[BioPython] Changes in NCBI BLAST output format !!??

[Michiel De Hoon]

I've updated Biopython in CVS with this fix. See Bio/Blast/NCBIWWW.py
revision 1.41. Please let me know if you find any problems. Thanks for
finding this solution.

--Michiel.

Michiel de Hoon
Center for Computational Biology and Bioinformatics
Columbia University
1150 St Nicholas Avenue
New York, NY 10032



-----Original Message-----
From: Alexander A. Morgan [mailto:amorgan at mitre.org]
Sent: Wed 8/3/2005 2:48 PM
To: Michiel De Hoon
Subject: Re: [BioPython] Changes in NCBI BLAST output format !!??
 
Michiel:

I couldn't find anything wrong with it.  The Blast record objects seem 
to be correct and have the right alignments.  However, I don't know of a 
thorough way to test it.  In general the parser is pretty fragile though 
and will break for even the most minor changes in NCBI format, but it 
would be very challenging to try to make it more robust.

Thanks,

-Alex





Michiel De Hoon wrote:

>Do you happen to know if this change can break anything else in the Blast
>parser? From running Biopython's tests for Blast, it seems that this change
>is OK. On the other hand, I don't use Blast much myself, so I don't trust my
>own judgement in this matter.
>If making this change does not cause any new bugs, I'd be happy to include
it
>in CVS.
>
>--Michiel.
>
>
>Michiel de Hoon
>Center for Computational Biology and Bioinformatics
>Columbia University
>1150 St Nicholas Avenue
>New York, NY 10032
>
>
>
>-----Original Message-----
>From: biopython-bounces at portal.open-bio.org on behalf of Alexander A. Morgan
>Sent: Tue 8/2/2005 4:07 PM
>To: aurelie.bornot at free.fr
>Cc: biopython at biopython.org
>Subject: Re: [BioPython] Changes in NCBI BLAST output format !!??
> 
>Hello:
>    I've just run into the same problem, and I haven't seen a suggested 
>fix go by, so I apologize if this is redundant information, but it seems 
>that the files I've been getting from NCBI have a <p> removed from the 
>header between the "RID: " line and the "<p>Query" line, and it is just 
>a blank line now.  If you edit Bio.Blast.NCBIWWW to not look for the 
>"<p>", it seems to work okay.
> 
>class _Scanner:
>....
>   def _scan_header(self, uhandle, consumer):
>....
>
>    change:
>        attempt_read_and_call(uhandle, consumer.noevent, start='<p>')
>    to: 
>        attempt_read_and_call(uhandle, consumer.noevent)
>
>
>
>
>
>aurelie.bornot at free.fr wrote:
>
>  
>
>>Thank you very much Jessica !!!
>>
>>Unfortunately, I need a lot of thing in the BLAST reports.....
>>It will be difficult to do the same thing as you did....
>>
>>I will try to do something in the code of parser of Python.
>>But it will be difficult for me..
>>so if you or someone has advices !!!
>>
>>Thanks a lot again for your answer Jessica !
>>Aurélie
>>
>>
>>--------------
>>Aurelie BORNOT
>>MNHN
>>Paris
>>
>>
>>_______________________________________________
>>BioPython mailing list  -  BioPython at biopython.org
>>http://biopython.org/mailman/listinfo/biopython
>> 
>>
>>    
>>
>
>
>_______________________________________________
>BioPython mailing list  -  BioPython at biopython.org
>http://biopython.org/mailman/listinfo/biopython
>
>  
>