Feature #3639: Create DataPackage from ID of metadata or data objects - DataONE R Client - DataONE Tasks

Feature #3639

Create DataPackage from ID of metadata or data objects

Added by Matthew Jones over 11 years ago. Updated about 10 years ago.

Status:

Closed

Priority:

Normal

Assignee:

Target version:

DataONE R Client 1.1.0

Start date:

2013-03-05

Due date:

2014-09-10

% Done:

100%

Milestone:

None

Story Points:

Sprint:

Related issues

History

#1 Updated by Matthew Jones over 11 years ago

Due date set to 2013-03-05
Start date set to 2013-03-05
Target version set to DataONE R Client 1.1.0

#2 Updated by Matthew Jones over 11 years ago

Currently DataPackage instances can be created with the ID of the package of interest. However, users often have the ID of a metadata object or data object -- provide a methods in D1Client to get DataPackage given the ID of one of its components. The difficult part of this is to determine which package to instantiate when an object is a member of multiple packages. One common case will be when there are multiple versions of a package due to edits in other files in the package; in this case, the best algorithm is probably to return the most recent revision in the package obsolescence chain that contains the id of interest. Other cases where an object is contained in multiple packages that are not in the same versioning chain are harder to decide on what to return.

#3 Updated by Rob Nahf over 11 years ago

Due date changed from 2013-03-05 to 2013-03-06

Yes, unfortunately the number of packages that a data / metadata object can be part of only increases over time, and membership is out of the control of the object owner, which further complicates what is meant by the "latest package" for a given object. So, we'd have to apply some limits so that the package that gets returned is easily understood and expected.

I think that if we can rely on the obsoletes chain of resource maps to meaningfully define what replaces what, we can move forward. Users could decide whether they want the latest package containing that id, or the latest package containing that id or it's replacement. In the latter case, I'm not sure what to do if the object is archived instead of replaced.

I think for the first, we follow the obsoletes chain of the object to the latest, then find it's package with a solr query, and for the second, we do the solr query, then follow the resourceMap's obsoletes chain to the latest.

#4 Updated by Rob Nahf over 11 years ago

fyi, in anticipation of this need, I put in place a client-side method in d1_libclient_java to build an ObsoletesChain for a given object. D1Client.listUpdateHistory(Identifier pid).

Ben relates that he has tried a client-side implementation for Morpho and it is slow, so a server-side impl is probably needed. But in the meantime, what's already in libclient v1.2.0 (and included in dataonelibs project) should get the job done.