def test_nlmsaslice_cache(self):
"NLMSASlice sequence caching & removal"
# set up sequences
dnaseq = testutil.datafile('dnaseq.fasta')
db = SequenceFileDB(dnaseq, autoGC=-1) # use pure WeakValueDict...
try:
gc.collect()
assert len(
db._weakValueDict) == 0, '_weakValueDict should be empty'
seq1, seq2 = db['seq1'], db['seq2']
assert len(db._weakValueDict)==2, \
'_weakValueDict should have 2 seqs'
# build referencing NLMSA
mymap = NLMSA('test', 'memory', db, pairwiseMode=True)
mymap += seq1
mymap[seq1] += seq2
mymap.build()
# check: no cache
assert not hasattr(db, '_cache'), 'should be no cache yet'
seq1, seq2 = db['seq1'], db['seq2'] # re-retrieve
# now retrieve a NLMSASlice, forcing entry of seq into cache
ival = seq1[5:10]
x = mymap[ival]
assert len(db._cache.values()) != 0
n1 = len(db._cache)
assert n1 == 1, "should be exactly one cache entry, not %d" % \
(n1, )
# ok, now trash referencing arguments & make sure of cleanup
del x
gc.collect()
assert len(db._cache.values()) == 0
n2 = len(db._cache)
assert n2 == 0, '%d objects remain; cache memory leak!' % n2
# FAIL because of __dealloc__ error in cnestedlist.NLMSASlice.
# Drop our references, the cache should empty.
del mymap, ival, seq1, seq2
gc.collect()
# check that db._weakValueDict cache is empty
assert len(
db._weakValueDict) == 0, '_weakValueDict should be empty'
finally:
db.close()
def test_nlmsaslice_cache(self):
"NLMSASlice sequence caching & removal"
# set up sequences
dnaseq = testutil.datafile('dnaseq.fasta')
db = SequenceFileDB(dnaseq, autoGC=-1) # use pure WeakValueDict...
try:
gc.collect()
assert len(db._weakValueDict)==0, '_weakValueDict should be empty'
seq1, seq2 = db['seq1'], db['seq2']
assert len(db._weakValueDict)==2, \
'_weakValueDict should have 2 seqs'
# build referencing NLMSA
mymap = NLMSA('test', 'memory', db, pairwiseMode=True)
mymap += seq1
mymap[seq1] += seq2
mymap.build()
# check: no cache
assert not hasattr(db, '_cache'), 'should be no cache yet'
seq1, seq2 = db['seq1'], db['seq2'] # re-retrieve
# now retrieve a NLMSASlice, forcing entry of seq into cache
ival = seq1[5:10]
x = mymap[ival]
assert len(db._cache.values()) != 0
n1 = len(db._cache)
assert n1 == 1, "should be exactly one cache entry, not %d" % \
(n1, )
# ok, now trash referencing arguments & make sure of cleanup
del x
gc.collect()
assert len(db._cache.values()) == 0
n2 = len(db._cache)
assert n2 == 0, '%d objects remain; cache memory leak!' % n2
# FAIL because of __dealloc__ error in cnestedlist.NLMSASlice.
# Drop our references, the cache should empty.
del mymap, ival, seq1, seq2
gc.collect()
# check that db._weakValueDict cache is empty
assert len(db._weakValueDict)==0, '_weakValueDict should be empty'
finally:
db.close()
def nlmsaslice_cache_test(self):
"Test NLMSASlice sequence caching & removal"
# set up sequences
db = SequenceFileDB('dnaseq', autoGC=-1) # use pure WeakValueDict...
gc.collect()
assert len(db._weakValueDict)==0, '_weakValueDict should be empty'
seq1, seq2 = db['seq1'], db['seq2']
assert len(db._weakValueDict)==2, '_weakValueDict should have 2 seqs'
# build referencing NLMSA
mymap = NLMSA('test', 'memory', db, pairwiseMode=True)
mymap += seq1
mymap[seq1] += seq2
mymap.build()
# check: no cache
assert not hasattr(db, '_cache'), 'should be no cache yet'
seq1, seq2 = db['seq1'], db['seq2'] # re-retrieve
# now retrieve a NLMSASlice, forcing entry of seq into cache
ival = seq1[5:10]
x = mymap[ival]
print 'this should not be empty:', db._cache.values()
n1 = len(db._cache)
assert n1 == 1, "should be exactly one cache entry, not %d" % (n1,)
# ok, now trash referencing arguments & make sure of cleanup
del x
gc.collect()
print 'this should be empty:', db._cache.values()
n2 = len(db._cache)
assert n2 == 0, '%d objects remain; cache memory leak!' % n2
# FAIL because of __dealloc__ error in cnestedlist.NLMSASlice.
del mymap, ival, seq1, seq2 # drop our references, cache should empty
gc.collect()
# check that db._weakValueDict cache is empty
assert len(db._weakValueDict)==0, '_weakValueDict should be empty'
#!user/bin/env python
import pygr.Data
from pygr.seqdb import BlastDB, AnnotationDB
seqdb = BlastDB('/somepath/hg18.fa')
exon_slices = pygr.Data.Collection(path='exon_slices.db',mode='c',writeback=False) # Store on disk
exon_db = AnnotationDB(exon_slices, seqdb,sliceAttrDict=dict(id=0, exon_id=1, orientation=2, gene_id=3, start=4, stop=5))
from pygr.cnestedlist import NLMSA
nlmsa = NLMSA('exonAnnot','w',use_virtual_lpo=True,bidirectional=False)
ifile = file('exonslice.txt')
for line in ifile:
row = [x for x in line.split()] # CONVERT TO LIST SO MUTABLE
row[1] = int(row[1]) # CONVERT FROM STRING TO INTEGER
exon_slices[row[1]] = row
exon = exon_db[row[1]] # GET THE ANNOTATION OBJECT FOR THIS EXON
nlmsa.addAnnotation(exon) # SAVE IT TO GENOME MAPPING
exon_db.clear()
nlmsa.build() # FINALIZE GENOME ALIGNMENT INDEXES
ifile.close()
