def crossTabulateHits(hitFile1, hitFile2,
skipHeader1=True, skipHeader2=True,
readCol1=0, readCol2=0,
hitCol1=-1, hitCol2=-1,
hitSep1=None, hitSep2=None):
"""
Given two hit tables, cross reference reads to build a 2D DataFrame
of hit counts
"""
# get a map from reads to hits in file 1
hitmap = tupleIteratorToMap(
parseHits(
hitFile1,
readCol1,
hitCol1,
skipHeader1,
hitSep1))
# get an iterator over hits in file 2
crossHits = parseHits(hitFile2, readCol2, hitCol2, skipHeader2, hitSep2)
counts = {}
types2 = set()
readsOnlyIn2 = []
readMatchCount = 0
for read, hits in crossHits:
try:
hits2 = hitmap.pop(read)
readMatchCount += 1
except KeyError:
readsOnlyIn2.append(read)
continue
for h2 in hits2:
for h1 in hits:
h1counts = counts.setdefault(h1, {})
h1counts[h2] = h1counts.setdefault(h2, 0) + 1
types2.add(h2)
# generate dataframe
logger.warn(
"%d of %d reads only in file 1" %
(len(hitmap), readMatchCount))
logger.warn(
"%d of %d reads only in file 2" %
(len(readsOnlyIn2), readMatchCount))
return pandas.DataFrame(counts)
def crossTabulateHits(hitFile1,
hitFile2,
skipHeader1=True,
skipHeader2=True,
readCol1=0,
readCol2=0,
hitCol1=-1,
hitCol2=-1,
hitSep1=None,
hitSep2=None):
"""
Given two hit tables, cross reference reads to build a 2D DataFrame
of hit counts
"""
# get a map from reads to hits in file 1
hitmap = tupleIteratorToMap(
parseHits(hitFile1, readCol1, hitCol1, skipHeader1, hitSep1))
# get an iterator over hits in file 2
crossHits = parseHits(hitFile2, readCol2, hitCol2, skipHeader2, hitSep2)
counts = {}
types2 = set()
readsOnlyIn2 = []
readMatchCount = 0
for read, hits in crossHits:
try:
hits2 = hitmap.pop(read)
readMatchCount += 1
except KeyError:
readsOnlyIn2.append(read)
continue
for h2 in hits2:
for h1 in hits:
h1counts = counts.setdefault(h1, {})
h1counts[h2] = h1counts.setdefault(h2, 0) + 1
types2.add(h2)
# generate dataframe
logger.warn("%d of %d reads only in file 1" %
(len(hitmap), readMatchCount))
logger.warn("%d of %d reads only in file 2" %
(len(readsOnlyIn2), readMatchCount))
return pandas.DataFrame(counts)
def main():
""" set up the command line interface """
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("-1", "--input_file_1",
default=None,
type=argparse.FileType('r'),
metavar=("INPUT_TABLE_1"),
help="Input table 1")
parser.add_argument("-2", "--input_file_2",
default=None,
type=argparse.FileType('r'),
metavar=("INPUT_TABLE_2"),
help="Input table 2")
parser.add_argument("-m", "--multiplier",
default=None,
metavar=("MULTIPLIER_TABLE"),
help=("Table of values to multiply each sequence. "
"EG assembly coverages."))
parser.add_argument("-T", "--total_reads",
default=0,
metavar="TOTAL_READS",
type=int,
help="Total number of reads to expect. (This allows "
"the reporting of unknown read count)")
parser.add_argument(
"-o",
"--outfile",
dest="outfile",
type=argparse.FileType('w'),
default=sys.stdout,
metavar="OUTFILE",
help="Write count table to OUTFILE. (Defaults to STDOUT")
parser.add_argument(
"-L",
"--long_output",
default=False,
action="store_true",
help="Print one number per row (prefixed by two keys) instead "
"of a table with one seet of keys as column names and one "
"set as row names.")
parser.add_argument(
"-H",
"--hitCol1",
dest="hitCol1",
type=int,
default=-1,
help="Index (starting at 0) of column in file 1 with hit name, -1 "
"is default meaning all columns that are not the read name are "
"hit names.",
metavar="HITCOL")
parser.add_argument(
"-I",
"--hitCol2",
dest="hitCol2",
type=int,
default=-1,
help="Index (starting at 0) of column in file 2 with hit name, -1 "
"is default meaning all columns that are not the read name "
"are hit names.",
metavar="HITCOL")
parser.add_argument(
"-S",
"--skipFirstRow",
action="store_true",
default=False,
help="hit tables have a header row which needs to be skipped")
add_universal_arguments(parser)
arguments = parser.parse_args()
setup_logging(arguments)
if arguments.input_file_1 is None or arguments.input_file_2 is None:
parser.error("Please supply two input files")
logging.info("reading hits from %s", arguments.input_file_1.name)
hits1 = parseHits(arguments.input_file_1,
0,
arguments.hitCol1,
arguments.skipFirstRow,
None)
logging.info("reading hits from %s", arguments.input_file_2.name)
hits2 = parseHits(arguments.input_file_2,
0,
arguments.hitCol2,
arguments.skipFirstRow,
None)
hits1 = tupleIteratorToMap(hits1)
hits2 = tupleIteratorToMap(hits2)
if arguments.multiplier is not None:
multipliers = parseMapFile(arguments.multiplier, valueType=float)
else:
multipliers = None
logging.info("counting hits")
(table, cols) = combine_counts(hits1, hits2, multipliers,
total_reads=arguments.total_reads)
# print out hit table
logging.info("printing table to " + arguments.outfile.name)
print_table(arguments.outfile, table, cols,
is_multiplied=multipliers is not None,
long_output=arguments.long_output)
def main():
""" set up the command line interface """
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("-1", "--input_file_1",
default=None,
type=argparse.FileType('r'),
metavar=("INPUT_TABLE_1"),
help="Input table 1")
parser.add_argument("-2", "--input_file_2",
default=None,
type=argparse.FileType('r'),
metavar=("INPUT_TABLE_2"),
help="Input table 2")
parser.add_argument("-m", "--multiplier",
default=None,
metavar=("MULTIPLIER_TABLE"),
help=("Table of values to multiply each sequence. "
"EG assembly coverages."))
parser.add_argument("-T", "--total_reads",
default=0,
metavar="TOTAL_READS",
type=int,
help="Total number of reads to expect. (This allows "
"the reporting of unknown read count)")
parser.add_argument(
"-o",
"--outfile",
dest="outfile",
type=argparse.FileType('w'),
default=sys.stdout,
metavar="OUTFILE",
help="Write count table to OUTFILE. (Defaults to STDOUT")
parser.add_argument(
"-L",
"--long_output",
default=False,
action="store_true",
help="Print one number per row (prefixed by two keys) instead "
"of a table with one seet of keys as column names and one "
"set as row names.")
parser.add_argument(
"-H",
"--hitCol1",
dest="hitCol1",
type=int,
default=-1,
help="Index (starting at 0) of column in file 1 with hit name, -1 "
"is default meaning all columns that are not the read name are "
"hit names.",
metavar="HITCOL")
parser.add_argument(
"-I",
"--hitCol2",
dest="hitCol2",
type=int,
default=-
1,
help="Index (starting at 0) of column in file 2 with hit name, -1 "
"is default meaning all columns that are not the read name "
"are hit names.",
metavar="HITCOL")
parser.add_argument(
"-S",
"--skipFirstRow",
action="store_true",
default=False,
help="hit tables have a header row which needs to be skipped")
add_universal_arguments(parser)
arguments = parser.parse_args()
setup_logging(arguments)
if arguments.input_file_1 is None or arguments.input_file_2 is None:
parser.error("Please supply two input files")
logging.info("reading hits from %s", arguments.input_file_1.name)
hits1 = parseHits(arguments.input_file_1,
0,
arguments.hitCol1,
arguments.skipFirstRow,
None)
logging.info("reading hits from %s", arguments.input_file_2.name)
hits2 = parseHits(arguments.input_file_2,
0,
arguments.hitCol2,
arguments.skipFirstRow,
None)
hits1 = tupleIteratorToMap(hits1)
hits2 = tupleIteratorToMap(hits2)
if arguments.multiplier is not None:
multipliers = parseMapFile(arguments.multiplier, valueType=float)
else:
multipliers = None
logging.info("counting hits")
(table, cols) = combine_counts(hits1, hits2, multipliers,
total_reads=arguments.total_reads)
# print out hit table
logging.info("printing table to " + arguments.outfile.name)
print_table(arguments.outfile, table, cols,
is_multiplied=multipliers is not None,
long_output=arguments.long_output)