1 | #!/usr/bin/env python |
---|
2 | """ |
---|
3 | Input: fasta, minimal length, maximal length |
---|
4 | Output: fasta |
---|
5 | Return sequences whose lengths are within the range. |
---|
6 | """ |
---|
7 | |
---|
8 | import sys, os |
---|
9 | |
---|
10 | assert sys.version_info[:2] >= ( 2, 4 ) |
---|
11 | |
---|
12 | def stop_err( msg ): |
---|
13 | sys.stderr.write( msg ) |
---|
14 | sys.exit() |
---|
15 | |
---|
16 | def __main__(): |
---|
17 | input_filename = sys.argv[1] |
---|
18 | try: |
---|
19 | min_length = int( sys.argv[2] ) |
---|
20 | except: |
---|
21 | stop_err( "Minimal length of the return sequence requires a numerical value." ) |
---|
22 | try: |
---|
23 | max_length = int( sys.argv[3] ) |
---|
24 | except: |
---|
25 | stop_err( "Maximum length of the return sequence requires a numerical value." ) |
---|
26 | output_filename = sys.argv[4] |
---|
27 | output_handle = open( output_filename, 'w' ) |
---|
28 | tmp_size = 0 #-1 |
---|
29 | tmp_buf = '' |
---|
30 | at_least_one = 0 |
---|
31 | for line in file(input_filename): |
---|
32 | if not line or line.startswith('#'): |
---|
33 | continue |
---|
34 | if line[0] == '>': |
---|
35 | if min_length <= tmp_size <= max_length or (min_length <= tmp_size and max_length == 0): |
---|
36 | output_handle.write(tmp_buf) |
---|
37 | at_least_one = 1 |
---|
38 | tmp_buf = line |
---|
39 | tmp_size = 0 |
---|
40 | else: |
---|
41 | if max_length == 0 or tmp_size < max_length: |
---|
42 | tmp_size += len(line.rstrip('\r\n')) |
---|
43 | tmp_buf += line |
---|
44 | # final flush of buffer |
---|
45 | if min_length <= tmp_size <= max_length or (min_length <= tmp_size and max_length == 0): |
---|
46 | output_handle.write(tmp_buf.rstrip('\r\n')) |
---|
47 | at_least_one = 1 |
---|
48 | output_handle.close() |
---|
49 | if at_least_one == 0: |
---|
50 | print "There is no sequence that falls within your range." |
---|
51 | |
---|
52 | if __name__ == "__main__" : __main__() |
---|