Fix DIAMOND hsp processing exception

[terrycojones]

The following exception occurred:

normalizeHSP error:
  queryLen: 99
  Original HSP:
    bits: 62.0
    btop: '2KP5EQ3ST1TVAG1IV1SK3FY6DANDVI1AGFYQTLPAM1AGLPHMFLYAPSTAQN1TGSTAPEV*QSKFLSTWLHD2WQSEFLPS*I1IPATPLPHEGFLGADNPHVD1*LFSMLYPIY1KLAQ5PGPYTSSTCW2PS1'
    expect: 1.8e-09
    frame: 2
    query_end: 301
    query_start: 2
    sbjct: 'IGPFARATQIGATAVGSVAKLFGYTNVPVIADIHGYTPMNGPMLASANIGTPVQKLTLDPKQELSIDPTLHGLANHDELSLPYLLQKESYFGYSTWSTSN'
    query: 'ATAGGTAAATTTGCTAGAGCTACAGAAATTGGGGCGAGTGCTACTGCATCGATAGCATCGCTATTTGGTTTTACCAATGTGCCTGTTATCGATAACGTGCATGCTTTTCAGCTGGCAAATGCACTCCATTTTTATCCGACTCAGATTACATCTGCCGAATAGAGCTTTTCTTGGCATCCTAAATGGAGTTTTCCATAAGACATTGCACCACCTGAATTTGGGGATCCAGTGGAATAATTTATGTATATTCTTAAAGCAAAAGAGAGTTATTTTCCTCCCACAAGTTGTTCCACTCCGAAT'
    sbjct_end: 437
    sbjct_start: 338
  Local variables:
    diamondTask: blastx
    initiallyIgnored: 1
    queryEnd: 100
    queryEndInSubject: 436
    queryGaps: 0
    queryLen: 99
    queryReversed: False
    queryStart: 0
    queryStartInSubject: 337
    subjectEnd: 437
    subjectGaps: 0
    subjectStart: 337
    unmatchedQueryLeft: 0
[Python backtrace deleted]
AssertionError: queryEndInSubject (436) < subjectEnd (437)

The error happens when matching YP_009337257.1 with query

@M01437:363:000000000-C7JVL:1:2119:19593:14570 2:N:0:GAATTCGT+CAGGACGT
TATAGGTAAATTTGCTAGAGCTACAGAAATTGGGGCGAGTGCTACTGCATCGATAGCATCGCTATTTGGTTTTACCAATGTGCCTGTTATCGATAACGTGCATGCTTTTCAGCTGGCAAATGCACTCCATTTTTATCCGACTCAGATTACATCTGCCGAATAGAGCTTTTCTTGGCATCCTAAATGGAGTTTTCCATAAGACATTGCACCACCTGAATTTGGGGATCCAGTGGAATAATTTATGTATATTCTTAAAGCAAAAGAGAGTTATTTTCCTCCCACAAGTTGTTCCACTCCGAAT
+
@-C@-6EA,ECEFCF9C,,E,@6@@ECCCE,ECE@F7BF@D,FC<<,C,CE7@E,6CEF86,9B@<FF,E@F7CFA<C<,6@<,9CF,C,,C<F=CFC6=,BEF<F,<F,99,,?,C85=FGGE?,,4,995,9,,,,++6++,,,9,,4,,,,44:+++++,,,33,63,3C=,,,,,,3@3,,36,7@D;,,3,35,7<,,,,,,,+,34*,,*35<22=*****0295*,;/**2+1++2+++++++1=+=9+++)29;A3(*);++;A7+1*+1)0);>5.69C7(18CA3*0/(,-

The DIAMOND output seems to be correct.

I'm working on it...

[terrycojones]

This is due to duplicate read ids, but it's not clear how. I thought it might be coming from DIAMOND (see bbuchfink/diamond#254) but it's not.