python开发_thread_线程_搜索本地文件

在之前的blog中，曾经写到过关于搜索本地文件的技术文章

如：

java开发_快速搜索本地文件_小应用程序

python开发_搜索本地文件信息写入文件

下面说说python中关于线程来搜索本地文件

利用多个线程处理搜索的问题，我们可以发现他很快....

========================================================

下面是代码部分：

========================================================

  1 # A parallelized "find(1)" using the thread module.
  2 
  3 # This demonstrates the use of a work queue and worker threads.
  4 # It really does do more stats/sec when using multiple threads,
  5 # although the improvement is only about 20-30 percent.
  6 # (That was 8 years ago.  In 2002, on Linux, I can't measure
  7 # a speedup. :-( )
  8 
  9 # I'm too lazy to write a command line parser for the full find(1)
 10 # command line syntax, so the predicate it searches for is wired-in,
 11 # see function selector() below.  (It currently searches for files with
 12 # world write permission.)
 13 
 14 # Usage: parfind.py [-w nworkers] [directory] ...
 15 # Default nworkers is 4
 16 
 17 
 18 import sys
 19 import getopt
 20 import time
 21 import os
 22 from stat import *
 23 import _thread as thread
 24 
 25 
 26 # Work queue class.  Usage:
 27 #   wq = WorkQ()
 28 #   wq.addwork(func, (arg1, arg2, ...)) # one or more calls
 29 #   wq.run(nworkers)
 30 # The work is done when wq.run() completes.
 31 # The function calls executed by the workers may add more work.
 32 # Don't use keyboard interrupts!
 33 
 34 class WorkQ:
 35 
 36     # Invariants:
 37 
 38     # - busy and work are only modified when mutex is locked
 39     # - len(work) is the number of jobs ready to be taken
 40     # - busy is the number of jobs being done
 41     # - todo is locked iff there is no work and somebody is busy
 42 
 43     def __init__(self):
 44         self.mutex = thread.allocate()
 45         self.todo = thread.allocate()
 46         self.todo.acquire()
 47         self.work = []
 48         self.busy = 0
 49 
 50     def addwork(self, func, args):
 51         job = (func, args)
 52         self.mutex.acquire()
 53         self.work.append(job)
 54         self.mutex.release()
 55         if len(self.work) == 1:
 56             self.todo.release()
 57 
 58     def _getwork(self):
 59         self.todo.acquire()
 60         self.mutex.acquire()
 61         if self.busy == 0 and len(self.work) == 0:
 62             self.mutex.release()
 63             self.todo.release()
 64             return None
 65         job = self.work[0]
 66         del self.work[0]
 67         self.busy = self.busy + 1
 68         self.mutex.release()
 69         if len(self.work) > 0:
 70             self.todo.release()
 71         return job
 72 
 73     def _donework(self):
 74         self.mutex.acquire()
 75         self.busy = self.busy - 1
 76         if self.busy == 0 and len(self.work) == 0:
 77             self.todo.release()
 78         self.mutex.release()
 79 
 80     def _worker(self):
 81         time.sleep(0.00001)     # Let other threads run
 82         while 1:
 83             job = self._getwork()
 84             if not job:
 85                 break
 86             func, args = job
 87             func(*args)
 88             self._donework()
 89 
 90     def run(self, nworkers):
 91         if not self.work:
 92             return # Nothing to do
 93         for i in range(nworkers-1):
 94             thread.start_new(self._worker, ())
 95         self._worker()
 96         self.todo.acquire()
 97 
 98 
 99 # Main program
100 
101 def main():
102     nworkers = 4
103     #print(getopt.getopt(sys.argv[1:], '-w:'))
104     opts, args = getopt.getopt(sys.argv[1:], '-w:')
105     for opt, arg in opts:
106         if opt == '-w':
107             nworkers = int(arg)
108     if not args:
109         #print(os.curdir)
110         args = [os.curdir]
111 
112     wq = WorkQ()
113     for dir in args:
114         wq.addwork(find, (dir, selector, wq))
115 
116     t1 = time.time()
117     wq.run(nworkers)
118     t2 = time.time()
119 
120     sys.stderr.write('Total time %r sec.
' % (t2-t1))
121 
122 
123 # The predicate -- defines what files we look for.
124 # Feel free to change this to suit your purpose
125 
126 def selector(dir, name, fullname, stat):
127     # Look for world writable files that are not symlinks
128     return (stat[ST_MODE] & 0o002) != 0 and not S_ISLNK(stat[ST_MODE])
129 
130 
131 # The find procedure -- calls wq.addwork() for subdirectories
132 
133 def find(dir, pred, wq):
134     try:
135         names = os.listdir(dir)
136     except os.error as msg:
137         print(repr(dir), ':', msg)
138         return
139     for name in names:
140         if name not in (os.curdir, os.pardir):
141             fullname = os.path.join(dir, name)
142             try:
143                 stat = os.lstat(fullname)
144             except os.error as msg:
145                 print(repr(fullname), ':', msg)
146                 continue
147             if pred(dir, name, fullname, stat):
148                 print(fullname)
149             if S_ISDIR(stat[ST_MODE]):
150                 if not os.path.ismount(fullname):
151                     wq.addwork(find, (fullname, pred, wq))
152 
153 
154 # Call the main program
155 
156 main()

========================================================