view mytools/sampline.py @ 7:f0dc65e7f6c0

Uploaded
author xuebing
date Fri, 09 Mar 2012 19:59:07 -0500
parents
children
line wrap: on
line source

#!/usr/bin/env python

"""
Sampling random records from a file. Each record is defined by a fixed number of lines.

Usage: sampline.py [options] 

Options:
  -h, --help            show this help message and exit
  -r, --replacement     Sampling with replacement
  -i INPUT, --input=INPUT
                        Input file
  -o OUTPUT, --output=OUTPUT
                        Output file
  -k NSAMPLE, --nSample=NSAMPLE
                        (required) number of records to be sampled/output
  -m RECSIZE, --recSize=RECSIZE
                        (default=1) number of lines spanned by each record
  -n NSKIP, --nSkip=NSKIP
                        (default=0) number of comment lines to skip at the
                        beginning

example:
    python sampline.py -i test10000.fastq -o out.txt --nSample=5 --recSize=4 --nSkip=0 --replacement
"""

import optparse, string, random,sys,math,itertools

assert sys.version_info[:2] >= ( 2, 4 )

def main():

    # Parse command line    
    parser = optparse.OptionParser( usage="%prog [options] " )
    parser.add_option( "-r", "--replacement",  action="store_true", dest="replacement",default=False,
                       help="Sampling with replacement" )
    parser.add_option( "-i", "--input",  dest="input",  default=None,
                       help="Input file" )
    parser.add_option( "-o", "--output", dest="output", default=None,
                       help="Output file" )
    parser.add_option("-k","--nSample", type='int',dest="nSample",default=None,
                      help="(required) number of records to be sampled/output" )
    parser.add_option("-m","--recSize", type='int',dest="recSize",default=1,
                      help="(default=1) number of lines spanned by each record" )     
    parser.add_option("-n","--nSkip", type='int',dest="nSkip",default=0,
                      help="(default=0) number of comment lines to skip at the beginning")    
    options, args = parser.parse_args()
    #assert options.region in ( 'coding', 'utr3', 'utr5', 'transcribed' ), "Invalid region argument"
    
    sampline(options.input,options.output,options.nSample,options.recSize,options.nSkip,options.replacement)

def sample_wr(population, k):
    "Chooses k random elements (with replacement) from a population"
    n = len(population)
    _random, _int = random.random, int  # speed hack
    return [_int(_random() * n) for i in itertools.repeat(None, k)]

# num of lines
def readinput(filename):
    try:
        f = open (filename)
    except:
        print >> sys.stderr, "can't open file "+str(filename)
        sys.exit(0)

    nline = 0
    for line in f:
        nline = nline + 1
    f.close()    
    return nline

def sampline(infile,outfile,nSample,recSize,nSkip,replacement):
    # sample nSample records from file 
    # each record contains recSize lines
    # skip the top nSkip lines  
    
    nLine = readinput(infile)
    print 'num of lines in input: ',nLine
    print 'avoid sampling the first ',nSkip,' lines'
    print 'lines per record: ',recSize

    if (nLine-nSkip) % recSize:
        print >> sys.stderr, "the number of lines is not dividable by record size!"
        sys.exit(0)

    nTotalRecords = (nLine-nSkip) / recSize
    print "total number of records: ",nTotalRecords
            
    if replacement or nTotalRecords < nSample:
        sel = sample_wr(range(nTotalRecords),nSample)
    else:
        sel = random.sample(range(nTotalRecords),nSample)
    
    #print len(sel), sorted(sel)

    # output
    try:
        fout = open (outfile,'w')
    except:
        print >> sys.stderr, "can't open file "+str(outfile)
        sys.exit(0)
    fin = open(infile)
    n = 0 # index of line
    rec = "" # to store all content of a record
    nrepeat = 0 # number of times a record is sampled
    curr_rec = -1           
    for line in fin:
        if n < nSkip:
            n = n + 1
            fout.write(line)
            continue
        
        if not (n-nSkip) % recSize:# a new record
            # print the previous sampled record
            for i in range(nrepeat):
                fout.write(rec)
            curr_rec = (n-nSkip)/recSize
            nrepeat = sel.count(curr_rec)
            if nrepeat: # sampled            
                rec = line
                #print curr_rec,nrepeat           
        elif (n-nSkip)/recSize == curr_rec:
            rec = rec + line
        n = n + 1
    # if the last record is selected
    if curr_rec == nTotalRecords-1:
        for i in range(nrepeat):
            fout.write(rec)
    fin.close()
    fout.close()          
        

if __name__ == "__main__": main()