class _BamReaderBase(object):
"""
The BamReader class provides a high-level interface to PacBio BAM
files. If a PacBio BAM index (bam.pbi file) is present and the
user instantiates the BamReader using the reference FASTA as the
second argument, the BamReader will provide an interface
compatible with CmpH5Reader.
"""
def _loadReferenceInfo(self):
refRecords = self.peer.header["SQ"]
refNames = [r["SN"] for r in refRecords]
refLengths = [r["LN"] for r in refRecords]
refMD5s = [r["M5"] for r in refRecords]
refIds = map(self.peer.gettid, refNames)
nRefs = len(refRecords)
self._referenceInfoTable = np.rec.fromrecords(zip(
refIds,
refIds,
refNames,
refNames,
refLengths,
refMD5s,
np.zeros(nRefs, dtype=np.uint32),
np.zeros(nRefs, dtype=np.uint32)),
dtype=[('ID', '<i8'), ('RefInfoID', '<i8'),
('Name', 'O'), ('FullName', 'O'),
('Length', '<i8'), ('MD5', 'O'),
('StartRow', '<u4'), ('EndRow', '<u4')])
self._referenceDict = {}
self._referenceDict.update(zip(refIds, self._referenceInfoTable))
self._referenceDict.update(zip(refNames, self._referenceInfoTable))
def _loadReadGroupInfo(self):
rgs = self.peer.header["RG"]
readGroupTable_ = []
pulseFeaturesInAll_ = frozenset(PULSE_FEATURE_TAGS.keys())
for rg in rgs:
# Regarding RG ID: BLASR currently outputs a hex digest of
# 10 nibbles, instead of the 8 which would fit into a
# 32-bit word. So we truncate here for the purposes of
# cross-referencing within this API and the PacBioBamIndex
# API. We do check for a collision below.
rgID = int(rg["ID"][:8], 16)
rgName = rg["PU"]
ds = dict([pair.split("=") for pair in rg["DS"].split(";") if pair != ""])
triple = ds["BINDINGKIT"], ds["SEQUENCINGKIT"], ds["SOFTWAREVERSION"]
rgChem = decodeTriple(*triple)
rgReadType = ds["READTYPE"]
readGroupTable_.append((rgID, rgName, rgReadType, rgChem))
pulseFeaturesInAll_ = pulseFeaturesInAll_.intersection(ds.keys())
self._readGroupTable = np.rec.fromrecords(
readGroupTable_,
dtype=[("ID" , np.uint32),
("MovieName" , "O"),
("ReadType" , "O"),
("SequencingChemistry", "O")])
assert len(set(self._readGroupTable.ID)) == len(self._readGroupTable), \
"First 8 chars of read group IDs must be unique!"
self._readGroupDict = { rg.ID : rg
for rg in self._readGroupTable }
self._pulseFeaturesAvailable = pulseFeaturesInAll_
def _loadProgramInfo(self):
# TODO: guarantee that these fields are nonoptional in our bams --- check with Marcus
# TODO: are we interesting in the PP info?
self._programTable = np.rec.fromrecords(
[ (pg["ID"], pg.get("VN", None), pg.get("CL", None))
for pg in self.peer.header["PG"] ],
dtype=[("ID" , "O"),
("Version", "O"),
("CommandLine", "O")])
def _loadReferenceFasta(self, referenceFastaFname):
ft = FastaTable(referenceFastaFname)
# Verify that this FASTA is in agreement with the BAM's
# reference table---BAM should be a subset.
fastaIdsAndLens = set((c.id, c.length) for c in ft)
bamIdsAndLens = set((c.Name, c.Length) for c in self.referenceInfoTable)
if not bamIdsAndLens.issubset(fastaIdsAndLens):
raise ReferenceMismatch, "FASTA file must contain superset of reference contigs in BAM"
self.referenceFasta = ft
def __init__(self, fname, referenceFastaFname=None):
self.filename = fname = abspath(expanduser(fname))
self.peer = Samfile(fname, "rb")
# Check for sortedness, index.
# There doesn't seem to be a "public" way to do this right
# now, but that's fine because we're going to have to rewrite
# it all anyway once the pysam rewrite lands.
if not self.peer._hasIndex:
raise ValueError, "Specified bam file lacks a bam index---required for this API"
self._loadReferenceInfo()
self._loadReadGroupInfo()
self._loadProgramInfo()
self.referenceFasta = None
if referenceFastaFname is not None:
self._loadReferenceFasta(referenceFastaFname)
@property
def isIndexLoaded(self):
return self.index is not None
@property
def isReferenceLoaded(self):
return self.referenceFasta is not None
def attach(self, fofnFilename):
self.basH5Collection = BasH5Collection(fofnFilename)
@property
def moviesAttached(self):
return (self.basH5Collection is not None)
@property
def alignmentIndex(self):
raise UnavailableFeature("BAM has no alignment index")
#TODO: change concept to readGroupTable in cmp.h5
@property
def movieInfoTable(self):
raise Unimplemented()
# TODO: change to read group accessor, this is semantically wrong now
def movieInfo(self, movieId):
raise Unimplemented()
@property
def movieNames(self):
return set([mi.MovieName for mi in self.readGroupTable])
@property
def readGroupTable(self):
return self._readGroupTable
def readGroup(self, readGroupId):
return self._readGroupDict[readGroupId]
@property
def sequencingChemistry(self):
"""
List of the sequencing chemistries by movie. Order is
unspecified.
"""
return list(self.readGroupTable.SequencingChemistry)
#TODO: elide "Info" innames?
@property
def referenceInfoTable(self):
return self._referenceInfoTable
#TODO: standard? how about subread instead? why capitalize ccs?
# can we standardize this? is cDNA an additional possibility
@property
def readType(self):
"""
Either "standard", "CCS", "mixed", or "unknown", to represent the
type of PacBio reads aligned in this BAM file.
"""
readTypes = self.readGroupTable.ReadType
if all(readTypes == "SUBREAD"):
return "standard"
elif all(readTypes == "CCS"):
return "CCS"
elif all((readTypes == "CCS") | (readTypes == "SUBREAD")):
return "mixed"
else:
return "unknown"
#TODO: Marcus needs to put something in the spec for this
@property
def version(self):
raise Unimplemented()
#TODO: Marcus needs to put something in the spec for this
def versionAtLeast(self, minimalVersion):
raise Unimplemented()
def softwareVersion(self, programName):
raise Unimplemented()
@property
def isSorted(self):
return True
@property
def isBarcoded(self):
raise Unimplemented()
@property
def isEmpty(self):
return (len(self) == 0)
# TODO: make this private in cmp.h5 reader
def alignmentGroup(self, alnGroupId):
raise UnavailableFeature("BAM has no HDF5 groups")
def referenceInfo(self, key):
return self._referenceDict[key]
def atOffset(self, offset):
self.peer.seek(offset)
return BamAlignment(self, next(self.peer))
def hasPulseFeature(self, featureName):
return featureName in self._pulseFeaturesAvailable
def pulseFeaturesAvailable(self):
return self._pulseFeaturesAvailable
@property
def barcode(self):
raise Unimplemented()
@property
def barcodeName(self):
raise Unimplemented()
@property
def barcodes(self):
raise Unimplemented()
def __repr__(self):
return "<%s for %s>" % (type(self).__name__, self.filename)
def __len__(self):
return self.peer.mapped
def close(self):
if hasattr(self, "file") and self.file is not None:
self.file.close()
self.file = None
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
self.close()
class _BamReaderBase(ReaderBase):
"""
The BamReader class provides a high-level interface to PacBio BAM
files. If a PacBio BAM index (bam.pbi file) is present and the
user instantiates the BamReader using the reference FASTA as the
second argument, the BamReader will provide an interface
compatible with CmpH5Reader.
"""
def _loadReferenceInfo(self):
refRecords = self.peer.header["SQ"]
refNames = [r["SN"] for r in refRecords]
refLengths = [r["LN"] for r in refRecords]
refMD5s = [r["M5"] for r in refRecords]
refIds = map(self.peer.gettid, refNames)
nRefs = len(refRecords)
if nRefs > 0:
self._referenceInfoTable = np.rec.fromrecords(
zip(refIds, refIds, refNames, refNames, refLengths, refMD5s,
np.zeros(nRefs, dtype=np.uint32),
np.zeros(nRefs, dtype=np.uint32)),
dtype=[('ID', '<i8'), ('RefInfoID', '<i8'), ('Name', 'O'),
('FullName', 'O'), ('Length', '<i8'), ('MD5', 'O'),
('StartRow', '<u4'), ('EndRow', '<u4')])
self._referenceDict = {}
self._referenceDict.update(zip(refIds, self._referenceInfoTable))
self._referenceDict.update(zip(refNames, self._referenceInfoTable))
else:
self._referenceInfoTable = None
self._referenceDict = None
def _loadReadGroupInfo(self):
rgs = self.peer.header["RG"]
readGroupTable_ = []
pulseFeaturesInAll_ = frozenset(PULSE_FEATURE_TAGS.keys())
for rg in rgs:
rgID = rgAsInt(rg["ID"])
rgName = rg["PU"]
ds = dict([
pair.split("=") for pair in rg["DS"].split(";") if pair != ""
])
# spec: we only consider first two components of basecaller version
# in "chem" lookup
basecallerVersion = ".".join(
ds["BASECALLERVERSION"].split(".")[0:2])
triple = ds["BINDINGKIT"], ds["SEQUENCINGKIT"], basecallerVersion
rgChem = decodeTriple(*triple)
rgReadType = ds["READTYPE"]
# TODO(dalexander): need FRAMERATEHZ in RG::DS!
#rgFrameRate = ds["FRAMERATEHZ"]
rgFrameRate = 75.0
readGroupTable_.append(
(rgID, rgName, rgReadType, rgChem, rgFrameRate))
pulseFeaturesInAll_ = pulseFeaturesInAll_.intersection(ds.keys())
self._readGroupTable = np.rec.fromrecords(readGroupTable_,
dtype=[
("ID", np.int32),
("MovieName", "O"),
("ReadType", "O"),
("SequencingChemistry",
"O"),
("FrameRate", float)
])
assert len(set(self._readGroupTable.ID)) == len(self._readGroupTable), \
"First 8 chars of read group IDs must be unique!"
self._readGroupDict = {rg.ID: rg for rg in self._readGroupTable}
self._pulseFeaturesAvailable = pulseFeaturesInAll_
def _loadProgramInfo(self):
pgRecords = [(pg["ID"], pg.get("VN", None), pg.get("CL", None))
for pg in self.peer.header.get("PG", [])]
if len(pgRecords) > 0:
self._programTable = np.rec.fromrecords(pgRecords,
dtype=[("ID", "O"),
("Version", "O"),
("CommandLine", "O")
])
else:
self._programTable = None
def _loadReferenceFasta(self, referenceFastaFname):
ft = FastaTable(referenceFastaFname)
# Verify that this FASTA is in agreement with the BAM's
# reference table---BAM should be a subset.
fastaIdsAndLens = set((c.id, len(c)) for c in ft)
bamIdsAndLens = set(
(c.Name, c.Length) for c in self.referenceInfoTable)
if not bamIdsAndLens.issubset(fastaIdsAndLens):
raise ReferenceMismatch, "FASTA file must contain superset of reference contigs in BAM"
self.referenceFasta = ft
def _checkFileCompatibility(self):
# Verify that this is a "pacbio" BAM file of version at least
# 3.0b3
try:
checkedVersion = self.version
except:
raise IncompatibleFile(
"This BAM file is incompatible with this API " +
"(only PacBio BAM files version >= 3.0b3 are supported)")
def __init__(self, fname, referenceFastaFname=None):
self.filename = fname = abspath(expanduser(fname))
self.peer = Samfile(fname, "rb", check_sq=False)
self._checkFileCompatibility()
self._loadReferenceInfo()
self._loadReadGroupInfo()
self._loadProgramInfo()
self.referenceFasta = None
if referenceFastaFname is not None:
if self.isUnmapped:
raise ValueError, "Unmapped BAM file--reference FASTA should not be given as argument to BamReader"
self._loadReferenceFasta(referenceFastaFname)
@property
def isIndexLoaded(self):
return self.index is not None
@property
def isReferenceLoaded(self):
return self.referenceFasta is not None
@property
def isUnmapped(self):
return not (self.isMapped)
@property
def isMapped(self):
return len(self.peer.header["SQ"]) > 0
@property
def alignmentIndex(self):
raise UnavailableFeature("BAM has no alignment index")
@property
def movieNames(self):
return set([mi.MovieName for mi in self.readGroupTable])
@property
def readGroupTable(self):
return self._readGroupTable
def readGroupInfo(self, readGroupId):
return self._readGroupDict[readGroupId]
@property
def sequencingChemistry(self):
"""
List of the sequencing chemistries by movie. Order is
unspecified.
"""
return list(self.readGroupTable.SequencingChemistry)
@property
def referenceInfoTable(self):
return self._referenceInfoTable
#TODO: standard? how about subread instead? why capitalize ccs?
# can we standardize this? is cDNA an additional possibility
@property
def readType(self):
"""
Either "standard", "CCS", "mixed", or "unknown", to represent the
type of PacBio reads aligned in this BAM file.
"""
readTypes = self.readGroupTable.ReadType
if all(readTypes == "SUBREAD"):
return "standard"
elif all(readTypes == "CCS"):
return "CCS"
elif all((readTypes == "CCS") | (readTypes == "SUBREAD")):
return "mixed"
else:
return "unknown"
@property
def version(self):
return self.peer.header["HD"]["pb"]
def versionAtLeast(self, minimalVersion):
raise Unimplemented()
def softwareVersion(self, programName):
raise Unimplemented()
@property
def isSorted(self):
return self.peer.header["HD"]["SO"] == "coordinate"
@property
def isBarcoded(self):
raise Unimplemented()
@property
def isEmpty(self):
return (len(self) == 0)
def referenceInfo(self, key):
return self._referenceDict[key]
def atOffset(self, offset):
self.peer.seek(offset)
return BamAlignment(self, next(self.peer))
def hasPulseFeature(self, featureName):
return featureName in self._pulseFeaturesAvailable
def pulseFeaturesAvailable(self):
return self._pulseFeaturesAvailable
@property
def barcode(self):
raise Unimplemented()
@property
def barcodeName(self):
raise Unimplemented()
@property
def barcodes(self):
raise Unimplemented()
@requiresBai
def __len__(self):
return self.peer.mapped + self.peer.unmapped
def close(self):
if hasattr(self, "file") and self.file is not None:
self.file.close()
self.file = None
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
self.close()
class _BamReaderBase(ReaderBase):
"""
The BamReader class provides a high-level interface to PacBio BAM
files. If a PacBio BAM index (bam.pbi file) is present and the
user instantiates the BamReader using the reference FASTA as the
second argument, the BamReader will provide an interface
compatible with CmpH5Reader.
"""
def _loadReferenceInfo(self):
refRecords = self.peer.header["SQ"]
refNames = [r["SN"] for r in refRecords]
refLengths = [r["LN"] for r in refRecords]
refMD5s = [r["M5"] for r in refRecords]
refIds = map(self.peer.gettid, refNames)
nRefs = len(refRecords)
if nRefs > 0:
self._referenceInfoTable = np.rec.fromrecords(zip(
refIds,
refIds,
refNames,
refNames,
refLengths,
refMD5s,
np.zeros(nRefs, dtype=np.uint32),
np.zeros(nRefs, dtype=np.uint32)),
dtype=[('ID', '<i8'), ('RefInfoID', '<i8'),
('Name', 'O'), ('FullName', 'O'),
('Length', '<i8'), ('MD5', 'O'),
('StartRow', '<u4'), ('EndRow', '<u4')])
self._referenceDict = {}
self._referenceDict.update(zip(refIds, self._referenceInfoTable))
self._referenceDict.update(zip(refNames, self._referenceInfoTable))
else:
self._referenceInfoTable = None
self._referenceDict = None
def _loadReadGroupInfo(self):
rgs = self.peer.header["RG"]
readGroupTable_ = []
pulseFeaturesInAll_ = frozenset(PULSE_FEATURE_TAGS.keys())
for rg in rgs:
rgID = rgAsInt(rg["ID"])
rgName = rg["PU"]
ds = dict([pair.split("=") for pair in rg["DS"].split(";") if pair != ""])
# spec: we only consider first two components of basecaller version
# in "chem" lookup
basecallerVersion = ".".join(ds["BASECALLERVERSION"].split(".")[0:2])
triple = ds["BINDINGKIT"], ds["SEQUENCINGKIT"], basecallerVersion
rgChem = decodeTriple(*triple)
rgReadType = ds["READTYPE"]
# TODO(dalexander): need FRAMERATEHZ in RG::DS!
#rgFrameRate = ds["FRAMERATEHZ"]
rgFrameRate = 75.0
readGroupTable_.append((rgID, rgName, rgReadType, rgChem, rgFrameRate))
pulseFeaturesInAll_ = pulseFeaturesInAll_.intersection(ds.keys())
self._readGroupTable = np.rec.fromrecords(
readGroupTable_,
dtype=[("ID" , np.int32),
("MovieName" , "O"),
("ReadType" , "O"),
("SequencingChemistry", "O"),
("FrameRate", float)])
assert len(set(self._readGroupTable.ID)) == len(self._readGroupTable), \
"First 8 chars of read group IDs must be unique!"
self._readGroupDict = { rg.ID : rg
for rg in self._readGroupTable }
self._pulseFeaturesAvailable = pulseFeaturesInAll_
def _loadProgramInfo(self):
pgRecords = [ (pg["ID"], pg.get("VN", None), pg.get("CL", None))
for pg in self.peer.header.get("PG", []) ]
if len(pgRecords) > 0:
self._programTable = np.rec.fromrecords(
pgRecords,
dtype=[("ID" , "O"),
("Version", "O"),
("CommandLine", "O")])
else:
self._programTable = None
def _loadReferenceFasta(self, referenceFastaFname):
ft = FastaTable(referenceFastaFname)
# Verify that this FASTA is in agreement with the BAM's
# reference table---BAM should be a subset.
fastaIdsAndLens = set((c.id, len(c)) for c in ft)
bamIdsAndLens = set((c.Name, c.Length) for c in self.referenceInfoTable)
if not bamIdsAndLens.issubset(fastaIdsAndLens):
raise ReferenceMismatch, "FASTA file must contain superset of reference contigs in BAM"
self.referenceFasta = ft
def _checkFileCompatibility(self):
# Verify that this is a "pacbio" BAM file of version at least
# 3.0b3
try:
checkedVersion = self.version
except:
raise IncompatibleFile(
"This BAM file is incompatible with this API " +
"(only PacBio BAM files version >= 3.0b3 are supported)")
def __init__(self, fname, referenceFastaFname=None):
self.filename = fname = abspath(expanduser(fname))
self.peer = Samfile(fname, "rb", check_sq=False)
self._checkFileCompatibility()
self._loadReferenceInfo()
self._loadReadGroupInfo()
self._loadProgramInfo()
self.referenceFasta = None
if referenceFastaFname is not None:
if self.isUnmapped:
raise ValueError, "Unmapped BAM file--reference FASTA should not be given as argument to BamReader"
self._loadReferenceFasta(referenceFastaFname)
@property
def isIndexLoaded(self):
return self.index is not None
@property
def isReferenceLoaded(self):
return self.referenceFasta is not None
@property
def isUnmapped(self):
return not(self.isMapped)
@property
def isMapped(self):
return len(self.peer.header["SQ"]) > 0
@property
def alignmentIndex(self):
raise UnavailableFeature("BAM has no alignment index")
@property
def movieNames(self):
return set([mi.MovieName for mi in self.readGroupTable])
@property
def readGroupTable(self):
return self._readGroupTable
def readGroupInfo(self, readGroupId):
return self._readGroupDict[readGroupId]
@property
def sequencingChemistry(self):
"""
List of the sequencing chemistries by movie. Order is
unspecified.
"""
return list(self.readGroupTable.SequencingChemistry)
@property
def referenceInfoTable(self):
return self._referenceInfoTable
#TODO: standard? how about subread instead? why capitalize ccs?
# can we standardize this? is cDNA an additional possibility
@property
def readType(self):
"""
Either "standard", "CCS", "mixed", or "unknown", to represent the
type of PacBio reads aligned in this BAM file.
"""
readTypes = self.readGroupTable.ReadType
if all(readTypes == "SUBREAD"):
return "standard"
elif all(readTypes == "CCS"):
return "CCS"
elif all((readTypes == "CCS") | (readTypes == "SUBREAD")):
return "mixed"
else:
return "unknown"
@property
def version(self):
return self.peer.header["HD"]["pb"]
def versionAtLeast(self, minimalVersion):
raise Unimplemented()
def softwareVersion(self, programName):
raise Unimplemented()
@property
def isSorted(self):
return self.peer.header["HD"]["SO"] == "coordinate"
@property
def isBarcoded(self):
raise Unimplemented()
@property
def isEmpty(self):
return (len(self) == 0)
def referenceInfo(self, key):
return self._referenceDict[key]
def atOffset(self, offset):
self.peer.seek(offset)
return BamAlignment(self, next(self.peer))
def hasPulseFeature(self, featureName):
return featureName in self._pulseFeaturesAvailable
def pulseFeaturesAvailable(self):
return self._pulseFeaturesAvailable
@property
def barcode(self):
raise Unimplemented()
@property
def barcodeName(self):
raise Unimplemented()
@property
def barcodes(self):
raise Unimplemented()
@requiresBai
def __len__(self):
return self.peer.mapped + self.peer.unmapped
def close(self):
if hasattr(self, "file") and self.file is not None:
self.file.close()
self.file = None
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
self.close()
class _BamReaderBase(object):
"""
The BamReader class provides a high-level interface to PacBio BAM
files. If a PacBio BAM index (bam.pbi file) is present and the
user instantiates the BamReader using the reference FASTA as the
second argument, the BamReader will provide an interface
compatible with CmpH5Reader.
"""
def _loadReferenceInfo(self):
refRecords = self.peer.header["SQ"]
refNames = [r["SN"] for r in refRecords]
refLengths = [r["LN"] for r in refRecords]
refMD5s = [r["M5"] for r in refRecords]
refIds = map(self.peer.gettid, refNames)
nRefs = len(refRecords)
self._referenceInfoTable = np.rec.fromrecords(
zip(refIds, refIds, refNames, refNames, refLengths, refMD5s,
np.zeros(nRefs, dtype=np.uint32),
np.zeros(nRefs, dtype=np.uint32)),
dtype=[('ID', '<i8'), ('RefInfoID', '<i8'), ('Name', 'O'),
('FullName', 'O'), ('Length', '<i8'), ('MD5', 'O'),
('StartRow', '<u4'), ('EndRow', '<u4')])
self._referenceDict = {}
self._referenceDict.update(zip(refIds, self._referenceInfoTable))
self._referenceDict.update(zip(refNames, self._referenceInfoTable))
def _loadReadGroupInfo(self):
rgs = self.peer.header["RG"]
readGroupTable_ = []
pulseFeaturesInAll_ = frozenset(PULSE_FEATURE_TAGS.keys())
for rg in rgs:
# Regarding RG ID: BLASR currently outputs a hex digest of
# 10 nibbles, instead of the 8 which would fit into a
# 32-bit word. So we truncate here for the purposes of
# cross-referencing within this API and the PacBioBamIndex
# API. We do check for a collision below.
rgID = int(rg["ID"][:8], 16)
rgName = rg["PU"]
ds = dict([
pair.split("=") for pair in rg["DS"].split(";") if pair != ""
])
triple = ds["BINDINGKIT"], ds["SEQUENCINGKIT"], ds[
"SOFTWAREVERSION"]
rgChem = decodeTriple(*triple)
rgReadType = ds["READTYPE"]
readGroupTable_.append((rgID, rgName, rgReadType, rgChem))
pulseFeaturesInAll_ = pulseFeaturesInAll_.intersection(ds.keys())
self._readGroupTable = np.rec.fromrecords(readGroupTable_,
dtype=[
("ID", np.uint32),
("MovieName", "O"),
("ReadType", "O"),
("SequencingChemistry",
"O")
])
assert len(set(self._readGroupTable.ID)) == len(self._readGroupTable), \
"First 8 chars of read group IDs must be unique!"
self._readGroupDict = {rg.ID: rg for rg in self._readGroupTable}
self._pulseFeaturesAvailable = pulseFeaturesInAll_
def _loadProgramInfo(self):
# TODO: guarantee that these fields are nonoptional in our bams --- check with Marcus
# TODO: are we interesting in the PP info?
self._programTable = np.rec.fromrecords(
[(pg["ID"], pg.get("VN", None), pg.get("CL", None))
for pg in self.peer.header["PG"]],
dtype=[("ID", "O"), ("Version", "O"), ("CommandLine", "O")])
def _loadReferenceFasta(self, referenceFastaFname):
ft = FastaTable(referenceFastaFname)
# Verify that this FASTA is in agreement with the BAM's
# reference table---BAM should be a subset.
fastaIdsAndLens = set((c.id, c.length) for c in ft)
bamIdsAndLens = set(
(c.Name, c.Length) for c in self.referenceInfoTable)
if not bamIdsAndLens.issubset(fastaIdsAndLens):
raise ReferenceMismatch, "FASTA file must contain superset of reference contigs in BAM"
self.referenceFasta = ft
def __init__(self, fname, referenceFastaFname=None):
self.filename = fname = abspath(expanduser(fname))
self.peer = Samfile(fname, "rb")
# Check for sortedness, index.
# There doesn't seem to be a "public" way to do this right
# now, but that's fine because we're going to have to rewrite
# it all anyway once the pysam rewrite lands.
if not self.peer._hasIndex:
raise ValueError, "Specified bam file lacks a bam index---required for this API"
self._loadReferenceInfo()
self._loadReadGroupInfo()
self._loadProgramInfo()
self.referenceFasta = None
if referenceFastaFname is not None:
self._loadReferenceFasta(referenceFastaFname)
@property
def isIndexLoaded(self):
return self.index is not None
@property
def isReferenceLoaded(self):
return self.referenceFasta is not None
def attach(self, fofnFilename):
self.basH5Collection = BasH5Collection(fofnFilename)
@property
def moviesAttached(self):
return (self.basH5Collection is not None)
@property
def alignmentIndex(self):
raise UnavailableFeature("BAM has no alignment index")
#TODO: change concept to readGroupTable in cmp.h5
@property
def movieInfoTable(self):
raise Unimplemented()
# TODO: change to read group accessor, this is semantically wrong now
def movieInfo(self, movieId):
raise Unimplemented()
@property
def movieNames(self):
return set([mi.MovieName for mi in self.readGroupTable])
@property
def readGroupTable(self):
return self._readGroupTable
def readGroup(self, readGroupId):
return self._readGroupDict[readGroupId]
@property
def sequencingChemistry(self):
"""
List of the sequencing chemistries by movie. Order is
unspecified.
"""
return list(self.readGroupTable.SequencingChemistry)
#TODO: elide "Info" innames?
@property
def referenceInfoTable(self):
return self._referenceInfoTable
#TODO: standard? how about subread instead? why capitalize ccs?
# can we standardize this? is cDNA an additional possibility
@property
def readType(self):
"""
Either "standard", "CCS", "mixed", or "unknown", to represent the
type of PacBio reads aligned in this BAM file.
"""
readTypes = self.readGroupTable.ReadType
if all(readTypes == "SUBREAD"):
return "standard"
elif all(readTypes == "CCS"):
return "CCS"
elif all((readTypes == "CCS") | (readTypes == "SUBREAD")):
return "mixed"
else:
return "unknown"
#TODO: Marcus needs to put something in the spec for this
@property
def version(self):
raise Unimplemented()
#TODO: Marcus needs to put something in the spec for this
def versionAtLeast(self, minimalVersion):
raise Unimplemented()
def softwareVersion(self, programName):
raise Unimplemented()
@property
def isSorted(self):
return True
@property
def isBarcoded(self):
raise Unimplemented()
@property
def isEmpty(self):
return (len(self) == 0)
# TODO: make this private in cmp.h5 reader
def alignmentGroup(self, alnGroupId):
raise UnavailableFeature("BAM has no HDF5 groups")
def referenceInfo(self, key):
return self._referenceDict[key]
def atOffset(self, offset):
self.peer.seek(offset)
return BamAlignment(self, next(self.peer))
def hasPulseFeature(self, featureName):
return featureName in self._pulseFeaturesAvailable
def pulseFeaturesAvailable(self):
return self._pulseFeaturesAvailable
@property
def barcode(self):
raise Unimplemented()
@property
def barcodeName(self):
raise Unimplemented()
@property
def barcodes(self):
raise Unimplemented()
def __repr__(self):
return "<%s for %s>" % (type(self).__name__, self.filename)
def __len__(self):
return self.peer.mapped
def close(self):
if hasattr(self, "file") and self.file is not None:
self.file.close()
self.file = None
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
self.close()