Mercurial > repos > iuc > ngsutils_bam_filter

comparison filter.py @ 3:9b9ae5963d3c draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ngsutils commit 4a3c9f195ba5d899b1a1ce5e80281cdf230f456a
author iuc
date Mon, 23 Oct 2017 13:27:02 -0400
parents 7a68005de299
children 49bf71fbc574
comparison
equal deleted inserted replaced
2:7a68005de299 3:9b9ae5963d3c
105 MAPQ Mapping quality (defined as part of SAM spec) 105 MAPQ Mapping quality (defined as part of SAM spec)
106 106
107 The tag type (:i, :f, :Z) is optional. 107 The tag type (:i, :f, :Z) is optional.
108 108
109 """ 109 """
110 from __future__ import print_function
110 111
111 import os 112 import os
112 import sys 113 import sys
113 114
114 import pysam 115 import pysam
116 from ngsutils.bed import BedFile 117 from ngsutils.bed import BedFile
117 from ngsutils.support.dbsnp import DBSNP 118 from ngsutils.support.dbsnp import DBSNP
118 119
119 120
120 def usage(): 121 def usage():
121 print __doc__ 122 print(__doc__)
122 print """ 123 print("""
123 Usage: bamutils filter in.bam out.bam {-failed out.txt} criteria... 124 Usage: bamutils filter in.bam out.bam {-failed out.txt} criteria...
124 125
125 Options: 126 Options:
126 -failed fname A text file containing the read names of all reads 127 -failed fname A text file containing the read names of all reads
127 that were removed with filtering 128 that were removed with filtering
129 Example: 130 Example:
130 bamutils filter filename.bam output.bam -mapped -gte AS:i 1000 131 bamutils filter filename.bam output.bam -mapped -gte AS:i 1000
131 132
132 This will remove all unmapped reads, as well as any reads that have an AS:i 133 This will remove all unmapped reads, as well as any reads that have an AS:i
133 value less than 1000. 134 value less than 1000.
134 """ 135 """)
135 sys.exit(1) 136 sys.exit(1)
136 137
137 138
138 class Unique(object): 139 class Unique(object):
139 def __init__(self, length=None): 140 def __init__(self, length=None):
758 self.tag = tag 759 self.tag = tag
759 760
760 # guess at type... 761 # guess at type...
761 try: 762 try:
762 self.value = int(value) 763 self.value = int(value)
763 except: 764 except ValueError:
764 try: 765 try:
765 self.value = float(value) 766 self.value = float(value)
766 except: 767 except ValueError:
767 self.value = value 768 self.value = value
768 769
769 def get_value(self, read): 770 def get_value(self, read):
770 if self.tag == 'MAPQ': 771 if self.tag == 'MAPQ':
771 return read.mapq 772 return read.mapq
958 infile = arg 959 infile = arg
959 elif not outfile: 960 elif not outfile:
960 outfile = arg 961 outfile = arg
961 elif arg[0] == '-': 962 elif arg[0] == '-':
962 if not arg[1:] in _criteria: 963 if not arg[1:] in _criteria:
963 print "Unknown criterion: %s" % arg 964 print("Unknown criterion: %s" % arg)
964 fail = True 965 fail = True
965 if crit_args: 966 if crit_args:
966 criteria.append(_criteria[crit_args[0][1:]](*crit_args[1:])) 967 criteria.append(_criteria[crit_args[0][1:]](*crit_args[1:]))
967 crit_args = [arg, ] 968 crit_args = [arg, ]
968 elif crit_args: 969 elif crit_args:
969 crit_args.append(arg) 970 crit_args.append(arg)
970 else: 971 else:
971 print "Unknown argument: %s" % arg 972 print("Unknown argument: %s" % arg)
972 fail = True 973 fail = True
973 974
974 if not fail and crit_args: 975 if not fail and crit_args:
975 criteria.append(_criteria[crit_args[0][1:]](*crit_args[1:])) 976 criteria.append(_criteria[crit_args[0][1:]](*crit_args[1:]))
976 977
977 if fail or not infile or not outfile or not criteria: 978 if fail or not infile or not outfile or not criteria:
978 if not infile and not outfile and not criteria: 979 if not infile and not outfile and not criteria:
979 usage() 980 usage()
980 981
981 if not infile: 982 if not infile:
982 print "Missing: input bamfile" 983 print("Missing: input bamfile")
983 if not outfile: 984 if not outfile:
984 print "Missing: output bamfile" 985 print("Missing: output bamfile")
985 if not criteria: 986 if not criteria:
986 print "Missing: filtering criteria" 987 print("Missing: filtering criteria")
987 usage() 988 usage()
988 else: 989 else:
989 bam_filter(infile, outfile, criteria, failed, verbose) 990 bam_filter(infile, outfile, criteria, failed, verbose)