view tools/blastxml_to_top_descr/blastxml_to_top_descr.py @ 11:98f8431dab44 draft

Uploaded v0.1.0, now also handles extended tabular BLAST output.

#!/usr/bin/env python
"""Convert a BLAST XML file to a top hits description table.

Takes three command line options, input BLAST XML filename, output tabular
BLAST filename, number of hits to collect the descriptions of.

Assumes the hits are pre-sorted, so "best" 3 hits gives first 3 hits.
"""
import os
import sys
import re
from optparse import OptionParser

if "-v" in sys.argv or "--version" in sys.argv:
    print "v0.1.0"
    sys.exit(0)

if sys.version_info[:2] >= ( 2, 5 ):
    import xml.etree.cElementTree as ElementTree
else:
    from galaxy import eggs
    import pkg_resources; pkg_resources.require( "elementtree" )
    from elementtree import ElementTree

def stop_err( msg ):
    sys.stderr.write("%s\n" % msg)
    sys.exit(1)

usage = """Use as follows:

$ blastxml_to_top_descr.py [-t 3] -o example.tabular input.xml

Or,

$ blastxml_to_top_descr.py [--topN 3] --output example.tabular input.xml

This will take the top 3 BLAST descriptions from the input BLAST XML file,
writing them to the specified output file in tabular format.
"""

parser = OptionParser(usage=usage)
parser.add_option("-t", "--topN", dest="topN", default=3,
                  help="Number of descriptions to collect (in order from file)")
parser.add_option("-o", "--output", dest="out_file", default=None,
                  help="Output filename for tabular file",
                  metavar="FILE")
parser.add_option("-f", "--format", dest="format", default="blastxml",
                  help="Input format (blastxml or tabular)")
parser.add_option("-q", "--qseqid", dest="qseqid", default="1",
                  help="Column for query 'qseqid' (for tabular input; default 1)")
parser.add_option("-s", "--sseqid", dest="sseqid", default="2",
                  help="Column for subject 'sseqid' (for tabular input; default 2)")
parser.add_option("-d", "--salltitles", dest="salltitles", default="25",
                  help="Column for descriptions 'salltitles' (for tabular input; default 25)")
(options, args) = parser.parse_args()

if len(sys.argv) == 4 and len(args) == 3 and not options.out_file:
    stop_err("""The API has changed, replace this:

$ python blastxml_to_top_descr.py input.xml output.tab 3

with:

$ python blastxml_to_top_descr.py -o output.tab -t 3 input.xml

Sorry.
""")

if not args:
    stop_err("Input filename missing, try -h")
if len(args) > 1:
    stop_err("Expects a single argument, one input filename")
in_file = args[0]
out_file = options.out_file
topN = options.topN

try:
    topN = int(topN)
except ValueError:
    stop_err("Number of hits  argument should be an integer (at least 1)")
if topN < 1:
    stop_err("Number of hits  argument should be an integer (at least 1)")

if not os.path.isfile(in_file):
    stop_err("Missing input file: %r" % in_file)


def get_column(value):
    """Convert column number on command line to Python index."""
    if value.startswith("c"):
        # Ignore c prefix, e.g. "c1" for "1"
        value = value[1:]
    try:
        col = int(value)
    except:
        stop_err("Expected an integer column number, not %r" % value)
    if col < 1:
        stop_err("Expect column numbers to be at least one, not %r" % value)
    return col - 1 # Python counting!

def tabular_hits(in_file, qseqid, sseqid, salltitles):
    """Parse key data from tabular BLAST output.

    Iterator returning tuples (qseqid, list_of_subject_description)
    """
    current_query = None
    current_hits = []
    with open(in_file) as input:
        for line in input:
            parts = line.rstrip("\n").split("\t")
            query = parts[qseqid]
            descr = "%s %s" % (parts[sseqid], parts[salltitles])
            if current_query is None:
                # First hit
                current_query = query
                current_hits = [descr]
            elif current_query == query:
                # Another hit
                current_hits.append(descr)
            else:
                # New query
                yield current_query, current_hits
                current_query = query
                current_hits = [descr]
    if current_query is not None:
        # Final query
        yield current_query, current_hits

def blastxml_hits(in_file):
    """Parse key data from BLAST XML output.

    Iterator returning tuples (qseqid, list_of_subject_description)
    """
    try:
        context = ElementTree.iterparse(in_file, events=("start", "end"))
    except:
        with open(in_file) as handle:
            header = handle.read(100)
        stop_err("Invalid data format in XML file %r which starts: %r" % (in_file, header))
    # turn it into an iterator
    context = iter(context)
    # get the root element
    try:
        event, root = context.next()
    except:
        with open(in_file) as handle:
            header = handle.read(100)
        stop_err("Unable to get root element from XML file %r which starts: %r" % (in_file, header))

    re_default_query_id = re.compile("^Query_\d+$")
    assert re_default_query_id.match("Query_101")
    assert not re_default_query_id.match("Query_101a")
    assert not re_default_query_id.match("MyQuery_101")
    re_default_subject_id = re.compile("^Subject_\d+$")
    assert re_default_subject_id.match("Subject_1")
    assert not re_default_subject_id.match("Subject_")
    assert not re_default_subject_id.match("Subject_12a")
    assert not re_default_subject_id.match("TheSubject_1")

    count = 0
    pos_count = 0
    current_query = None
    hit_descrs = []
    for event, elem in context:
        # for every <Iteration> tag
        if event == "end" and elem.tag == "Iteration":
            # Expecting either this, from BLAST 2.2.25+ using FASTA vs FASTA
            # <Iteration_query-ID>sp|Q9BS26|ERP44_HUMAN</Iteration_query-ID>
            # <Iteration_query-def>Endoplasmic reticulum resident protein 44 OS=Homo sapiens GN=ERP44 PE=1 SV=1</Iteration_query-def>
            # <Iteration_query-len>406</Iteration_query-len>
            # <Iteration_hits></Iteration_hits>
            #
            # Or, from BLAST 2.2.24+ run online
            # <Iteration_query-ID>Query_1</Iteration_query-ID>
            # <Iteration_query-def>Sample</Iteration_query-def>
            # <Iteration_query-len>516</Iteration_query-len>
            # <Iteration_hits>...
            qseqid = elem.findtext("Iteration_query-ID")
            if qseqid is None:
                stop_err("Missing <Iteration_query-ID> (could be really old BLAST XML data?)")
            if re_default_query_id.match(qseqid):
                #Place holder ID, take the first word of the query definition
                qseqid = elem.findtext("Iteration_query-def").split(None,1)[0]
            if current_query is None:
                # First hit
                current_query = qseqid
                hit_descrs = []
            elif current_query != qseqid:
                # New hit
                yield current_query, hit_descrs
                current_query = qseqid
                hit_descrs = []
            else:
                # Continuation of previous query
                # i.e. This BLAST XML did not use one <Iteration> per query
                # sys.stderr.write("Multiple <Iteration> blocks for %s\n" % qseqid)
                pass
            # for every <Hit> within <Iteration>
            for hit in elem.findall("Iteration_hits/Hit"):
                # Expecting either this,
                # <Hit_id>gi|3024260|sp|P56514.1|OPSD_BUFBU</Hit_id>
                # <Hit_def>RecName: Full=Rhodopsin</Hit_def>
                # <Hit_accession>P56514</Hit_accession>
                # or,
                # <Hit_id>Subject_1</Hit_id>
                # <Hit_def>gi|57163783|ref|NP_001009242.1| rhodopsin [Felis catus]</Hit_def>
                # <Hit_accession>Subject_1</Hit_accession>
                #
                #apparently depending on the parse_deflines switch
                sseqid = hit.findtext("Hit_id").split(None,1)[0]
                hit_def = sseqid + " " + hit.findtext("Hit_def")
                if re_default_subject_id.match(sseqid) \
                and sseqid == hit.findtext("Hit_accession"):
                    #Place holder ID, take the first word of the subject definition
                    hit_def = hit.findtext("Hit_def")
                    sseqid = hit_def.split(None,1)[0]
                assert hit_def not in hit_descrs
                hit_descrs.append(hit_def)
            # prevents ElementTree from growing large datastructure
            root.clear()
            elem.clear()
    if current_query is not None:
        # Final query
        yield current_query, hit_descrs

if options.format == "blastxml":
    hits = blastxml_hits(in_file)
elif options.format == "tabular":
    qseqid = get_column(options.qseqid)
    sseqid = get_column(options.sseqid)
    salltitles = get_column(options.salltitles)
    hits = tabular_hits(in_file, qseqid, sseqid, salltitles)
else:
    stop_err("Unsupported format: %r" % options.format)


def best_hits(descriptions, topN):
    if len(descriptions) < topN:
        return descriptions +  [""] * (topN - len(descriptions))
    else:
        return descriptions[:topN]

count = 0
if out_file is None:
    outfile = sys.stdout
else:
    outfile = open(out_file, 'w')
outfile.write("#Query\t%s\n" % "\t".join("BLAST hit %i" % (i+1) for i in range(topN)))
for query, descrs in hits:
    count += 1
    outfile.write("%s\t%s\n" % (query, "\t".join(best_hits(descrs, topN))))
if out_file is not None:
    outfile.close()
# Queries with no hits are not present in tabular BLAST output
print("%i queries with BLAST results" % count)