This plan makes sense to me. I believe we need to update the hibernate definition to recursively crawl community2community. I am new to hibernate, but I presume that this is possible.
Terry On Fri, May 20, 2016 at 10:25 AM, Tim Donohue <tdono...@duraspace.org> wrote: > After some more thought on this, I agree with Monika. > > Since we have no known use cases for having a Collection appear in > multiple Communities, let's leave the REST API as-is in 6.0. It doesn't > need to surface this "feature" of the underlying Java API, until we > actually have a reason to do so. > > I also agree that eventually we might want to just simplify these REST API > params to "parent" and "parentList". But, that can be a future task (not > for 6.0). > > So, Terry, that should answer your #2. Don't bother updating the REST API > for this. As for #1, I think we just need to make "parentCommunityList" > act similar to 5.x in the 6.x codebase, which should solve > https://jira.duraspace.org/browse/DS-2766 > > - Tim > On 5/18/2016 5:11 PM, Monika Mevenkamp wrote: > > Well I guess _ i rephrase - I’d vote for hiding :-) > > On another note - would it make sense to get away from names like > parentCommunity > parentCollection > > in favor of parent / parentList > > again looking ahead to the bright future where collections and communities > become the same > > Monika > > — > Monika Mevenkamp > Digital Repository Infrastructure Developer > Princeton University > Phone: 609-258-4161 > Skype: mo-meven > > > > On May 18, 2016, at 5:58 PM, Tim Donohue < <tdono...@duraspace.org> > tdono...@duraspace.org> wrote: > > Hi Monika, > > Those are good questions. To clarify, I'm not recommending we add a new > "map collections" feature. I was just pointing out to Terry, that *at the > Java API level*, DSpace has supported the idea of a Collection having > multiple parents for some time. That is not a change, it's just how the > Java API has worked. However, at the UI level, this feature has never been > implemented because we never had a clear use case for it (and I'm not > exactly sure myself why the API was built this way to begin with). > > When it comes to the REST API, we'd then have a decision. Do we let it act > more like the Java API or do we just "hide" this feature of the Java API > and make the REST API look more like the current UIs. I think this is the > basics of what Terry is asking in his #2. I'm not sold on the fact that we > need to let the REST API even support this feature (it's definitely not a > high priority). > > I think the real issue here is just finding a solution to > https://jira.duraspace.org/browse/DS-2766, which is more about just > figuring out what "parentCommunityList" was meant to do. > > - Tim > > On 5/18/2016 3:08 PM, Monika Mevenkamp wrote: > > Right now DSpace has a conceptually clean structure, an almost perfect > hierarchy, except where items are mapped to multiple collections. So for > example: it is clear what it means to show a collection in the UI. For a > mapped item that is not necessarily true. Image styling different > collections in different ways in a future flexible UI - what style should > be picked for a mapped item ? In other words - mapping things - means > moving away from a hierarchy - means life will be more complicated. > > My question here: Mapping Collection to multiple communities — Is it > worth it ? > > What is the use case for mapping collections to multiple communities ? > > Can the same be achieved with solr and ‘virtual’ collections which could > be defined as whatever matches some kind of query ? > > What is the implication in terms of getting rid of the distinction between > communities and collections ? > That is a feature I am looking forward to: we have several communities > with a single collection with few items. > > Where does that position the DSPACE model in relation to the Portland Data > Model ? > > My gut tells me to not to map collections. I don’t have a good overview of > the wider DSPace world though. > But it would be worthwhile discussing before hacking … > > > Monika > — > Monika Mevenkamp > Digital Repository Infrastructure Developer > Princeton University > Phone: 609-258-4161 > Skype: mo-meven > > > > On May 18, 2016, at 12:33 PM, Terry Brady <terry.br...@georgetown.edu> > wrote: > > Tim, > > Thanks for the explanation. > > To implement #1, I recommend that we add the necessary hibernate > configuration logic to recursively traverse community2community. Who could > advise on that change? > > To implement #2, we would probably want to change the method return type > to return a list of lists: one list for each immediate community parent. > > Terry > > On Wed, May 18, 2016 at 7:39 AM, Tim Donohue < <tdono...@duraspace.org> > tdono...@duraspace.org> wrote: > >> Hi Terry, >> >> Some quick answers... >> >> I'm not sure of the correct answer to #1 myself (Peter Dietz, did you >> create this behavior?). Since it's called a "List" though, it sounds to me >> like the behavior in DSpace 5.x is more correct (that the intension may be >> to locate the full location of the collection, similar to breadcrumbs). >> >> As for #2, the API / database has always supported the "idea" of a >> Collection existing in multiple locations (i.e. a "linked" Collection). >> However, we've never found a reason to fully build this out with unit >> tests, etc. So, it's a partially created concept, and it's never really >> been proven to "work". In addition, the UIs have never supported this >> idea...at the UI level, a Collection only ever belongs to a single >> Community. >> >> So, essentially, #2 would/should be treated as a new feature. The >> groundwork is there in the API / database layer. But, it has never been >> fully built out. It'd need a volunteer to dig in and implement this concept >> so that it's actually usable in the UIs. It also may require some >> refactoring at the REST API layer, as you've noted. >> >> - Tim >> On 5/17/2016 1:03 PM, Terry Brady wrote: >> >> I am investigating the following ticket: >> <https://jira.duraspace.org/browse/DS-2766> >> https://jira.duraspace.org/browse/DS-2766 >> >> Here are my questions. >> 1. In the collections endpoint, of the REST api, what is the intended >> behavior of expand=parentCommunityList? >> 2. In DSpace, can a collection exist in multiple communities? >> >> Question 1. expand=parentCommunityList >> >> In DSpace5x, the expand=parentCommunityList option returns the full >> hierarchy of owing communities for a collection (all the way to the top >> level community). >> >> In DSpace 6x, this option returns the subcommunity that is an immediate >> parent of the collection. >> >> Assuming that this behavior is incorrect in DSpace 6, the following >> configuration will need to be updated to follow the community2community >> hierarchy: >> <https://github.com/DSpace/DSpace/blob/master/dspace-api/src/main/java/org/dspace/content/Collection.java#L81-L86> >> https://github.com/DSpace/DSpace/blob/master/dspace-api/src/main/java/org/dspace/content/Collection.java#L81-L86 >> >> Question 2: Can a collection exist in multiple communities? >> >> The following method in DSpace 5x could imply that a collection can exist >> in multiple communities: >> <https://github.com/DSpace/DSpace/blob/dspace-5_x/dspace-api/src/main/java/org/dspace/content/Collection.java#L1402> >> https://github.com/DSpace/DSpace/blob/dspace-5_x/dspace-api/src/main/java/org/dspace/content/Collection.java#L1402 >> >> But, the REST API implies that there is a single "parentCommunity". >> >> <https://github.com/DSpace/DSpace/blob/dspace-5_x/dspace-rest/src/main/java/org/dspace/rest/common/Collection.java#L136> >> https://github.com/DSpace/DSpace/blob/dspace-5_x/dspace-rest/src/main/java/org/dspace/rest/common/Collection.java#L136 >> >> and a separate parentCommunityList >> >> <https://github.com/DSpace/DSpace/blob/dspace-5_x/dspace-rest/src/main/java/org/dspace/rest/common/Collection.java#L156> >> https://github.com/DSpace/DSpace/blob/dspace-5_x/dspace-rest/src/main/java/org/dspace/rest/common/Collection.java#L156 >> >> What is the intended set of relationships to be supported? >> >> If a collection can exist in multiple communities, how should the REST >> api respond to "expand=parentCommunityList" option? >> >> If a collection can exist in multiple communities, how can this >> relationship be created in our UIs? >> >> Terry >> >> -- >> Terry Brady >> Applications Programmer Analyst >> Georgetown University Library Information Technology >> <https://www.library.georgetown.edu/lit/code> >> https://www.library.georgetown.edu/lit/code >> 425-298-5498 (Seattle, WA) >> -- >> You received this message because you are subscribed to the Google Groups >> "DSpace Technical Support" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to <dspace-tech+unsubscr...@googlegroups.com> >> dspace-tech+unsubscr...@googlegroups.com. >> To post to this group, send email to dspace-tech@googlegroups.com. >> Visit this group at <https://groups.google.com/group/dspace-tech> >> https://groups.google.com/group/dspace-tech. >> For more options, visit <https://groups.google.com/d/optout> >> https://groups.google.com/d/optout. >> >> >> -- >> Tim Donohue >> Technical Lead for DSpace & DSpaceDirectDuraSpace.org >> <http://duraspace.org/> | DSpace.org <http://dspace.org/> | DSpaceDirect.org >> <http://dspacedirect.org/> >> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "DSpace Technical Support" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to <dspace-tech+unsubscr...@googlegroups.com> >> dspace-tech+unsubscr...@googlegroups.com. >> To post to this group, send email to dspace-tech@googlegroups.com. >> Visit this group at <https://groups.google.com/group/dspace-tech> >> https://groups.google.com/group/dspace-tech. >> For more options, visit <https://groups.google.com/d/optout> >> https://groups.google.com/d/optout. >> > > > > -- > Terry Brady > Applications Programmer Analyst > Georgetown University Library Information Technology > <https://www.library.georgetown.edu/lit/code> > https://www.library.georgetown.edu/lit/code > 425-298-5498 (Seattle, WA) > > -- > You received this message because you are subscribed to the Google Groups > "DSpace Technical Support" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to <dspace-tech+unsubscr...@googlegroups.com> > dspace-tech+unsubscr...@googlegroups.com. > To post to this group, send email to <dspace-tech@googlegroups.com> > dspace-tech@googlegroups.com. > Visit this group at https://groups.google.com/group/dspace-tech. > For more options, visit <https://groups.google.com/d/optout> > https://groups.google.com/d/optout. > > > > -- > Tim Donohue > Technical Lead for DSpace & DSpaceDirectDuraSpace.org | DSpace.org | > DSpaceDirect.org > > > > -- > Tim Donohue > Technical Lead for DSpace & DSpaceDirect > DuraSpace.org | DSpace.org | DSpaceDirect.org > > -- Terry Brady Applications Programmer Analyst Georgetown University Library Information Technology https://www.library.georgetown.edu/lit/code 425-298-5498 (Seattle, WA) -- You received this message because you are subscribed to the Google Groups "DSpace Technical Support" group. To unsubscribe from this group and stop receiving emails from it, send an email to dspace-tech+unsubscr...@googlegroups.com. To post to this group, send email to dspace-tech@googlegroups.com. Visit this group at https://groups.google.com/group/dspace-tech. For more options, visit https://groups.google.com/d/optout.
