | awk \'// {print $5,$9;}\' > '
|
| 38 |
cmd += tmpfile
|
| 39 |
#print cmd
|
| 40 |
os.system(cmd)
|
| 41 |
|
| 42 |
#keep track of whether we made the directory or not
|
| 43 |
alreadymadedir = 0
|
| 44 |
|
| 45 |
f = open(tmpfile,'r')
|
| 46 |
|
| 47 |
#to deal with the arbitrary extension to the filename that crab
|
| 48 |
#now adds, I will add a second associative array to keep track of the
|
| 49 |
#whole file name. (completefilenames)
|
| 50 |
#The only subtle part is where the file names are sorted.
|
| 51 |
#Hopefully this does not create any bugs
|
| 52 |
indexdict = {}
|
| 53 |
completefilenames = {}
|
| 54 |
|
| 55 |
stub = 'stub'
|
| 56 |
|
| 57 |
#do some accounting
|
| 58 |
nmoved = 0
|
| 59 |
|
| 60 |
for line in f:
|
| 61 |
mypair = line.split()
|
| 62 |
#file size
|
| 63 |
size = int(mypair[0])
|
| 64 |
#parse filename
|
| 65 |
result = re.match('(.*)_([0-9]*)_([0-9]*)_(.*).root',mypair[1])
|
| 66 |
if size > 0:
|
| 67 |
if stub=='stub':
|
| 68 |
stub = result.group(1)
|
| 69 |
elif stub!=result.group(1):
|
| 70 |
print "WARNING -- there is more than one group of filenames here!"
|
| 71 |
|
| 72 |
if result.group(2) in indexdict:
|
| 73 |
#so the key already exists
|
| 74 |
indexdict[result.group(2)].append(result.group(3))
|
| 75 |
completefilenames[result.group(2)].append(mypair[1])
|
| 76 |
else:
|
| 77 |
indexdict[result.group(2)] = [result.group(3)]
|
| 78 |
completefilenames[result.group(2)] = [mypair[1]]
|
| 79 |
# print result.group(0)
|
| 80 |
# print result.group(1)
|
| 81 |
# print result.group(2)
|
| 82 |
# print result.group(3)
|
| 83 |
|
| 84 |
#print indexdict
|
| 85 |
|
| 86 |
for ii in indexdict:
|
| 87 |
if len(indexdict[ii])==1:
|
| 88 |
print "nothing to do for ",ii
|
| 89 |
else:
|
| 90 |
indexdict[ii].sort()
|
| 91 |
#will this work? in my one test case, yes
|
| 92 |
completefilenames[ii].sort()
|
| 93 |
goodindex = indexdict[ii].pop()
|
| 94 |
goodfilename = completefilenames[ii].pop()
|
| 95 |
print ii, ": keeping index ",goodindex
|
| 96 |
print ii, ": corresponds to file ",goodfilename
|
| 97 |
for jj in completefilenames[ii]:
|
| 98 |
if alreadymadedir == 0:
|
| 99 |
print mkdircommand
|
| 100 |
#this will give a harmless error if the dir already exists
|
| 101 |
os.system(mkdircommand)
|
| 102 |
|
| 103 |
s=inputdir
|
| 104 |
s+=jj
|
| 105 |
print "moving ", s
|
| 106 |
cpcmd = 'rfcp '
|
| 107 |
cpcmd += s
|
| 108 |
cpcmd += ' '
|
| 109 |
cpcmd += outputdir
|
| 110 |
print cpcmd
|
| 111 |
os.system(cpcmd)
|
| 112 |
rmcmd = 'rfrm '
|
| 113 |
rmcmd += s
|
| 114 |
print rmcmd
|
| 115 |
os.system(rmcmd)
|
| 116 |
nmoved = nmoved+1
|
| 117 |
|
| 118 |
f.close()
|
| 119 |
os.remove(tmpfile)
|
| 120 |
|
| 121 |
print "----------------------------"
|
| 122 |
print "I moved this many files: ",nmoved
|
| 123 |
print "----------------------------"
|
