...

Inferencing can be very time-consuming.  If you are loading a large set of data (thousands of people and tens of thousands of publication, that is, sets of triple triples in the 15M and above size) you may benefit from turning off inferencing during load and recomputing inferences (see below) after initial load.  And if you have added the inferred triples using your own processes, before loading, you may not need to recompute the inferences.

Yo You turn off inferencing at start-up:

1. Find the file WEB-INF/resources/startup_listeners.txt
2. Edit the file and comment out two lines as shown below
3. Restart VIVO.  VIVO will no longer do reasoning at start-up
4. Load data.  Data will not be inferenced.
5. Once data is loaded, you may wish to use the the SysAdmin Site Admin interface to recompute interfacesinferences.  You may also recompute interfaces recompute inferences using the URL shown in the next section.

...

As data is loaded, VIVO updates its search index.  This may take considerable time if the number of triples being loaded is high (millions of triples) .  In such cases, there may be an interest in turning off indexing during the data load and rebuilding the search index as shown in the next section at a more convenient time separated from data loading.  If data is loaded and indexing is not performed, the new data will not be found or displayed using the VVO application.(to be written)

To disable indexing, edit WEB-INF/resources/startup_listeners.txt, comment out the line below, and restart VIVO.

# edu.cornell.mannlib.vitro.webapp.searchindex.SearchIndexerSetup

Re-building the search index

...