Mercurial > repos > nick > duplex

annotate swalign.py @ 4:af383638de66 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
author nick
date Mon, 23 Nov 2015 18:44:23 -0500
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
4
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
1 import os
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
2 import ctypes
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
3
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
4 script_dir = os.path.dirname(os.path.realpath(__file__))
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
5 swalign = ctypes.cdll.LoadLibrary(os.path.join(script_dir, 'swalignc.so'))
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
6
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
7
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
8 # C struct for ctypes
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
9 class SeqPairC(ctypes.Structure):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
10 _fields_ = [
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
11 ('a', ctypes.c_char_p),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
12 ('alen', ctypes.c_uint),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
13 ('b', ctypes.c_char_p),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
14 ('blen', ctypes.c_uint),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
15 ]
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
16
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
17
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
18 # C struct for ctypes
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
19 class AlignC(ctypes.Structure):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
20 _fields_ = [
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
21 ('seqs', ctypes.POINTER(SeqPairC)),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
22 ('start_a', ctypes.c_int),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
23 ('start_b', ctypes.c_int),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
24 ('end_a', ctypes.c_int),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
25 ('end_b', ctypes.c_int),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
26 ('matches', ctypes.c_int),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
27 ('score', ctypes.c_double),
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
28 ]
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
29
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
30
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
31 # The Python version
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
32 class Align(object):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
33 def __init__(self, align_c):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
34 self.target = align_c.seqs.contents.a
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
35 self.query = align_c.seqs.contents.b
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
36 self.start_target = align_c.start_a
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
37 self.start_query = align_c.start_b
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
38 self.end_target = align_c.end_a
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
39 self.end_query = align_c.end_b
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
40 self.matches = align_c.matches
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
41 self.score = align_c.score
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
42
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
43
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
44 # Initialize functions (define types).
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
45 swalign.smith_waterman.restype = ctypes.POINTER(AlignC)
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
46 swalign.revcomp.restype = ctypes.c_char_p
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
47
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
48
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
49 def smith_waterman(target, query):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
50 seq_pair = SeqPairC(target, len(target), query, len(query))
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
51 align_c = swalign.smith_waterman(ctypes.pointer(seq_pair), 1).contents
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
52 return Align(align_c)
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
53
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
54
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
55 def revcomp(seq):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
56 """WARNING: This will alter the input string in-place!"""
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
57 swalign.revcomp(seq)