Mercurial > repos > matces > carpet_toolsuite

annotate carpet-src-1/tools/CARPET/TSS_distance.py @ 0:cdd489d98766

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
author matces
date Tue, 07 Jun 2011 16:50:41 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
1 #!/usr/bin/env python
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
2
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
3 # Copyright 2009 Matteo Cesaroni, Lucilla Luzi
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
4 #
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
5 # This program is free software; ; you can redistribute it and/or modify
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
6 # it under the terms of the GNU Lesser General Public License as published by
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
7 # the Free Software Foundation; either version 3 of the License, or (at your
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
8 # option) any later version.
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
9 #
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
10 # This program is distributed in the hope that it will be useful,
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
13 # GNU General Public License for more details.
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
14
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
15
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
16 import sys
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
17 from rpy import *
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
18
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
19
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
20 def stop_err(msg):
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
21 sys.stderr.write(msg)
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
22 sys.exit()
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
23
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
24 def main():
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
25
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
26 # Handle input params
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
27 in_fname = sys.argv[1]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
28 out_fname = sys.argv[2]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
29 try:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
30 column = int( sys.argv[3] ) - 1
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
31 except:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
32 stop_err( "..Column not specified, your query does not contain a column of numerical data." )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
33 title = sys.argv[4]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
34 xlab = sys.argv[5]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
35 breaks = int( sys.argv[6] )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
36 if breaks == 0: breaks = "Sturges"
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
37 if sys.argv[7] == "true": density = True
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
38 else: density = False
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
39
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
40
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
41
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
42 matrix = []
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
43 skipped_lines = 0
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
44 first_invalid_line = 0
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
45 invalid_value = ''
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
46
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
47 for i, line in enumerate( file( in_fname ) ):
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
48 valid = True
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
49 line = line.rstrip('\r\n')
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
50 # Skip comments
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
51 if line and not line.startswith( '#' ):
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
52 # Extract values and convert to floats
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
53 row = []
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
54 try:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
55 fields = line.split( "\t" )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
56 val = fields[column]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
57 if val.lower() == "na":
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
58 row.append( float( "nan" ) )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
59 if float(val) > float(xlab):
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
60 val = (float(xlab)+2000)
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
61
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
62 row.append( float( val ) )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
63 except:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
64 valid = False
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
65 skipped_lines += 1
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
66 if not first_invalid_line:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
67 first_invalid_line = i+1
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
68 else:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
69 try:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
70 row.append( float( val ) )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
71 except ValueError:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
72 valid = False
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
73 skipped_lines += 1
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
74 if not first_invalid_line:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
75 first_invalid_line = i+1
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
76 invalid_value = fields[column]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
77 else:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
78 valid = False
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
79 skipped_lines += 1
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
80 if not first_invalid_line:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
81 first_invalid_line = i+1
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
82
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
83 if valid:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
84 matrix.append( row )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
85
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
86 if skipped_lines < i:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
87 print "..on columnn %s" %sys.argv[3]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
88 try:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
89 a = array( matrix )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
90 r.pdf( out_fname, 8, 8 )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
91 r.hist( a, probability=True, main=title, xlab="TSS distance", breaks=breaks )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
92 if density:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
93 r.lines( r.density( a ) )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
94 r.dev_off()
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
95 except exc:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
96 stop_err("Building histogram resulted in error: %s." %str( exc ))
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
97 else:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
98 print "..all values in column %s are non-numeric." %sys.argv[3]
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
99
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
100 if skipped_lines > 0:
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
101 print "..skipped %d invalid lines starting with line #%d. Value '%s' is not numeric." % ( skipped_lines, first_invalid_line, invalid_value )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
102
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
103 r.quit( save="no" )
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
104
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
105 if __name__ == "__main__":
cdd489d98766 Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
matces
parents:
diff changeset
106 main()