abox/tbox - what can be done, what defines success
Don - work on item 1, documentation, correlating files with the graph, use cases like label changes
Brian is onboard with documenting this, does this serve the use case or not. Need to know how it works before we change
Item 3 - moving sdb to tdb
Don - sdb/tdb is content model, not config model, would like to debate config vs content model
Ralph – both tied together
Mike - sdb vs tdb – this needs to be a discussion. SDB is deprecated
Andrew - some of this discussion is captured in a ticket. Should sdb/tdb be part of abox/tbox sprint
Michael - if tdb is default, need to document this. Need to decide what to do with SDB being deprecated. Let’s have a calendar discussion on SDB/TDB
Huda - SDB/TDB is a different conversion
Don - agrees with Huda
Don - abox/tbox is more about config vs content model instead of sdb/tdb
Andrew – documentation is a measure of success. Correlate files to the graph. Detail use cases. Don has some. Walk through how to make updates in order to support use cases. Document ontology process and the loading of graphs.
Brian - agrees - these all go together
Andrew - anybody else with abox/tbox
Mike - can help
Elastic/Solr
2 items
Solr upgrade
Swapping SOLR with ES
Ralph: interested in upgrading SOLR
How to document and upgrade
What is the process to swap out SOLR for ES
SOLR upgrade is necessary due to a SOLR bug
Can look at this as a ticket, regarding the security piece.
Andrew - Asim is interested in helping out. Not sure how available he is.
Mike - is purpose to use current version of SOLR? Will sites have to install SOLR separately ( or ES ) and then work on how to integrate with VIVO.
Jim/Huda - concur
Huda - can we get VIVO to work with other search indexes. The code for ensuring that the search index can communicate with the rest of VIVO needs to be looked at more. Also, look at schema updates from v4 to v7. So deployment is part of this eventually, not immediately.
Andrew: work has been started to upgrading SOLR, but not in a place that we just need to document, still need to explore.
Huda: missing piece to work on is API piece, got it that there are no more smoketest errors.
Andrew: Jim demo’d ES search swap in. Jim will not be in sprint. Given that, does Jim see a productive course of action for sprint team to continue.
Jim: will be on email to answer questions. Provided code link to get ES to work. So this comes back to deployment question. Is this something we want people to use, if so what instructions to provide.
Mike - given the new defaults, we know that embedded SOLR can NOT be the default anymore. So we need to work on deployment and doc such that VIVO builds properly.
Ralph: concurs, go with SOLR 7.
Mike: so we need to tell people that they need to install SOLR or ES
Jim: we don’t tell people how to install tomcat or mysql, do we need to handhold for the index
Mike: needs some sort of highlight that there’s a new step for people.
Andrew: what to scope to this sprint and things that we do scope to sprint will be rolled to an upcoming release. So do we imagine/hope that having an external search engine will be part of the next release.
Mike; thought sprint item was to have a selectable index as part of next release.
Steve M: using default SOLR at Brown. Still confused regarding the total scope of the effort
Andrew: so names for this sprint —
Asim, Ralph, Don (part time), Mike said Graham, Steve McCauley, Ralph
Don - did someone reach out to Duke
Andrew: reached out, but Duke will not be participating in this sprint
Don: can this include item 3 of having a rich json document.
Andrew: based on the time and commitments, would be reluctant to integrate
Huda: as requirements get figured out, then have an effort on how to map rdf to the product evolution document. Need infrastructure in place, but the question of how to populate the product evolution index document can be done separately. Suggesting that the mapping of which sparql queries will populate the PE index can be done separately. Then, depending on the PE document spec, we can determine if SOLR or ES can deliver that document format.
Brian: the idea of of designing rich documents for the VIVO index is like designing an API. Need to have a architectural discussion first.
Mike: there’s a discussion of how to design the json doc in PE, but it needs to be designed first by the PE group.
Brian: how do we expect to have these things designed in the future, if PE is doing something in the future, how will that work with the vivo core. Does PE want core team to give them stuff.
Ralph: PE group process is matured, they are designing the doc now. It’s not a replacement but just an example of what they could do.
Andrew: coordinate between groups. So there are different efforts that are moving along at their own pace. But hard to reconcile at this point.
Mike: PE was intended to be an advisory group to the leadership, still looking for advice. They are pushing for modernization.
Brian: concerned that situations will happen like vivo-isf, isf was better, but it impacted everybody. These things make him nervous because things get in the pipeline, but if we look later we might find that with more design we might improve things.
Huda: rationale on what we want vivo to be and do. Have more rationale. How does this tie to the larger objective.
Andrew: so we replicate the experience that Jim created ( separate indexes). This will give us a sense on what deployment looks like and could lead to documentation.
Internationalization
Capability map
Andrew inclined to remove the map depth from sprint.
Kitio concurs
Home page and headers of profile
Team: kitio, christian, ralph, perhaps manuel schwarz), and mike of course for documenting
Andrew will send out an email to coordinate kicking off for the sprint
Other
Don wants to submit a pull request for schema fix.
No objections from group. Just try to also do this for 1.10
Andrew: looking for signoff on 1501
Andrew: yinlin wants support for handles ( or dois ). Looks like this should move forward on agenda item 10 #1560.
Previous Actions
Alex Viggio will bring news of Elasticsearch instead of Solr up with Product Evolution. Might there be consequences for the September sprint.