[DAS2] DAS/2 GET requests

[Andrew Dalke]

Nomi:
> i have gotten my apollo das/2 adapter to read das2xml, and now i want 
> to
> write the GET part.  is the GET protocol now stable, and is there a
> reference server implementing it?  i've been using the affy server
> because i couldn't get data from the biodas server (only metadata).
>
> if GET is not yet stable, can i have an estimate of when it will be?

I was out the last couple days for a local bioinformatics meeting
and not checking email.  I think I caught you just before Friday 5pm.


There are no servers which support it.  The spec is still being
written, including for the XML formats.

Today I worked on the stylesheet data. That's now in CVS.
My working data is in the

   das/das2/scratch

"scratch" because I'm likely to move things around and I don't
like CVS's filename delete support.

These formats are support with a nearly complete RNC

== "sources.rnc" is used for "application/x-das-sources+xml"

I expect at most minor changes (probably in the "ASSEMBLY"
field and perhaps an element or attribute name.


== "residues.rnc" is used for "application/x-das-residues+xml"
(I think this should be renamed "segments" since the it's hard
to make a plural of "residues".)

I expect no changes except perhaps to add extensibility.

  == "types.rnc" is used for "application/x-das-types+xml"

I mostly have questions about how to refer to the ontology
(which ontology? what naming scheme?)

I also don't understand how the "source" attribute is used.
The one example uses "curated".  Are there other properties?

   == "feature.rnc" is used for "application/x-das-features+xml"

At most minor changes, probably a few attributes added
when people start use the features.



There will be some changes for the style information.  These
will come up during the sprint.  I think they can wait.


I have XML examples for the different data types.  These are
made by hand and are scientifically non-realistic.  They are
of the forms:

features*.xml
residues*.xml
sources*.xml
styles*.xml
types*.xml

I worked on a test script to make sure these documents were
valid.  It's a new (but simple) language to load an XML
file and verify it against the given type, then ask it
for some simple properties.  The idea is that I or others
can reimplement the script commands in another language
and/or data structure and verify that the loaded data is
processed correctly.  (For example, correct handling of Unicode)

The interpreter is named "verify_examples.py" (in the scratch
area).  The commands are of the form "*_cmds.txt".  For example,
here is part of "sources_cmds.txt"


# Simple maintainer, one genome, no versions
load SOURCES "sources2.xml"
= Ok
count das:MAINTAINER
= 1
attrib das:MAINTAINER name
= Yoyodyne DNA Systems
count das:SOURCE
= 1



What remains is:
   - the locking information (I'm waiting for the sprint for this)
   - working more on the formal writeup (notes are
        in ~das/das2/new_spec.txt)
   - updating the query interface

NOTE: the query interface is nearly unchanged.  Well, except
for the "regions" and "properties" subtrees which no longer exist.
The feature filters should be unchanged.

					Andrew
					dalke at dalkescientific.com