Fix chromosome bug

scripts/write-snps.py

@@ -9,49 +9,54 @@
 import glob
 from collections import defaultdict
 import numpy
-import argparse
 
 def main(argv):
-
-   parser=argparse.ArgumentParser()
-   parser.add_argument("outfile", help="Enter output file name") 
-   args=parser.parse_args()
-   outfp=args.outfile
+   try:
+      opts, args = getopt.getopt(argv,"hi:o:",["ofile="])
+   except getopt.GetoptError:
+      print 'write-snps.py  -o <outputfile>'
+      sys.exit(2)
+   for opt, arg in opts:
+      if opt == '-h':
+         print 'write-snps.py  -o <outputfile>'
+         sys.exit()
+      elif opt in ("-o", "--ofile"):
+         outfp = arg
    nones=0
    snps=0
    index=1
    vcfs=list()
    filelist = glob.glob('*.vcf')
    dic={}
-   outfp=open(sys.argv[1], 'w') 
+   outfp=open(sys.argv[1], 'w')  
    j=0
    for every in filelist:
        name=every.split('.')
        vcf=name[0]
        vcfs.append(vcf)
    mat=numpy.zeros((2000000,len(vcfs)+2),int)
-   indeces= []
-   indeces.append('zero')
    for r1 in filelist:
       print 'processing', r1, index
+      lines = open(r1, "r" ).readlines()
+      j=j+1 
       i=0
-      j=j+1
-      for line in open(r1):
-        i=i+1 
-        rec=line.split('\t')
-        if line.startswith('#'):  
-            continue 
-
-        else:
-               ivalue=str(rec[0])+'-'+str(rec[1])
-               if(ivalue in indeces):
-                          oldindex=indeces.index(ivalue)
+      lines = tuple(lines)
+      while i < len(lines):
+          rec=lines[i].split('\t')
+          if lines[i][0] in '#':
+                 i+=1
+                 pass
+          else:
+               i+=1
+               if str(rec[1]) in dic:
+                          oldindex=dic.get(rec[1])
                           mat[oldindex][j]=1
                else: 
+                          mat[index][0]=str(rec[1])
                           mat[index][j]=1
-                          indeces.append(ivalue)
+                          dic[str(rec[1])]=index
                           index+=1
-   print len(indeces)
+   snps=index
    deli=" "
    zero="0"
    one="1"
@@ -62,11 +67,11 @@ def main(argv):
              outfp.write(str(deli).rstrip('\n'))
    print >> outfp ,'\n'
    itr1=1
-   while itr1 < len(indeces) :
+   while itr1 <= snps:
           itr2=1 
-          outfp.write(str(indeces[itr1]).rstrip('\n')) 
+          outfp.write(str(mat[itr1][0]).rstrip('\n')) 
           outfp.write(str(deli).rstrip('\n'))
-          while itr2 <= j: 
+          while itr2 <= 11: 
                   outfp.write(str(mat[itr1][itr2]).rstrip('\n'))
                   outfp.write(str(deli).rstrip('\n'))
                   itr2+=1