view SumTest.py @ 5:635c4cfd83ee draft default tip

Uploaded
author elindfor
date Wed, 15 Jan 2014 04:33:21 -0500
parents d0fbc0d93d05
children
line wrap: on
line source

#!/usr/bin/env python
#Erno Lindfors

import sys
from rpy import *

def stop_err(msg):
    sys.stderr.write(msg)
    sys.exit()

def main():

    in_fname = sys.argv[1]
    out_fname = sys.argv[2]
    try:
        #columns = int( sys.argv[3] ) - 1, int( sys.argv[4] ) - 1
        columns = int( sys.argv[3] ) - 1
    except:
        stop_err( "Columns not specified, your query does not contain a column of numerical data." )
    
    matrix = []
    skipped_lines = 0
    first_invalid_line = 0
    invalid_value = ''
    invalid_column = 0
    i = 0
    for i, line in enumerate( file( in_fname ) ):
        valid = True
        line = line.rstrip( '\r\n' )
        if line and not line.startswith( '#' ): 
            row = []
            fields = line.split( "\t" )
            for column in columns:
                try:
                    val = fields[column]
                    if val.lower() == "na": 
                        row.append( float( "nan" ) )
                    else:
                        row.append( float( fields[column] ) )
                except:
                    valid = False
                    skipped_lines += 1
                    if not first_invalid_line:
                        first_invalid_line = i + 1
                        try:
                            invalid_value = fields[column]
                        except:
                            invalid_value = ''
                        invalid_column = column + 1
                    break
        else:
            valid = False
            skipped_lines += 1
            if not first_invalid_line:
                first_invalid_line = i+1

        if valid:
            matrix.append( row )

    if skipped_lines < i:
        try:
            #r.pdf( out_fname, 8, 8 )
            #r.plot( array( matrix ), type="p", main=title, xlab=xlab, ylab=ylab, col="blue", pch=19 )
            #r.dev_off()
            r.sum(inputNumbers)
            r.write.table(
              sum(matrix[,1],
              out_fname,
              row.names = FALSE,
              col.names = c("Sum"),
              quote = FALSE
            )
        except Exception, exc:
            stop_err( "%s" %str( exc ) )
    else:
        stop_err( "All values in column %s are non-numeric or empty." % (sys.argv[3]) )

    print "Scatter plot on column %s. " % (sys.argv[3])
    if skipped_lines > 0:
        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 )

    r.quit( save="no" )

if __name__ == "__main__":
    main()