Resource abstraction
Context
Currently sequence files are stored on S1 and accessible for download via an URL. The issue is the contextual data does not explicitly link to the sequence data or any other data source we may add in the future (by running some analysis ourselves for instance).
The solution is simply to create an extra attribute for each resource (one for .fa, one for .fna, ...) containing the URL of the resource. This way, the UI has direct access to the URL without need to compute it (in a non-configurable manner). This URL would be automatically assigned using the attribute's template.
Note: so that it is possible to change the location of the resources without need to modify all the entries in the database, the URL should be a (custom) CIRI.
To do
-
Modify pipeline so that .fa, .fna and .faa are zipped together (and do the zipping manually on S1!) -
Create a custom CIRI ( sfb_covid19_seq:<ID>
-->/srv/mar/SarsCovid19DB/genomes/<ID>/<ID>.tgz
) -
Create one attribute:
type
: use resource:sequence, ...
template
: use {acc:genbank@raw_id} (whatever the template should be) -
By running touch
, the new attributes should be created everywhere -
Remove the hard-coded Download sequences menuItem in record view -
Replace it with a generic a menuItem Download resources
having one option for eachresource:XXX
found in the config -
Download context should just be renamed. Download data/record/entry
for instance. -
Make graph use a defined set of types for the X type instead of .nonEmpty
(Set("cat", "num", "date")
)
Edited by mma227