#! /usr/bin/env python
# VERSION 1.0
# copyleft Harvey Greenberg hgreen@u.washington.edu
# This assumes that you have a series of files (.e.g. NAIP image files)
# which are broken into gigabyte chunks named fname.a, fname.b, etc.
# to be concatenated into fname.  The .a file is renamed, while the
# remaining files are appended to it.
# RHEL tests show a megabyte buffer slightly better than a kilobyte
#
# e.g. cat.py
# If python is not in your searchpath, you should add it:
#   That can be done in XP with Settings/ControlPanel/System/Advanced/EnvironmentalVariables.
# or call python explicitly, .e.g. c:\Python24\python cat.py
# ESRI users have python installed, generally in c:\Python24
# (or c:\Python21 for older versions).
# 
# There are no arguments.  This will append every set in this directory.

import string
import sys
import os

roots=[]
nroots=0
fl = os.listdir('.')
for fil in fl:
  if fil[-2:] != '.a':
    continue
  roots.append(fil[:-1])
  nroots=nroots+1
print "Files to process:",roots
for root in roots:
  count=0
  for fil in fl:
    if fil[:-1] == root:
      count = count+1
  print "There are",count,"pieces to",root
  size=os.stat(root+string.ascii_lowercase[count-1])[6]
# print "Last file is",root+string.ascii_lowercase[count-1]
  if size == 1073741824L:
    print "The last component "+root+string.ascii_lowercase[count-1]+" is one gigabyte, so there is probably a missing file." 
    print "We are skipping this set."
    continue
  for n in range(count):
    if n < (count-1):	# if not the last component
      size=os.stat(root+string.ascii_lowercase[n])[6]
#     print size,root+string.ascii_lowercase[n]
      if size != 1073741824L:
        print "Size of "+root+string.ascii_lowercase[n]+" is only "+`size`+", not one gigabyte."
        print "We are skipping this set."
        break
    if n == 0:
      os.rename(root+"a",root[:-1])
      g=open(root[:-1],"ab")
    else:
      print "Opening",string.ascii_lowercase[n],size
      f=open(root+string.ascii_lowercase[n],"rb")
      while 1:
        buffer=f.read(1048576) #RHEL tests show a megabyte buffer slightly better than a kilobyte
        g.write(buffer)
        if len(buffer) != 1048576:
          f.close()
          break
      if n == (count-1):	# last component
        g.close()