* Desription

  * Need VoID descriptions, class and its the entity number ,property and its triple number,distinct subject and object number information
  * For triple patterns with bound variables which are not covered in the VOID statistics, the system send a SPARQL ASK query to pre-selected(after predicated and type selection) endpoints --- no limitation of predicate bound
  * Dynamic Programming and prefer bushy trees.
 


* source code: rdffederator - Revision 192: /trunk/src
 
   http://code.google.com/p/rdffederator/

* step:

 1. make the corresponding VoID files

  * method 1: 

  1.1 data download:
    
   drugbank:  http://download.bio2rdf.org/release/3/drugbank/  drugbank.nq.gz	19-Dec-2014 13:23	52M
   
   omim:      http://download.bio2rdf.org/release/3/omim/     omim.nq.gz	11-Nov-2014 03:50	124M	 

   pharmgkb:  http://download.bio2rdf.org/release/3/pharmgkb/  diseases.nq.gz  genes.nq.gz     rsid.nq.gz  drugs.nq.gz     offsides.nq.gz  twosides.nq.gz    01-Jun-2014 17:54   2.4G or so

  1.2 ./generate_void.sh ../data/drugbank.nq drugbank_void.n3 

  
{{{
    counting triples and properties
    counting types and entities
    counting distinct objects
    counting distinct subjects
}}}


   
    drugbank: time taken: 91 seconds

    omim: time taken: 411 seconds



 * method 2: 
Made the void files with the statistics queries in the following page:
[https://github.com/bio2rdf/bio2rdf-scripts/wiki/Bio2RDF-Release-3-Summary-Statistics]

2. execute: ./SPLENDID.sh SPLENDID-config.n3 query/query1.txt


For the VoID files with method 1,the following errors occur:
{{{
20:13:36 [WARN ] [Rio error] Expected ':', found '>' (465, -1) (ParseErrorLogger.java:28)
20:13:36 [WARN ] [Rio error] Namespace prefix 'LBSL' used but not defined (465,-1) (ParseErrorLogger.java:28)
20:13:36 [ERROR] [Rio fatal] Not a valid (absolute) URI: null (465, -1) (ParseErrorLogger.java:32)
20:13:36 [ERROR] can not parse VOID file file:/opt/services/fsearch/app/splendid/void/omim.n3: Not a valid (absolute) URI: null [line 465] (VoidStatistics.java:407)
}}}

check the corresponding void files, and deal with the lines as follows. we simply do the following changes:

void:property fusion> =>  void:property <fusion>


Result:

case 1: 5 results

||Query\time(ms) ||time 1 || time 2 || time 3 ||time 4||time 5 ||
||case1 ||6724	||4763	||4766	||4765||	4763||

case 2: error: Caused by: org.openrdf.query.QueryEvaluationException: Virtuoso 37000 Error SP030: SPARQL compiler, line 2: syntax error at '"GLUCOCORTICOID-REMEDIABLE"' before  '"i"'

case 3:  part of results(216) returned and  timeout occurred when processing request  org.apache.commons.httpclient.HttpMethodDirector executeWithRetry.

when use "limit 100" ,it took(ms): 348080,341991,339716,339089,340024

case 4:  timeout after returned 76 results 

when use "limit 100" ,it took(ms) :822992,619742,
case 5:  timeout after returned 304 results

when use "limit 100" ,it took(ms) 387507,148426,130344,130165,130181