|
3
|
1 """
|
|
|
2 BlastXml class
|
|
|
3 """
|
|
|
4
|
|
|
5 from galaxy.datatypes.data import get_file_peek
|
|
|
6 from galaxy.datatypes.data import Text
|
|
|
7 from galaxy.datatypes.xml import GenericXml
|
|
|
8
|
|
|
9 class BlastXml( GenericXml ):
|
|
|
10 """NCBI Blast XML Output data"""
|
|
|
11 file_ext = "blastxml"
|
|
|
12
|
|
|
13 def set_peek( self, dataset, is_multi_byte=False ):
|
|
|
14 """Set the peek and blurb text"""
|
|
|
15 if not dataset.dataset.purged:
|
|
|
16 dataset.peek = get_file_peek( dataset.file_name, is_multi_byte=is_multi_byte )
|
|
|
17 dataset.blurb = 'NCBI Blast XML data'
|
|
|
18 else:
|
|
|
19 dataset.peek = 'file does not exist'
|
|
|
20 dataset.blurb = 'file purged from disk'
|
|
|
21 def sniff( self, filename ):
|
|
|
22 """
|
|
|
23 Determines whether the file is blastxml
|
|
|
24
|
|
|
25 >>> fname = get_test_fname( 'megablast_xml_parser_test1.blastxml' )
|
|
|
26 >>> BlastXml().sniff( fname )
|
|
|
27 True
|
|
|
28 >>> fname = get_test_fname( 'tblastn_four_human_vs_rhodopsin.xml' )
|
|
|
29 >>> BlastXml().sniff( fname )
|
|
|
30 True
|
|
|
31 >>> fname = get_test_fname( 'interval.interval' )
|
|
|
32 >>> BlastXml().sniff( fname )
|
|
|
33 False
|
|
|
34 """
|
|
|
35 #TODO - Use a context manager on Python 2.5+ to close handle
|
|
|
36 handle = open(filename)
|
|
|
37 line = handle.readline()
|
|
|
38 if line.strip() != '<?xml version="1.0"?>':
|
|
|
39 handle.close()
|
|
|
40 return False
|
|
|
41 line = handle.readline()
|
|
|
42 if line.strip() not in ['<!DOCTYPE BlastOutput PUBLIC "-//NCBI//NCBI BlastOutput/EN" "http://www.ncbi.nlm.nih.gov/dtd/NCBI_BlastOutput.dtd">',
|
|
|
43 '<!DOCTYPE BlastOutput PUBLIC "-//NCBI//NCBI BlastOutput/EN" "NCBI_BlastOutput.dtd">']:
|
|
|
44 handle.close()
|
|
|
45 return False
|
|
|
46 line = handle.readline()
|
|
|
47 if line.strip() != '<BlastOutput>':
|
|
|
48 handle.close()
|
|
|
49 return False
|
|
|
50 handle.close()
|
|
|
51 return True
|
|
|
52
|
|
|
53 def merge(split_files, output_file):
|
|
|
54 """Merging multiple XML files is non-trivial and must be done in subclasses."""
|
|
|
55 if len(split_files) == 1:
|
|
|
56 #For one file only, use base class method (move/copy)
|
|
|
57 return Text.merge(split_files, output_file)
|
|
|
58 out = open(output_file, "w")
|
|
|
59 h = None
|
|
|
60 for f in split_files:
|
|
|
61 h = open(f)
|
|
|
62 body = False
|
|
|
63 header = h.readline()
|
|
|
64 if not header:
|
|
|
65 out.close()
|
|
|
66 h.close()
|
|
|
67 raise ValueError("BLAST XML file %s was empty" % f)
|
|
|
68 if header.strip() != '<?xml version="1.0"?>':
|
|
|
69 out.write(header) #for diagnosis
|
|
|
70 out.close()
|
|
|
71 h.close()
|
|
|
72 raise ValueError("%s is not an XML file!" % f)
|
|
|
73 line = h.readline()
|
|
|
74 header += line
|
|
|
75 if line.strip() not in ['<!DOCTYPE BlastOutput PUBLIC "-//NCBI//NCBI BlastOutput/EN" "http://www.ncbi.nlm.nih.gov/dtd/NCBI_BlastOutput.dtd">',
|
|
|
76 '<!DOCTYPE BlastOutput PUBLIC "-//NCBI//NCBI BlastOutput/EN" "NCBI_BlastOutput.dtd">']:
|
|
|
77 out.write(header) #for diagnosis
|
|
|
78 out.close()
|
|
|
79 h.close()
|
|
|
80 raise ValueError("%s is not a BLAST XML file!" % f)
|
|
|
81 while True:
|
|
|
82 line = h.readline()
|
|
|
83 if not line:
|
|
|
84 out.write(header) #for diagnosis
|
|
|
85 out.close()
|
|
|
86 h.close()
|
|
|
87 raise ValueError("BLAST XML file %s ended prematurely" % f)
|
|
|
88 header += line
|
|
|
89 if "<Iteration>" in line:
|
|
|
90 break
|
|
|
91 if len(header) > 10000:
|
|
|
92 #Something has gone wrong, don't load too much into memory!
|
|
|
93 #Write what we have to the merged file for diagnostics
|
|
|
94 out.write(header)
|
|
|
95 out.close()
|
|
|
96 h.close()
|
|
|
97 raise ValueError("BLAST XML file %s has too long a header!" % f)
|
|
|
98 if "<BlastOutput>" not in header:
|
|
|
99 out.close()
|
|
|
100 h.close()
|
|
|
101 raise ValueError("%s is not a BLAST XML file:\n%s\n..." % (f, header))
|
|
|
102 if f == split_files[0]:
|
|
|
103 out.write(header)
|
|
|
104 old_header = header
|
|
|
105 elif old_header[:300] != header[:300]:
|
|
|
106 #Enough to check <BlastOutput_program> and <BlastOutput_version> match
|
|
|
107 out.close()
|
|
|
108 h.close()
|
|
|
109 raise ValueError("BLAST XML headers don't match for %s and %s - have:\n%s\n...\n\nAnd:\n%s\n...\n" \
|
|
|
110 % (split_files[0], f, old_header[:300], header[:300]))
|
|
|
111 else:
|
|
|
112 out.write(" <Iteration>\n")
|
|
|
113 for line in h:
|
|
|
114 if "</BlastOutput_iterations>" in line:
|
|
|
115 break
|
|
|
116 #TODO - Increment <Iteration_iter-num> and if required automatic query names
|
|
|
117 #like <Iteration_query-ID>Query_3</Iteration_query-ID> to be increasing?
|
|
|
118 out.write(line)
|
|
|
119 h.close()
|
|
|
120 out.write(" </BlastOutput_iterations>\n")
|
|
|
121 out.write("</BlastOutput>\n")
|
|
|
122 out.close()
|
|
|
123 merge = staticmethod(merge)
|
|
|
124
|
