def genRandomSeqDbCat(path, length, nCopies):
"""Generate random reference sequences as SeqDbFasta.
The first sequence is generated as random, and each next
one is a concatenation of N copies of the first."""
seqDb = SeqDbFasta(path, save=True)
seqBase = randomSeq(length)
# assert len(set(lengths)) == len(lengths),"Reference lengths must be unique"
seqLen = {}
for iIcm in xrange(nCopies):
idIcm = iIcm
writer = seqDb.fastaWriter(idIcm)
seq = n.tile(seqBase, iIcm + 1)
writer.record(str(idIcm), seq)
writer.close()
seqLen[idIcm] = len(seq)
seqDb.opt.seqLen = seqLen
seqDb.save()
def genRandomSeqDbIid(path, lengths):
"""Generate random reference sequences as SeqDbFasta.
Each generated sequence is generated independently"""
seqDb = SeqDbFasta(path, save=True)
# assert len(set(lengths)) == len(lengths),"Reference lengths must be unique"
seqLen = {}
for (iIcm, length) in enumerate(lengths):
idIcm = iIcm
writer = seqDb.fastaWriter(idIcm)
seq = randomSeq(length)
writer.record(str(idIcm), seq)
writer.close()
seqDb.finById(idIcm)
seqLen[idIcm] = length
seqDb.opt.seqLen = seqLen
seqDb.save()
def genInput():
genRandomSeqDbIid(seqDbPath, seqDbLengths)
genRandomQueryZeroMarkov(queryPath, queryLengths)
if len(sys.argv) == 1:
genInput()
opt = Struct()
opt.runMode = "batchDep"
# training opts
opt.seqDb = seqDbPath
seqDb = SeqDbFasta(opt.seqDb)
ids = seqDb.getIdList()
immIdToSeqIds = dict(((id, [id]) for id in ids))
immIdsFile = pjoin(workDir, "test.immapp.seqids.pkl")
dumpObj(immIdToSeqIds, immIdsFile)
opt.immIdToSeqIds = immIdsFile
opt.immDb = pjoin(workDir, "icm")
# scoring opts
immIds = ImmStore.open(path=opt.immDb, mode="r").listImmIdsWithIdScoreIdent()
immIdsFile = "test.immapp.immids.pkl"
dumpObj(immIds, immIdsFile)
opt.immIds = immIdsFile
opt.nImmBatches = 10
opt.inpSeq = queryPath
from MGT.SeqDbFasta import *
seqDbPath = pjoin(options.testDataDir,"seqdb-fasta")
seqDb = SeqDbFasta(seqDbPath)
ids = seqDb.getIdList()
for id in ids:
seqDb.writeFastaBothStrands([id],"%s.tmp.fna.gz" % (id,))
from MGT.ImmApp import *
from MGT.SeqDbFasta import *
seqDbPath = pjoin(options.testDataDir, "seqdb-fasta")
jobs = []
opt = Struct()
opt.runMode = "inproc" # "batchDep"
opt.seqDb = seqDbPath
opt.immDb = pabs("test.immdb")
seqDb = SeqDbFasta(opt.seqDb)
ids = seqDb.getIdList()
for id in ids:
seqDb.finById(id=id)
immIdToSeqIds = dict(((id, [id]) for id in ids))
immIdToSeqIdsFile = pabs("test.immapp.seqids.pkl")
dumpObj(immIdToSeqIds, immIdToSeqIdsFile)
opt.immIdToSeqIds = immIdToSeqIdsFile
opt.mode = "train"
ImmApp.fillWithDefaultOptions(opt)
# imm = ImmApp(opt=opt)
# jobs = imm.run()
for (reduceScoresEarly, cwd) in ((1, "imm.test.red_early_1"), (0, "imm.test.red_early_0")):
immIds = ImmStore.open(path=opt.immDb, mode="r").listImmIdsWithIdScoreIdent()
immIdsFile = pabs("test.immapp.immids.pkl")
from MGT.TaxaTreeUtils import *
"""Count taxonomic groups represented in SeqDbFasta object.
Produces an SQLite DB file containing a table with a "Linnean"
(in a broad sense) lineage for each SeqDb record, as well as
the aggregate count tables for each taxonomic level."""
refDataDir = "../../atovtchi/shannon_viral_IO_paper/shannon_viral_paper_2011.v.1"
workDir = refDataDir
#dbPath = pjoin(workDir,"refseq-taxa")
dbPath = "seq-db.mic"
#taxonomyDir = pjoin(workDir,"taxonomy")
topTaxids = micTaxids
dbSeq = SeqDbFasta(path=dbPath)
taxids = dbSeq.getTaxaList()
#taxaTree = loadTaxaTree(ncbiDumpFile=pjoin(taxonomyDir,"nodes.dmp"),
# ncbiNamesDumpFile=pjoin(taxonomyDir,"names.dmp"))
taxaTree = loadTaxaTree()
topNodes = [ taxaTree.getNode(topTaxid) for topTaxid in topTaxids ]
linwr = LinnWriter(taxaTree=taxaTree)
wr = linwr.newWriter()
for taxid in taxids:
node = taxaTree.getNode(taxid)
if sum(( node.isUnder(topNode) for topNode in topNodes )):
wr.send(dict(taxid=taxid,weight=dbSeq.seqLengths(taxid)["len"].sum()))
