import os, sys
import md5

import win32file

INCLUDE_EXTENSIONS = (".doc", ".xls", ".ppt", ".txt", ".sql", ".py", ".txt")
IGNORE_FILES = ("readme.txt", "index.html")

def hash (filename, n_chunks=0):
  CHUNK_SIZE = 128 * 1024
  hash = md5.new ()
  f = open (filename)
  try:
    n_chunk = 1
    while 1:
      data = f.read (CHUNK_SIZE)
      if data == '': break
      hash.update (data)
      n_chunk += 1
      if n_chunks and n_chunk > n_chunks: break
  finally:
    f.close ()
    return hash.hexdigest ()

def visit (files, dirname, fnames):
  """Keep track of reused filenames and the directory
   they appeared in.
  """
  sys.stderr.write ("%s\r" % (dirname + (80 * " "))[:80])
  for fname in fnames:
    if fname.lower () not in IGNORE_FILES:
      root, ext = os.path.splitext (fname)
      if ext in INCLUDE_EXTENSIONS:
        checksum = hash (os.path.join (dirname, fname))
        files.setdefault (fname, {}).setdefault (checksum, []).append (dirname)

def main (args):
  roots = [os.path.abspath (p.strip ()) for p in args]
  files = {}
  for root in roots:
    os.path.walk (root, visit, files)
  
  print
  print

  duplicates = {}
  for fname in files.keys ():
    for size in files[fname].keys ():
      dirnames = files[fname][size]
      if len (dirnames) > 1:
        master_dirname = dirnames[0]
        master_filepath = os.path.join (master_dirname, fname)
        print "Master is:", master_filepath
        for dirname in dirnames[1:]:
          slave_filepath = os.path.join (dirname, fname)
          win32file.DeleteFile (slave_filepath)
          win32file.CreateHardLink (slave_filepath, master_filepath)
          print "Linked:", slave_filepath
  
if __name__ == '__main__':
  main (sys.argv[1:])