The URI should be resolvable against the base URI. In
the biopackages DAS/2 some of the URIs are actually
created by slicing and dicing strings. This should
really not be done.
This is really talking about the filters in the DAS/2
server. Currently, the filters use ids. But these really
should be fully resolvable URIs i.e.
type=http://foo.com/genome...
rather than
type=MA:0000345
Apparently the following filters should be converted:
type
sequence/region (all location based)
segment
overlaps
inside
exact
A first pass on this would extract the correct value from a
das.biopackages.net URL and then use that normally. In the
future, it won't be possible to make assumptions about where
the URIs point and what they resolve to.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
After talking with Gregg and Allen about this one there are
two "solutions". The first (not so flexible) solution is to
mangle the URI and clip off the known ending and use that to
produce a response by the server. The problem with this is
it assumes the URI formats are all the same and that the
local DAS/2 server is using the same build of the genome(s).
The better solution is to use a mapping file based on
Lincoln's official URIs in the wiki. That way everyone
would use the same URI passed to, say, an overlaps query and
the DAS/2 server would know how to map that URI to an
internal ID used to locate and produce a response.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Logged In: YES
user_id=587499
Some notes:
This is really talking about the filters in the DAS/2
server. Currently, the filters use ids. But these really
should be fully resolvable URIs i.e.
type=http://foo.com/genome...
rather than
type=MA:0000345
Apparently the following filters should be converted:
type
sequence/region (all location based)
segment
overlaps
inside
exact
A first pass on this would extract the correct value from a
das.biopackages.net URL and then use that normally. In the
future, it won't be possible to make assumptions about where
the URIs point and what they resolve to.
Logged In: YES
user_id=587499
Some notes:
After talking with Gregg and Allen about this one there are
two "solutions". The first (not so flexible) solution is to
mangle the URI and clip off the known ending and use that to
produce a response by the server. The problem with this is
it assumes the URI formats are all the same and that the
local DAS/2 server is using the same build of the genome(s).
The better solution is to use a mapping file based on
Lincoln's official URIs in the wiki. That way everyone
would use the same URI passed to, say, an overlaps query and
the DAS/2 server would know how to map that URI to an
internal ID used to locate and produce a response.
Logged In: YES
user_id=587499
OK... this is "fixed".
The following need to use URIs:
* xid
* type
* segment
* coordinates
These work but xid isn't supported in our server (but in the
spec and needs to be implemented).