[Biopython-dev] Bio/SearchIO/BlastIO/blast_xml.py output under Python 2.7.4

[Wibowo Arindrarto]

Hi Peter, everyone,

This is reproducible in my machine (which just happen to update its default
Python2.x to 2.7.4 actually). Looks like they changed the underlying SAX
XML writer (xml.sax.saxutils.XMLGenerator ~ which the SearchIO blast_xml
writer subclasses). The new write() method calls the
io.TextIOWrapper.write() function, which expects unicode only.

See the comparison here:
Python2.7.3 (
http://hg.python.org/cpython/file/70274d53c1dd/Lib/xml/sax/saxutils.py#l84)
vs Python2.7.4 (
http://hg.python.org/cpython/file/026ee0057e2d/Lib/xml/sax/saxutils.py#l109).
Also, here's the io module doc page for reference:
http://docs.python.org/2/library/io.html

I'll write up a Python <=2.7.3-compatible patch to fix this soon, if
everybody's ok with it :).

Cheers,
Bow


On Mon, Apr 8, 2013 at 5:28 PM, Peter Cock <p.j.a.cock at googlemail.com>wrote:

> Hi Bow,
>
> I've just updated the 64bit Linux buildslave from Python 2.7.3 to
> 2.7.4 and there
> is a new unicode failure with the BLAST XML writing code in SearchIO
> (below).
>
> Can you reproduce this? Keep in mind this could be a new bug in Python
> itself ;)
>
> Separately the are some BGZF issues, see also this thread:
> http://lists.open-bio.org/pipermail/biopython/2013-April/008490.html
>
> Thanks,
>
> Peter
>
> ======================================================================
> ERROR: test_write_multiple_from_blastxml
> (test_SearchIO_write.BlastXmlWriteCases)
> Test blast-xml writing from blast-xml, BLAST 2.2.26+, multiple queries
> (xml_2226_blastp_001.xml)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "test_SearchIO_write.py", line 55, in
> test_write_multiple_from_blastxml
>     self.parse_write_and_compare(source, self.fmt, self.out, self.fmt)
>   File "test_SearchIO_write.py", line 27, in parse_write_and_compare
>     SearchIO.write(source_qresults, out_file, out_format, **kwargs)
>   File
> "/home_local/buildslave/repositories/biopython/build/lib.linux-x86_64-2.7/Bio/SearchIO/__init__.py",
> line 610, in write
>     writer.write_file(qresults)
>   File
> "/home_local/buildslave/repositories/biopython/build/lib.linux-x86_64-2.7/Bio/SearchIO/BlastIO/blast_xml.py",
> line 695, in write_file
>     xml.startDocument()
>   File
> "/home_local/buildslave/repositories/biopython/build/lib.linux-x86_64-2.7/Bio/SearchIO/BlastIO/blast_xml.py",
> line 612, in startDocument
>     self.write('<?xml version="1.0"?>\n'
>   File "/home_local/buildslave/lib/python2.7/xml/sax/saxutils.py",
> line 103, in write
>     super(UnbufferedTextIOWrapper, self).write(s)
> TypeError: must be unicode, not str
>
> ======================================================================
> ERROR: test_write_single_from_blastxml
> (test_SearchIO_write.BlastXmlWriteCases)
> Test blast-xml writing from blast-xml, BLAST 2.2.26+, single query
> (xml_2226_blastp_004.xml)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "test_SearchIO_write.py", line 49, in
> test_write_single_from_blastxml
>     self.parse_write_and_compare(source, self.fmt, self.out, self.fmt)
>   File "test_SearchIO_write.py", line 27, in parse_write_and_compare
>     SearchIO.write(source_qresults, out_file, out_format, **kwargs)
>   File
> "/home_local/buildslave/repositories/biopython/build/lib.linux-x86_64-2.7/Bio/SearchIO/__init__.py",
> line 610, in write
>     writer.write_file(qresults)
>   File
> "/home_local/buildslave/repositories/biopython/build/lib.linux-x86_64-2.7/Bio/SearchIO/BlastIO/blast_xml.py",
> line 695, in write_file
>     xml.startDocument()
>   File
> "/home_local/buildslave/repositories/biopython/build/lib.linux-x86_64-2.7/Bio/SearchIO/BlastIO/blast_xml.py",
> line 612, in startDocument
>     self.write('<?xml version="1.0"?>\n'
>   File "/home_local/buildslave/lib/python2.7/xml/sax/saxutils.py",
> line 103, in write
>     super(UnbufferedTextIOWrapper, self).write(s)
> TypeError: must be unicode, not str
>