litvinovg/vivo
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

1. Improvement of text for "refresh cache". Reviewed by Joseph B.

Browse source
This commit is contained in:
tankchintan 2011-07-21 21:08:14 +00:00
parent 54b390b851
commit 48b4ded3af
1 changed files with 12 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

29
productMods/templates/freemarker/visualization/tools.ftl
View file

@ -8,24 +8,19 @@
<a href="${refreshCacheURL}">Refresh Cached Models for Visualization</a> <a href="${refreshCacheURL}">Refresh Cached Models for Visualization</a>
<h5>Why</h5> <h5>Why</h5>
Certain class of visualizations like Temporal graph and Map of Science that work on a large scale Large-scale visualizations like the Temporal Graph or the Map of Science involve calculating total counts of publications or
attempt to calculate counts of publications (or grants) for all the organizations within an organization. of grants for some entity. Since this means checking also through of its all sub-entities, the underlying queries can be both
These queries tend to be both memory intensive and time consuming. In order to provide a good user experience when memory-intensive and time-consuming. For a faster user experience, we wish to save the results of these queries for later re-use.
these visualizations are served we decided to save the results of these queries to be reused later.
<h5>What</h5> <h5>What</h5>
To this end we have devised a caching solution which will store information about the organization To this end we have devised a caching solution which will retain information about the hierarchy of organizations-namely,
hierarchy, which publications were published from which organizations etc (among other things) in which publications are attributed to which organizations-by storing the RDF model. <br />
semantic way (i.e. we store the rdf model).<br/> We're currently caching these models in memory. The cache is built (only once) on the first user request after a server restart.
Because of this, the same model will be served until the next restart. This means that the data in these models may become stale
We're currently caching these models in-memory, which works nicely, for now. Currently the cache is only built depending upon when it was last created. This works well enough for now. In future releases we will improve this solution so that
once, on first user request after a server restart, so the visualization will essentially never be updated models are stored on disk and periodically updated.
until the server is restarted again. It is not being updated real-time or periodically. This means that the data in these
models might be stale depending upon when was it last created. In future releases we will come
up with a solution that stores these models on disk and be updated periodically.<br />
<h5>How</h5> <h5>How</h5>
To refresh these models either restart the server or click on "refresh cache" link above. We realize that The models are refreshed each time the server restarts. Since this is not generally practical on production instances,
in production VIVO instances it is not feasible to restart the server to refresh these models. Hence we administrators can instead use the “refresh cache” link above to do this without a restart.
recommend the above link. Administrators can use this link to trigger regeneration of all the existing <br />
models.<br />