[Biojava-l] blast xml parser

Simon Foote simon.foote@nrc.ca
Mon, 11 Jun 2001 10:33:05 -0400


--------------040601070205040202070208
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Thought I'd jump in here as I've found the blast xml parser quite useful 
and have created a partial blast object backend to it.  It was one of 
those quick and dirty implementations as I needed it quickly, so the 
code isn't of biojava quality at the moment, but I plan to fix it up 
once I get more time.  I actually based the blast object on a refinement 
of the pharmatools blast object model, but unfortunately I didn't 
include all the aspects of the blast report as I only needed certain 
portions at the time.

I've found the blast object quite useful as I can add filters and 
generate custom output (html, SQL, or whatever) separate from the 
parsing and the object creation.

Simon Foote

-- 
Bioinformatics Specialist
Institute for Biological Sciences
National Research Council of Canada
[T] 613-991-4342  [F] 613-952-9092

 

David Huen wrote:

>On Mon, 11 Jun 2001, Ewan Birney wrote:
>
>>On Mon, 11 Jun 2001, Wiepert, Mathieu wrote:
>>
>>>You know, you are more than welcome to add any functionality you desire.
>>>You have made a good point (as supported by Ewan), and besides that, no one
>>>is going to stop you from adding functionality to the code base.  This is
>>>open source code, so be open and code...
>>>
>>Hear hear.
>>
>Though do try to maintain a separation between the xml parsing and the
>backend object generation.  I have found this separation VERY useful
>recently when writing an XML parser.  My GAME frontend feeds straight into
>existing SeqIOListener backends allowing me to immediately instantiate
>biojava objects.  I was pleasantly surprised how fast and easy that
>was. 300 KB of sequence plus annotations in a blink. Later, I can always
>switch to direct entry to an SQL database without instantiating any
>biojava objects in the process merely with a new backend.
>
>But as for coding, I'd make a go of a blast object-making backend except
>I've got plenty of windmills of my own to tilt at just now [large bribes
>in brown paper bags can sway the priority... ;-) e.g., some hardware
>to biojava org to run an experimental DAS server would do nicely...]. Do
>look at the StAX framework of Thomas Down's. It's ever so nice for what
>you propose.
>
>Facetiously yours,
>David Huen, Dept. of Genetics, Univ. of Cambridge, U.K.
>P.S. Apologies for the commercial plug... I tend to be on the lookout for
>innovative and possibly ridiculous opportunities to fund open source
>developments.
>
>_______________________________________________
>Biojava-l mailing list  -  Biojava-l@biojava.org
>http://biojava.org/mailman/listinfo/biojava-l
>



--------------040601070205040202070208
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<html>
<head>
</head>
<body>
Thought I'd jump in here as I've found the blast xml parser quite useful
and have created a partial blast object backend to it. &nbsp;It was one of those
quick and dirty implementations as I needed it quickly, so the code isn't
of biojava quality at the moment, but I plan to fix it up once I get more
time. &nbsp;I actually based the blast object on a refinement of the pharmatools
blast object model, but unfortunately I didn't include all the aspects of
the blast report as I only needed certain portions at the time.<br>
<br>
I've found the blast object quite useful as I can add filters and generate
custom output (html, SQL, or whatever) separate from the parsing and the
object creation.<br>
<br>
Simon Foote<br>
<pre class="moz-signature" cols="$mailwrapcol">-- 
Bioinformatics Specialist
Institute for Biological Sciences
National Research Council of Canada
[T] 613-991-4342  [F] 613-952-9092</pre>
&nbsp;<br>
<br>
David Huen wrote:<br>
<blockquote type="cite" cite="mid:Pine.SOL.3.96.1010611142542.173A-100000@draco.cus.cam.ac.uk"><pre wrap="">On Mon, 11 Jun 2001, Ewan Birney wrote:<br><br></pre>
  <blockquote type="cite"><pre wrap="">On Mon, 11 Jun 2001, Wiepert, Mathieu wrote:<br><br></pre>
    <blockquote type="cite"><pre wrap="">You know, you are more than welcome to add any functionality you desire.<br>You have made a good point (as supported by Ewan), and besides that, no one<br>is going to stop you from adding functionality to the code base.  This is<br>open source code, so be open and code...<br></pre>
      </blockquote>
      <pre wrap="">Hear hear.<br><br></pre>
      </blockquote>
      <pre wrap=""><!---->Though do try to maintain a separation between the xml parsing and the<br>backend object generation.  I have found this separation VERY useful<br>recently when writing an XML parser.  My GAME frontend feeds straight into<br>existing SeqIOListener backends allowing me to immediately instantiate<br>biojava objects.  I was pleasantly surprised how fast and easy that<br>was. 300 KB of sequence plus annotations in a blink. Later, I can always<br>switch to direct entry to an SQL database without instantiating any<br>biojava objects in the process merely with a new backend.<br><br>But as for coding, I'd make a go of a blast object-making backend except<br>I've got plenty of windmills of my own to tilt at just now [large bribes<br>in brown paper bags can sway the priority... ;-) e.g., some hardware<br>to biojava org to run an experimental DAS server would do nicely...]. Do<br>look at the StAX framework of Thomas Down's. It's ever so nice for what<br>you propo!
se.<br><br>Facetiously yours,<br>David Huen, Dept. of Genetics, Univ. of Cambridge, U.K.<br>P.S. Apologies for the commercial plug... I tend to be on the lookout for<br>innovative and possibly ridiculous opportunities to fund open source<br>developments.<br><br>_______________________________________________<br>Biojava-l mailing list  -  <a class="moz-txt-link-abbreviated" href="mailto:Biojava-l@biojava.org">Biojava-l@biojava.org</a><br><a class="moz-txt-link-freetext" href="http://biojava.org/mailman/listinfo/biojava-l">http://biojava.org/mailman/listinfo/biojava-l</a><br></pre>
      </blockquote>
      <br>
      <pre class="moz-signature" cols="$mailwrapcol">
</pre>
      <br>
      </body>
      </html>

--------------040601070205040202070208--