Mercurial > repos > edward-kirton > velvet_toolsuite

comparison velvet/jgi_assembly.py @ 0:4afe13ac23b6 default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Migrated tool version 1.0.0 from old tool shed archive to new tool shed repository
author edward-kirton
date Tue, 07 Jun 2011 17:52:16 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:4afe13ac23b6
1 """
2 Assembly classes
3 """
4
5 import data
6 import logging
7 import re
8 import string
9 from cgi import escape
10 from galaxy.datatypes.metadata import MetadataElement
11 from galaxy.datatypes import metadata
12 import galaxy.model
13 from galaxy import util
14 from sniff import *
15
16 log = logging.getLogger(__name__)
17
18 class Assembly( data.Text ):
19 """Class describing an assembly"""
20
21 """Add metadata elements"""
22 MetadataElement( name="contigs", default=0, desc="Number of contigs", readonly=True, visible=False, optional=True, no_value=0 )
23 MetadataElement( name="reads", default=0, desc="Number of reads", readonly=True, visible=False, optional=True, no_value=0 )
24
25
26 class Ace(Assembly):
27 """Class describing an assembly Ace file"""
28
29 file_ext = "ace"
30
31 # def init_meta( self, dataset, copy_from=None ):
32 # Assembly.init_meta( self, dataset, copy_from=copy_from )
33
34 def set_meta( self, dataset, overwrite=True, **kwd ):
35 """
36 Set the number of assembled contigs and read sequences and the number of data lines in dataset.
37 """
38 contigs = 0
39 reads = 0
40 for line in file( dataset.file_name ):
41 line = line.strip()
42 if line and line.startswith( '#' ):
43 # Don't count comment lines
44 continue
45 if line and line.startswith( 'CO' ):
46 contigs += 1
47 if line and line.startswith( 'RD' ):
48 reads += 1
49 dataset.metadata.contigs = contigs
50 dataset.metadata.reads = reads
51
52 def set_peek( self, dataset, is_multi_byte=False ):
53 if not dataset.dataset.purged:
54 dataset.peek = data.get_file_peek( dataset.file_name, is_multi_byte=is_multi_byte )
55 if dataset.metadata.contigs:
56 dataset.blurb = "%s contigs" % util.commaify( str( dataset.metadata.contigs ) )
57 else:
58 dataset.blurb = data.nice_size( dataset.get_size() )
59 else:
60 dataset.peek = 'file does not exist'
61 dataset.blurb = 'file purged from disk'
62
63 def sniff( self, filename ):
64 """
65 Determines whether the file is in ace format
66
67 An ace file contains these sections
68 AS \d+ \d+
69
70 CO \S+ \d+ \d+ \d+ \w
71 [atcgATCGN\*]+
72
73 BQ
74 [\d\s]+
75
76 AF \S+ [CU] \-?\d+
77
78 BS \d+ \d+ \S+
79
80 RD \S+ \d+ \d+ \d+
81 [ATCGN\*]+
82
83 QA \d+ \d+ \d+ \d+
84 DS .*
85
86 Currently we only check if file begins with AS
87
88 >>> fname = get_test_fname( 'genome.ace' )
89 >>> Ace().sniff( fname )
90 True
91 >>> fname = get_test_fname( 'genome.fasta' )
92 >>> Ace().sniff( fname )
93 False
94 """
95
96 try:
97 fh = open( filename )
98 line = fh.readline()
99 line = line.strip()
100 if line:
101 if line.startswith( 'AS ' ):
102 fh.close()
103 return True
104 fh.close()
105 return False
106 except:
107 pass
108 return False
109
110 class Velveth(Assembly):
111 composite_type='basic'
112 file_ext = "txt"
113
114 def __init__(self,**kwd):
115 Assembly.__init__(self,**kwd)
116 self.add_composite_file('Roadmap')
117 self.add_composite_file('Sequences')