/sciologer

Visualization system for exploring relationships in data

Primary LanguageJavaScript

sciologer

Visualization system for exploring relationships in data

Version 0.1

Created by Michael Bales in 2008

Requirements

  • Tomcat
  • Icons needed for the visualization may be downloaded from the drawing folder

Implementing Sciologer

To implement Sciologer locally, do the following:

  • Copy the e folder to your machine in the location of your choosing
  • Modify file_settings.java to point to this folder
  • Install Tomcat
  • Configure Eclipse for Tomcat
  • Find the file httpd.conf, for example at /etc/apache2 and change the DocumentRoot to match the local folder where SciologerAdv.html is located
  • Restart Tomcat through Eclipse
  • Install Google Earth (may need to use sudo if installing on Linux)
  • Point browser to http://localhost/SciologerSVN/SciologerAdv.html
  • Run test query: F1275 (This query returns only four records from PubMed as of July 2017. Other test queries: neuroligins; neuregulins)
  • Open the resulting KML file (the browser should tell you where it is)

Troubleshooting

  • Make sure CVResults is set to upload file to the user
  • Verify that web.xml is configured properly
  • Check permissions in the e subfolders
  • If nodes are too large or too small, scaling may be modified in eleph1nt.java

Additional notes

  • Try several test queries to verify that advanced search is working properly
  • When running on a server, to make modifications to Sciologer it is sufficient to upload modified class files. These are located at: SciologerSVN\build\classes\bales

Descriptions of subfolders of the e folder

  • a: adjacencies
  • ab: abstracts
  • act: all corpus tokens
  • d: parsed
  • f: word frequencies
  • fc: full citations
  • fc_util: postprocessed full citations
  • k: kmz builder utility folder; stores files needed for building zipped kmz files which include the kml file, ground overlay, and relevant icons, models, photos, or drawings
  • li: links
  • np: noun phrases
  • o: output
  • p: node positions
  • pd: node positions sorted by degree
  • pt: node positions sorted by type
  • q: queries for batch mode
  • stemmed: stemmer output
  • tok: tokenizer output
  • tok2: postprocessed tokenizer output
  • u: umls utility file

Order of operations

  • Controll3r runs Entrez E-Utilities and generates files in: fc (full citations)
  • Optional -- if including tags
  • tok=tokens -- A utility folder for storing intermediate tokenizations
  • tok2=tokens2 -- A utility folder for storing intermediate tokenizations
  • act=all corpus tokens -- A list of common token frequencies for the entire corpus from which the search results are drawn, in the format token|frequency.
  • ti=token inclusions -- A utility folder for storing data on whether tokens are to be included
  • Optional -- if including UMLS lexical variants
  • u=umls lexical variants -- a folder for storing the file umls_with_lexical_variants_all.txt
  • (This file must be included on server upon implementation of Sciologer.)
  • Controll3r runs procfc which generates files in a (adjacencies), using the typepair list format: net "np_mutation" "np_gene FOXP2" 1
  • Controll3r runs ccvisu which generates files in p (positions) using this format: -28.165995 14.799589 1.2726835 5.0 new brain imaging study 65280 false
  • Controll3r runs NodeSortByType which generates files in pt (positions sorted by type)
  • Controll3r runs NodeSortByDegree which generates files in pd (positions sorted by degree)
  • Controll3r runs PrepLinks which generates files in li (links)
  • Controll3r runs eleph1nt which generates files in o (output)