|
0
|
1 #!/usr/bin/env python
|
|
|
2 #By, Guruprasad Ananda.
|
|
|
3
|
|
|
4 import sys, re
|
|
|
5
|
|
|
6 def stop_err(msg):
|
|
|
7 sys.stderr.write(msg)
|
|
|
8 sys.exit()
|
|
|
9
|
|
|
10 def main():
|
|
|
11 if len(sys.argv) != 5:
|
|
|
12 stop_err("usage: convert_characters infile from_char to_char outfile")
|
|
|
13
|
|
|
14 try:
|
|
|
15 fin = open(sys.argv[1],'r')
|
|
|
16 except:
|
|
|
17 stop_err("Input file cannot be opened for reading.")
|
|
|
18
|
|
|
19 from_char = sys.argv[2]
|
|
|
20 to_char = sys.argv[3]
|
|
|
21
|
|
|
22 try:
|
|
|
23 fout = open(sys.argv[4],'w')
|
|
|
24 except:
|
|
|
25 stop_err("Output file cannot be opened for writing.")
|
|
|
26
|
|
|
27 char_dict = {
|
|
|
28 'T': '\t',
|
|
|
29 's': '\s',
|
|
|
30 'Dt': '\.',
|
|
|
31 'C': ',',
|
|
|
32 'D': '-',
|
|
|
33 'U': '_',
|
|
|
34 'P': '\|',
|
|
|
35 'Co': ':',
|
|
|
36 'Sc': ';'
|
|
|
37 }
|
|
|
38 from_ch = char_dict[from_char] + '+' #making an RE to match 1 or more occurences.
|
|
|
39 to_char = char_dict[to_char]
|
|
|
40 skipped = 0
|
|
|
41
|
|
|
42 for line in fin:
|
|
|
43 line = line.strip()
|
|
|
44 try:
|
|
|
45 fout.write("%s\n" %(re.sub(from_ch, to_char, line)))
|
|
|
46 except:
|
|
|
47 skipped += 1
|
|
|
48
|
|
|
49 fout.close()
|
|
|
50 fin.close()
|
|
|
51 if skipped:
|
|
|
52 print "Skipped %d lines as invalid." %skipped
|
|
|
53
|
|
|
54 if __name__ == "__main__":
|
|
|
55 main()
|
