Updates
add more bullets as necessary
- Brown
- Colorado
- Cornell
- Duke
- Indiana – should we schedule a code review?
- Memorial
- NYU
- Penn
- PSU
- Stony Brook
- UCLA
- UCSF
- Weill Cornell
Implementation Fest outcomes and feedback
For those who attended, what did you find most useful? What did you learn? What impressed you about other VIVO implementations?
Brief report on Saturday coding activities
If you haven't yet, please provide feedback (anonymously if desired) at
Updates
- Brown (Ted) – rolled out some changes to templates and the sorting of the browse screens; preparing for the I-Fest
- Colorado (Stephen) – deployed release 1.5.2 to the staging environment and rolled it out after the data update today – easy with the 3-tier build to deploy to both Colorado Springs and Boulder campuses; so far no issues. Have also releases a new automated harvesting process to testing and working through the various components, but a first successful run yesterday.
- Cornell (Tim) – looking at performance of the SPARQL query to bring geographic focus onto a map within the home page; the query is trying to roll up numbers for U.S. states into the total for the U.S. on the world map. Has moved from a data getter to a controller paired with an Ajax call. Working on disambiguation of people based on frequency of last names, and hoping to be able to vary the strictness of matching so that unusual last names can be matched with lower exactness than very common last names.
- Duke (Richard) – preparing for release; looking at publications and characterizing people based on high, medium, or low publication counts to help with performance.
- Alex – there's a session next Friday morning on troubleshooting that will touch on performance issues, and will support remote call-in
- Indiana (Rohan) – Found a concurrency issue and have fixed it – hope to push the source code in a week or so after further testing and code review
- 20-23 minutes of loading time for the entire University of Florida, after which the Science Map visualizations load almost instantly
- Jim – would be happy to participate in the code review, as would Chin Hua
- NYU (Yin) – busy with creating a new summer school course; will be a special NIH study section on grants affected by Sandy; just got their computing cluster back up. Trying to get their Vmware server back up that had VIVO data on it
- May be able to participate in the Saturday development day remotely
- Has an algorithm in mind for modifying the search ranking and would like to focus on that and get some feedback
- Weill Cornell (Paul) – a couple further people to test on publications ingest, and will then be deploying to production as a lookup source f
- starting to do reports with MeSH terms – using it to look at a community of researchers to see their subject areas in common; looking at MeSH terms in published papers and comparing them to the block of MeSH terms associated with diseases; will also show how people fit into a hierarchy of MeSH terms. Will eventually want to show this for each term on an author's articles
- Noticing that there are a lot of duplicate terms in UMLS – 3 different hits for HIV, for example, with links to different source vocabularies – had been hoping to use the cleaner hierarchy of Snomed CT but that does not have as broad a coverage as UMLS
2013 VIVO Conference Call for Posters and Papers – due May 10
- Call for Papers abstracts are due May 10, 2013
- Call for Apps entries are due July 31
- Additional details on the VIVOWeb.org blog
Upcoming Webinars
- DuraSpace will be organizing and publicizing a series of 3 webinars in May and June focusing on VIVO
- The first (May 14) will be a general introduction
- The second (June 4) will include 3 case study presentations on VIVO implementations at Colorado, Brown, and Weill Cornell Medical College
- The final webinar will be a "technical deep dive" on June 11
Notable implementation and development list traffic
...
- Memory leak requires restarting Tomcat?
- (Brian L) – looking at what's happening on the VIVO Cornell server
- taking heap dumps to look for objects that are hanging around across garbage collections
- (Alex) – we restart MWF with data updates so have not had a problem, but will look for messages
- Other possible approaches:
- Verifying that what is being seen is actual memory usage, not just the size of the JVM as reported by top() – check the gc (garbage collection) logs
- adding a JSP page to force a full garbage collection may verify whether the baseline memory used by Tomcat is in fact increasing
- Check the possibilities for thread contention that could cause the application to freeze up
- Make sure your robots.txt blocks access to visualizations and CV generation (if you have that activated, which I think only UF does)
- What other sites are experiencing this problem? – checking the Cornell VIVO logs this morning
- Verifying that what is being seen is actual memory usage, not just the size of the JVM as reported by top() – check the gc (garbage collection) logs
- (Brian L) – looking at what's happening on the VIVO Cornell server
- Any way to batch convert Ontology URIs?
- (Brian) A hybrid approach might be workable
- could use the existing change namespace tool (designed for changing the namespace of the VIVO) to update all the ontology URIs in the instance data graph, if it's all in the main (KB-2) graph and accessible to that tool in its default configuration
- then you could export the ontology only, use a text editor to modify the URIs, and load it back in
- make a JIRA issue
- (Brian) A hybrid approach might be workable
Resolved but worth discussing
- (Mark) Bug in Tomcat 7 for Ubuntu – source of the "There is no <whatever> content in your VIVO" problem from last week
- (Jim) However, it's obvious that we need to make it easier to isolate and diagnose this sort of problem. As a first step, I have started a wiki page with suggestions for troubleshooting this sort of issue.
Still open?
Brown (Ted) -- Meeting with faculty departments to get feedback, e.g. faculty might want to upload a full CV and link to that from VIVO profile
the pairtree-derived file system set to store arbitrary files -- we happen to use it only for images so far
in the ontology is related to the person’s URI via a mainImage property and has a MIME type
but have never tried to store anything but
could also explore uploading to Fedora Commons (the repository, not the Linux OS)
Faculty are also interested in linking to full text of their publications.
Colorado (Alex and LIz) -- starting publications work. Stephen has been working on replacing a Selenium-based ingest with a Harvester-based approach; this week the results matched up when run in parallel so are ready to start the switchover. The Harvester-based approach is much faster because it does not work through the UI. Starting to work on visualizations showing collaboration around research interests, not just co-authors and co-investigators. This should be valuable for institutions not ready to undertake collecting publications data.
Using D3.js -- iDevice friendly visualizations
Example is a new college being created around media, technology, and art -- trying to find where those connections already exist on campus
Cornell (Jon and Jim)
working on 1.6 ontology changes, trying to keep migration patterns for context nodes like Positions and Advising Relationships
Jon and Jim at Implementation Fest
Jim looking into caching issue with help from Arve at Griffith and Ted from Brown
IFest internationalization work is being shared by JohnF at a UN meeting on collab of international agriculture researchers
Duke (Patrick and Sherri)
May 13 School of Medicine go live -- finalizing / curating data as needed, majority of faculty so a big group
Johns Hopkins
Indiana -- should we schedule a code review?
Jim has been iterating with Rohan et. al. based on their pull request
Memorial (Lisa)
back from the I-Fest safely and looking forward to the conference
NYU (Yin)
tried to join IFest code sprint, would still like to connect with some VIVO devs interactively to review his approach, use of VIVO GitHub, etc
Maybe see https://wiki.duraspace.org/display/VIVO/VIVO+Developer+Setup+101
Alex: is there a use case for submitting changes back to VIVO even if they are options that not all sites may want? A community branch? but that may get forked, and someone will need to integrate the changes back into the branch they originated from
Stony Brook (Erich and Tammy)
new university wide VIVO instance work starting
question about generating distinct URIs in a way consistent with VIVO’s approach (Brian Lowe: IndividualDaoJena code calls URI generation code)
UCSF (Eric)
next release of Profiles on their plate
ORNG gadgets (http://www.orng.info) for adding multimedia to Profile pages, considering Cornell’s VIVO and ontology they see “media contributions” meta data but they can’t see the media itself, ex. a video clip of a faculty interview on Channel 7 or a press release
ORNG understanding RDF on page?
Weill Cornell (Paul)
- Eliza more productive when Paul’s in Boulder :)
- working on automating publication ingest process to free up devs
- met with Grad School Office interested in tracking pubs by current students, question about alumni pubs
- Jon discussed use case of monitoring compliance with NIH public access mandate on a grant with 1,000+ co-investigators – PI could receive 115 notices from NIH system for investigators who have not uploaded copies of their NIH-funded publications to PubMed. NIH tracks by grant and Weill can track by publication to notify the correct authors
I-Fest outcomes and feedback
Apologies to those who tried to remote in --
For those who attended, what did you find most useful? What did you learn? What impressed you about other VIVO implementations?
Report on Saturday coding activities
- Ted, Stephen and Jim working on internationalization -- documented on the wiki (search “multiple language support”)
- Stephen ran into an issue after upgrading to Java 1.7 with Tomcat 6 -- resolved by upgrading to Tomcat 7
- Jon and Alex working on setting up Jon’s Mac and start of wiki page (https://wiki.duraspace.org/display/VIVO/VIVO+Developer+Setup+101)
- Pedro from USC working with Michael and Bill from PennState on using Karma to get data in
- will be trying Google Hangouts for small group coding and/or implementation support activities
If you haven't yet, please provide feedback (anonymously if desired) at http://goo.gl/z9lUr
Twitter stream: https://twitter.com/search?q=%23vivoifest13&src=hash
Slides and other materials are shared on SourceForge file repository -- also there were some notes captured as Google Docs (how do we collect these)
2013 VIVO Conference Call for Papers and Posters due May 10
Apps due July 31
Upcoming VIVO Webinars (all Tuesdays at 11)
May 14 -- Overview of VIVO
June 4 -- Case Studies: VIVO at Colorado, Brown, Duke and Weill Cornell Medical College
June 11 -- VIVO Technical Deep Dive
Notable implementation and development list issues
Harvester transfer dying unexpectedly (Giuseppe)
on a 64-bit machine I experience this strange issue of the harvester-transfer reporting
"killed". Here's the extract of the log.
2013-04-26 18:08:51.747 TRACE [o.v.h.u.r.JenaConnect] loading record: publication/75007
2013-04-26 18:08:53.570 TRACE [o.v.h.u.r.JenaConnect] loading record: publication/75008
/root/vivo/harvester/bin/harvester-transfer: line 29: 1328
Killed java $OPTS -Dprocess-task=Transfer org.vivoweb.harvester.transfer.Transfer "$@"
Note that this happens after hundreds of publications have been loaded.
Search in Chinese (Jianwei) – a local fix that will be added to VIVO 1.6
Locate the file vitro-core/webapp/src/edu/cornell/mannlib/vitro/webapp/controller/freemarker/UrlBuilder.java
In this file, search for the two occurrences of "ISO-8859-1" and replace them with "UTF-8"
Rerun "ant all" to rebuild VIVO
After restarting Tomcat, the links on the right side of the search results should work properly with Chinese characters
SQLNestedException: Cannot get a connection, pool error Timeout waiting for idle object (Tom T.)
related to MySQL max_connections configuration? – seems unlikely
checking vivo e 'show processlist' while performing a reindex if possible
may relate to the shift in connection pool libraries between VIVO 1.4 and VIVO 1.5
- Alex will follow up with Tom to see if someone from Symplectic is in the loop
...
Call-in Information
Topic: VIVO weekly call
Date: Every Thursday, no end date
Time: 1:00 pm, Eastern Daylight Time (New York, GMT-04:00)
Meeting Number: 641 641 825 891
To join the online meeting
Go to https://cornell.webex.com/cornell/e.php?AT=WMI&EventID=167096322&RT=MiM2
If requested, enter your name and email address.
Click "Join".
To view in other time zones or languages, please click the link: https://cornell.webex.com/cornell/globalcallin.php?serviceType=MC&ED=167096322&tollFree=1
If those links don't work, please visit the Cornell meeting page and look for a VIVO meeting.
To join the audio conference only
To receive a call back, provide your phone number when you join the meeting, or call the number below and enter the access code.
Call-in toll-free number (US/Canada): 1-855-244-8681
Call-in toll number (US/Canada): 1-650-479-3207
Global call-in numbers: https://cornell.webex.com/cornelluniversity/globalcallin.php?serviceType=MC&ED=161711167&tollFree=1
Toll-free dialing restrictions: http://www.webex.com/pdf/tollfree_restrictions.pdf
Access code:645 873 290