Mercurial > repos > devteam > intersect

comparison gops_intersect.py @ 3:5f72be09cfd3 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-devteam/tree/master/tool_collections/gops/intersect commit a1517c9d22029095120643bbe2c8fa53754dd2b7
author devteam
date Wed, 11 Nov 2015 12:48:44 -0500
parents 5b3c6135a982
children 33b3f3688db4
comparison
equal deleted inserted replaced
2:77641d5731c8 3:5f72be09cfd3
9 -m, --mincols=N: Require this much overlap (default 1bp) 9 -m, --mincols=N: Require this much overlap (default 1bp)
10 -p, --pieces: just print pieces of second set (after padding) 10 -p, --pieces: just print pieces of second set (after padding)
11 -G, --gff1: input 1 is GFF format, meaning start and end coordinates are 1-based, closed interval 11 -G, --gff1: input 1 is GFF format, meaning start and end coordinates are 1-based, closed interval
12 -H, --gff2: input 2 is GFF format, meaning start and end coordinates are 1-based, closed interval 12 -H, --gff2: input 2 is GFF format, meaning start and end coordinates are 1-based, closed interval
13 """ 13 """
14 import sys, traceback, fileinput 14 import fileinput
15 from warnings import warn 15 import sys
16 from bx.intervals import * 16 from bx.intervals.io import GenomicInterval, NiceReaderWrapper
17 from bx.intervals.io import * 17 from bx.intervals.operations.intersect import intersect
18 from bx.intervals.operations.intersect import *
19 from bx.cookbook import doc_optparse 18 from bx.cookbook import doc_optparse
20 from galaxy.tools.util.galaxyops import * 19 from bx.tabular.io import ParseError
20 from galaxy.tools.util.galaxyops import fail, parse_cols_arg, skipped
21 from utils.gff_util import GFFFeature, GFFReaderWrapper, convert_bed_coords_to_gff 21 from utils.gff_util import GFFFeature, GFFReaderWrapper, convert_bed_coords_to_gff
22 22
23 assert sys.version_info[:2] >= ( 2, 4 ) 23 assert sys.version_info[:2] >= ( 2, 4 )
24 24
25
25 def main(): 26 def main():
26 mincols = 1 27 mincols = 1
27 upstream_pad = 0
28 downstream_pad = 0
29 28
30 options, args = doc_optparse.parse( __doc__ ) 29 options, args = doc_optparse.parse( __doc__ )
31 try: 30 try:
32 chr_col_1, start_col_1, end_col_1, strand_col_1 = parse_cols_arg( options.cols1 ) 31 chr_col_1, start_col_1, end_col_1, strand_col_1 = parse_cols_arg( options.cols1 )
33 chr_col_2, start_col_2, end_col_2, strand_col_2 = parse_cols_arg( options.cols2 ) 32 chr_col_2, start_col_2, end_col_2, strand_col_2 = parse_cols_arg( options.cols2 )
34 if options.mincols: mincols = int( options.mincols ) 33 if options.mincols:
34 mincols = int( options.mincols )
35 pieces = bool( options.pieces ) 35 pieces = bool( options.pieces )
36 in1_gff_format = bool( options.gff1 ) 36 in1_gff_format = bool( options.gff1 )
37 in2_gff_format = bool( options.gff2 ) 37 in2_gff_format = bool( options.gff2 )
38 in_fname, in2_fname, out_fname = args 38 in_fname, in2_fname, out_fname = args
39 except: 39 except:
40 doc_optparse.exception() 40 doc_optparse.exception()
41 41
42 # Set readers to handle either GFF or default format. 42 # Set readers to handle either GFF or default format.
43 if in1_gff_format: 43 if in1_gff_format:
44 in1_reader_wrapper = GFFReaderWrapper 44 in1_reader_wrapper = GFFReaderWrapper
45 else: 45 else:
46 in1_reader_wrapper = NiceReaderWrapper 46 in1_reader_wrapper = NiceReaderWrapper
47 if in2_gff_format: 47 if in2_gff_format:
48 in2_reader_wrapper = GFFReaderWrapper 48 in2_reader_wrapper = GFFReaderWrapper
49 else: 49 else:
50 in2_reader_wrapper = NiceReaderWrapper 50 in2_reader_wrapper = NiceReaderWrapper
51 51
52 g1 = in1_reader_wrapper( fileinput.FileInput( in_fname ), 52 g1 = in1_reader_wrapper( fileinput.FileInput( in_fname ),
53 chrom_col=chr_col_1, 53 chrom_col=chr_col_1,
54 start_col=start_col_1, 54 start_col=start_col_1,
55 end_col=end_col_1, 55 end_col=end_col_1,
56 strand_col=strand_col_1, 56 strand_col=strand_col_1,
57 fix_strand=True ) 57 fix_strand=True )
58 if in1_gff_format: 58 if in1_gff_format:
59 # Intersect requires coordinates in BED format. 59 # Intersect requires coordinates in BED format.
60 g1.convert_to_bed_coord=True 60 g1.convert_to_bed_coord = True
61 g2 = in2_reader_wrapper( fileinput.FileInput( in2_fname ), 61 g2 = in2_reader_wrapper( fileinput.FileInput( in2_fname ),
62 chrom_col=chr_col_2, 62 chrom_col=chr_col_2,
63 start_col=start_col_2, 63 start_col=start_col_2,
64 end_col=end_col_2, 64 end_col=end_col_2,
65 strand_col=strand_col_2, 65 strand_col=strand_col_2,
66 fix_strand=True ) 66 fix_strand=True )
67 if in2_gff_format: 67 if in2_gff_format:
68 # Intersect requires coordinates in BED format. 68 # Intersect requires coordinates in BED format.
69 g2.convert_to_bed_coord=True 69 g2.convert_to_bed_coord = True
70 70
71 out_file = open( out_fname, "w" ) 71 out_file = open( out_fname, "w" )
72 try: 72 try:
73 for feature in intersect( [g1,g2], pieces=pieces, mincols=mincols ): 73 for feature in intersect( [g1, g2], pieces=pieces, mincols=mincols ):
74 if isinstance( feature, GFFFeature ): 74 if isinstance( feature, GFFFeature ):
75 # Convert back to GFF coordinates since reader converted automatically. 75 # Convert back to GFF coordinates since reader converted automatically.
76 convert_bed_coords_to_gff( feature ) 76 convert_bed_coords_to_gff( feature )
77 for interval in feature.intervals: 77 for interval in feature.intervals:
78 out_file.write( "%s\n" % "\t".join( interval.fields ) ) 78 out_file.write( "%s\n" % "\t".join( interval.fields ) )