[DAS2] DAS/2 writeback

[Helt,Gregg]

Thanks, that helped, the problem was on my end.  Looks like I was
POSTing but not making sure the data buffer was flushed out to the
server before I tried to read the server response.  I fixed that, now I
get a mapping document back.  I'm not sure how much effort to put into
parsing the mapping doc though - the next update of the spec is supposed
to change so that rather than a new mapping document type, the server
responds with the full feature XML of the created/updated features.
 
More progress - if I stick to the human genome (chr21), after the
writeback I'm able to retrieve the features via DAS/2 and visualize in
IGB.
 
            thanks again,
            Gregg
 
-----Original Message-----
From: allenday at gmail.com [mailto:allenday at gmail.com] On Behalf Of Allen
Day
Sent: Monday, June 26, 2006 3:04 PM
To: Helt,Gregg
Subject: Re: DAS/2 writeback
 
The content-type shouldn't matter.  I think I was submitting with
application/x-form-encoded, or something like that.

The CGI just reads whatever you sent it directly from STDIN though.
>From the error log it looks like you are not POSTing the document.
There is no web page there, you need to issue a POST request, XML doc as
the body, directly to that CGI.

-Allen


On 6/26/06, Helt,Gregg <Gregg_Helt at affymetrix.com> wrote:
That's my guess.
 
Also, I'm attempting to post variations of your XML example to
http://genomics.ctrl.ucla.edu/~allenday/cgi-bin/das2xml-parser/stable1.p
l
<http://genomics.ctrl.ucla.edu/%7Eallenday/cgi-bin/das2xml-parser/stable
1.pl> , 
but so far I've only gotten "500 Internal Server Error" responses back.
Can you tell what's happening?  Do I need to set the content-type for
the POST or some other header?
 
            gregg
 
-----Original Message----- 
From: allenday at gmail.com [mailto:allenday at gmail.com] On Behalf Of Allen
Day
Sent: Monday, June 26, 2006 2:40 PM
To: Helt,Gregg
Subject: Re: DAS/2 writeback
 
Ok, i will check into it.  Only showing human segments, is it?

-Allen
On 6/26/06, Helt,Gregg <Gregg_Helt at affymetrix.com> wrote:
But there's no entry in the segment query for chrI, which is the
chromosome you used in the the example XML you posted.  So I can't find
those annotations via IGB.
 
            gregg
 
-----Original Message----- 
From: allenday at gmail.com [mailto:allenday at gmail.com] On Behalf Of Allen
Day
Sent: Monday, June 26, 2006 2:11 PM
To: Helt,Gregg
Cc: DAS/2
Subject: Re: DAS/2 writeback
 
This datasource contains both yeast and human segments.  I set it up
this way so features can be written for either human or yeast, then
viewed at one  of:

http://das.biopackages.net/das/genome/yeast/S228C-writeback/feature 
http://das.biopackages.net/das/genome/human/17-writeback/feature

I thought this would be more useful so you can test viewing writeback
features alongside both "real" human and yeast features.

So you can just ignore the irrelevant segments, or if you'd prefer I can
delete one set of segments or the other.

-Allen
On 6/26/06, Helt,Gregg <Gregg_Helt at affymetrix.com> wrote:
Allen, thanks for getting the start of a writeback server up and
running!

I'm hoping to try writing back annotations later today.  However, I'm
having problems looking at the annotations in the writeback server via 
IGB.  It looks to me like the main issue is that
http://das.biopackages.net/das/genome/yeast/S228C-writeback/segment 
returns human chromosome ids in the uri attribute of the SEGMENT 
element, instead of yeast ids.  When IGB uses this to compose a query
like

http://das.biopackages.net/das/genome/yeast/S228C-writeback/feature?over

laps=chr1/10:15000;type=SO:centromere

it gets back an empty feature list.  But if I manually edit this to
replace "chr1" with "chrI",

http://das.biopackages.net/das/genome/yeast/S228C-writeback/feature?over

laps=chrI/10:15000;type=SO:centromere

I get back a list of feature that satisfies the query filters.

Any ideas?

        Thanks, 
        Gregg

> -----Original Message-----
> From: das2-bounces at lists.open-bio.org [mailto:das2-bounces at lists.open
-
> bio.org] On Behalf Of Allen Day
> Sent: Saturday, June 24, 2006 2:24 AM
> To: Chervitz, Steve
> Cc: DAS/2
> Subject: Re: [DAS2] Notes from the weekly DAS/2 teleconference, 19 Jun

> 2006
>
> You can see the features that are posted here:
>
> http://das.biopackages.net/das/genome/yeast/S228C-writeback/feature 
>
> It is fully compatible with the usual yeast source:
>
> http://das.biopackages.net/das/genome/yeast/S228C/feature
>
> All the usual feature filters apply.  The response at this URL is not
> cached
> to keep the content fresh, at the expense of ever-slower load times as
> written features accumulate.
> 
> -Allen
>
>
>
> On 6/24/06, Allen Day <allenday at ucla.edu> wrote:
> >
> > I have a temporary CGI set up to accept WRITEBACK documents: 
> >
> > http://genomics.ctrl.ucla.edu/~allenday/cgi-bin/das2xml-
<http://genomics.ctrl.ucla.edu/%7Eallenday/cgi-bin/das2xml-> 
> parser/stable1.pl< http://genomics.ctrl.ucla.edu/%7Eallenday/cgi-
<http://genomics.ctrl.ucla.edu/%7Eallenday/cgi-> 
> bin/das2xml-parser/stable1.pl>
> >
> > I have attached a das2xml document that POSTs cleanly for me using
the
> > lwp-request that is part of libwww-perl.  Please modify this 
document,
> post,
> > and let me know if anything breaks.
> >
> > This implementation accepts only new records.  It supports neither
> updates
> > nor deletes.  Furthermore, it only accepts new feature records.  It 
does
> not
> > support new type records, new region records, or any other type of
> record.
> >
> > Feature records may have 0 or more locations, 0 or more parents, 0
or
> more 
> > children, and 0 or more properties.  All parts/parents must be
present
> in
> > the document (no refering to existing features by URI), or it will
throw
> a
> > HTTP 500 error. 
> >
> > Next I will implement the update and delete support.  This should be
> > fairly straightforward, and may be doable over the weekend.
> >
> > -Allen
> >