comparison scatterplot.py @ 0:c12b0759203b draft

Imported from capsule None
author devteam
date Mon, 28 Jul 2014 11:56:22 -0400
parents
children 4f8b9e70fda0
comparison
equal deleted inserted replaced
-1:000000000000 0:c12b0759203b
1 #!/usr/bin/env python
2 #Greg Von Kuster
3
4 import sys
5 from rpy import *
6
7 def stop_err(msg):
8 sys.stderr.write(msg)
9 sys.exit()
10
11 def main():
12
13 in_fname = sys.argv[1]
14 out_fname = sys.argv[2]
15 try:
16 columns = int( sys.argv[3] ) - 1, int( sys.argv[4] ) - 1
17 except:
18 stop_err( "Columns not specified, your query does not contain a column of numerical data." )
19 title = sys.argv[5]
20 xlab = sys.argv[6]
21 ylab = sys.argv[7]
22
23 matrix = []
24 skipped_lines = 0
25 first_invalid_line = 0
26 invalid_value = ''
27 invalid_column = 0
28 i = 0
29 for i, line in enumerate( file( in_fname ) ):
30 valid = True
31 line = line.rstrip( '\r\n' )
32 if line and not line.startswith( '#' ):
33 row = []
34 fields = line.split( "\t" )
35 for column in columns:
36 try:
37 val = fields[column]
38 if val.lower() == "na":
39 row.append( float( "nan" ) )
40 else:
41 row.append( float( fields[column] ) )
42 except:
43 valid = False
44 skipped_lines += 1
45 if not first_invalid_line:
46 first_invalid_line = i + 1
47 try:
48 invalid_value = fields[column]
49 except:
50 invalid_value = ''
51 invalid_column = column + 1
52 break
53 else:
54 valid = False
55 skipped_lines += 1
56 if not first_invalid_line:
57 first_invalid_line = i+1
58
59 if valid:
60 matrix.append( row )
61
62 if skipped_lines < i:
63 try:
64 r.pdf( out_fname, 8, 8 )
65 r.plot( array( matrix ), type="p", main=title, xlab=xlab, ylab=ylab, col="blue", pch=19 )
66 r.dev_off()
67 except Exception, exc:
68 stop_err( "%s" %str( exc ) )
69 else:
70 stop_err( "All values in both columns %s and %s are non-numeric or empty." % ( sys.argv[3], sys.argv[4] ) )
71
72 print "Scatter plot on columns %s, %s. " % ( sys.argv[3], sys.argv[4] )
73 if skipped_lines > 0:
74 print "Skipped %d lines starting with line #%d, value '%s' in column %d is not numeric." % ( skipped_lines, first_invalid_line, invalid_value, invalid_column )
75
76 r.quit( save="no" )
77
78 if __name__ == "__main__":
79 main()