Minutes of Product Service Meeting F06G11301/02

01    SOLR implementation

Stats – Stefano will suggest appropriate stats to monitor. We will generate and receive monthly reports (as I remember you can set Google Analytics to do this automatically to selected mailboxes).

Interface/Texts – Stefano and Yves will examine and edit. Change “basic search” to “search”.


Deployment – Fabrizio says it is ready to deploy as soon as we have validation/editing of texts and accompanying translations.

Interface – onClick clear the keyword adding box.

News – we would like an AGRIS news stream that is generated from AIMS. This is probably low priority?

Documentation – Technical docs are complete and on network drive. (could package with code in SVN?)

Auto updating of Lucene indexes – could make a unix daemon to run it once a month, but does not remove need to validate records. (low priority?)

AgriDrupal plug-in – Fabrizio working with Valeria, is more or less finished. 1 week to go.


Interface/Texts – Stefano and Yves will examine and edit. Change “basic search” to “search”.

News item – publish an AIMS news item when deployment is ready.


Testing – ref library is testing, will give results by Dec. 15. They were given a document with some parameters for testing.


Translations – Spanish -> Maria, French -> Yves, Chinese -> Miriam?, Arabic -> Magdi?, Russian -> ?


02    Chinese Interface

One option is to simply send searches in Chinese to Agrovoc, get back the related terms in English and use those to search AGRIS.

Second option is to see how to re-use the work by Zhang on the AGROVOC-CAT mapper – should be possible to make it a plug-in for SOLR. Ideally it would be designed to work with any kind of SKOS file.

We decided to try option two, seeing if we could bring Zhang here for 3 months, provisionally Feb-Apr to do the work.

Fabrizio will package code and docs (let’s do it all on SVN) and Johannes will contact to make the proposal and ask to get their source code so we can look at it. Here is the design doc for Fabrizio: \\Kcew118\PWB-Material\AGRIS2010\Design_and_Implementation_of_CH2.doc


03    Linking Data

Following the SOLR release, we would like to test further run-time linking to other data providers such as we are now doing with Google, but focusing now on archives instead of search engines. Mentioned were EMBRAPA, PLoS (http://www.plos.org/), Springer Open, CGIAR Centers (http://cgvlibrary.wordpress.com/cg-center-data-bases/). The emphasis would be on archives with actual content not just metadata. Maybe also focus on who has an API that will return structured data instead of html so we can streamline results into a cohesive result. Second possibility would be to use a technology such as SiteMesh (http://www.opensymphony.com/sitemesh/).

We could also consider in future harvesting other archives’ metadata to add to our own Lucene indexes.


04    Triplification Experiements and possible Linked Data Extractions / 05 Preparation of December Workshop

The AGRIS workplan will await the results of the workshop and thus we will not provide one ahead of time.

The idea is that we lay out the problems and then we listen as people propose solutions.

We agreed to present an AGRIS wish list by Dec. 6th. Johannes and Stefano will start the page in AIMS.

Johannes will package together docs 01-08 to give participants a background on AGRIS future directions.

Fabrizio will prepare some triple store in Sesame examples and will improve the current prototype by adding in AGROVOC and Journal URIs

We agree to have a preliminary agenda ready by next Wednesday. The broad idea is that we will present the problem on Thursday the 16th and they will present the solutions the following day.

Some agenda items we discussed:

a)      Vision – Yves will present

b)      Naïve linking as currently implemented

c)      Journal disambiguation

d)      Links to AGROVOC


f)       Sesame triple store

g)      Rhizomik

h)      Inspire activities of this last year