comparison tools/ncbi_blast_plus/blastxml_to_tabular.py @ 22:6f386c5dc4fb draft

v0.2.01 add -max_hsps, -use_sw_tback; lists args; internal updates

21:7538e2bfcd41

This script attempts to produce identical output to what BLAST+ would have done.
However, check this with "diff -b ..." since BLAST+ sometimes includes an extra
space character (probably a bug).
"""


import os
import re
import sys

from optparse import OptionParser

if "-v" in sys.argv or "--version" in sys.argv:
    print "v0.2.00"
    sys.exit(0)

if sys.version_info[:2] >= (2, 5):
    try:
        from xml.etree import cElementTree as ElementTree

                        try:
                            sallseqid = ";".join(name.split(None, 1)[0] for name in hit_def.split(" >"))
                            salltitles = "<>".join(name.split(None, 1)[1] for name in hit_def.split(" >"))
                        except IndexError as e:
                            sys.exit("Problem splitting multuple hits?\n%r\n--> %s" % (hit_def, e))
                        # print hit_def, "-->", sallseqid
                        positive = hsp.findtext("Hsp_positive")
                        ppos = "%0.2f" % (100 * float(positive) / float(length))
                        qframe = hsp.findtext("Hsp_query-frame")
                        sframe = hsp.findtext("Hsp_hit-frame")
                        if blast_program == "blastp":

                                       salltitles,
                                       ])
                    if cols:
                        # Only a subset of the columns are needed
                        values = [values[colnames.index(c)] for c in cols]
                    # print "\t".join(values)
                    output_handle.write("\t".join(values) + "\n")
            # prevents ElementTree from growing large datastructure
            root.clear()
            elem.clear()

22:6f386c5dc4fb

This script attempts to produce identical output to what BLAST+ would have done.
However, check this with "diff -b ..." since BLAST+ sometimes includes an extra
space character (probably a bug).
"""

from __future__ import print_function

import os
import re
import sys

from optparse import OptionParser

if "-v" in sys.argv or "--version" in sys.argv:
    print("v0.2.01")
    sys.exit(0)

if sys.version_info[:2] >= (2, 5):
    try:
        from xml.etree import cElementTree as ElementTree

                        try:
                            sallseqid = ";".join(name.split(None, 1)[0] for name in hit_def.split(" >"))
                            salltitles = "<>".join(name.split(None, 1)[1] for name in hit_def.split(" >"))
                        except IndexError as e:
                            sys.exit("Problem splitting multuple hits?\n%r\n--> %s" % (hit_def, e))
                        # print(hit_def, "-->", sallseqid)
                        positive = hsp.findtext("Hsp_positive")
                        ppos = "%0.2f" % (100 * float(positive) / float(length))
                        qframe = hsp.findtext("Hsp_query-frame")
                        sframe = hsp.findtext("Hsp_hit-frame")
                        if blast_program == "blastp":

                                       salltitles,
                                       ])
                    if cols:
                        # Only a subset of the columns are needed
                        values = [values[colnames.index(c)] for c in cols]
                    # print("\t".join(values))
                    output_handle.write("\t".join(values) + "\n")
            # prevents ElementTree from growing large datastructure
            root.clear()
            elem.clear()