[DAS2] RE: Problem with parent/child features in biopackages server feature response

[Allen Day]

Okay, give it a shot now.  You uncovered a bug in my SQL query where the 
parent feature was outside your overlaps+inside ranges and wasn't being 
retrieved, leading to an error.  These two are now in my regression test 
suite.

-Allen


On Mon, 17 Oct 2005, Helt,Gregg wrote:

> Thanks for the quick fix!
> 
> Now I'm seeing another problem though.  Some feature queries with a
> combination of one "overlaps" and one "inside" filter are giving weird
> errors.  But some return correctly.
> 
> For example, 
> http://das.biopackages.net/das/genome/human/17/feature?overlaps=chr21/26
> 104961:26271647;inside=chr21/26104961:46976097;type=SO:mRNA
> returns correctly.
> 
> But
> http://das.biopackages.net/das/genome/human/17/feature?overlaps=chr21/25
> 854314:26027736;inside=chr21/0:26027736;type=SO:mRNA
> does not, instead returning these headers:
> 	HTTP/1.1 200 OK
> 	Date: Mon, 17 Oct 2005 22:40:54 GMT
> 	Server: Apache/2.0.51 (Fedora)
> 	X-DAS-Version: DAS/2.0
> 	X-DAS-Server: GMOD/0.0
> 	X-DAS-Content-Type: text/x-das-feature+xml
> 	X-DAS-Status-Details: Died at
> /usr/lib/perl5/site_perl/5.8.3/Package/Base/Devel.pm line 425.
> 
> and then what appears to be more headers but is designated as content
> (so I'm guessing a header-terminating blank line got inserted before
> these):
> 	X-DAS-Status: 500
> 	X-DAS-Status: 200
> 	Content-Length: 0
> 	Keep-Alive: timeout=15, max=100
> 	Connection: Keep-Alive
> 	Content-Type: text/xml
> 
> The combination of an "overlaps" and "inside" filter is an important
> part of IGB's client-side query optimizations.  I can turn this
> optimization off for now, but I'm hoping you can diagnose and fix the
> problem.
> 
> 	Thanks again,
> 	Gregg
> 
> > -----Original Message-----
> > From: Allen Day [mailto:allenday at ucla.edu]
> > Sent: Monday, October 17, 2005 2:50 PM
> > To: Helt,Gregg
> > Cc: das2 at portal.open-bio.org
> > Subject: Re: Problem with parent/child features in biopackages server
> > feature response
> > 
> > Hi Gregg,
> > 
> > There was a logic inversion in my code.  PARENT/PART relationships
> should
> > now be restored.
> > 
> > Andrew, I'd also like to know if there is some code already written
> that
> > can be written into my regression tests.
> > 
> > I understand that it's really irritating to have non-fatal errors of
> > different types continuously appearing.  Maybe I should spend some
> time
> > beefing up the regression test suite to do things like diff feature
> graphs
> > and make sure they are identical?  Something to discuss on the
> conference
> > call...
> > 
> > -Allen
> > 
> > On Mon, 17 Oct 2005, Helt,Gregg wrote:
> > 
> > > I finally got back to testing DAS/2 feature requests/responses in my
> IGB
> > > client.  I'm seeing a new problem in responses from the biopackages
> > > server, there are no <CHILD> or <PARENT> elements for features at
> all.
> > >
> > > See for example the returned XML from my standard test query:
> > >
> http://das.biopackages.net/das/genome/human/17/feature?overlaps=chr21/26
> > > 027736:26068042;type=SO:mRNA
> > >
> > > ...
> > > <FEATURE id="feature/BC001178" type="type/SO:mRNA" name="BC001178">
> > > 	<LOC pos="region/chr21/26018728:26029854:-1"/>
> > > 	...
> > > </FEATURE>
> > > <FEATURE id="feature/auto316726" type="type/SO:three_prime_UTR"
> > > name="auto316726">
> > > 	<LOC pos="region/chr21/26018728:26018820:-1"/>
> > > </FEATURE>
> > > ...
> > >
> > > But I know that "auto316726" should be a child of "BC001178".
> > >
> > > Andrew, would your DAS/2 validator catch problems like this?
> > >
> > >
> > > 	Thanks,
> > > 	Gregg
> > >
> > >
>