-
Notifications
You must be signed in to change notification settings - Fork 8
Concept editor interface
The concept editor interface is an implementation of CB-Ograph an abstraction of OrionDB's graph editor interface. The concept editor in polar insights enables users to tweak and build an ontology of interest which will drive extraction and analytics work flows.
###Persisting ontologies Once you've built an ontology you are satisfied with you need to save it onto local storage (browser memory). If not the ontology will disappear after refreshing the browser. Once the ontology is saved onto local storage refreshing the browser will load the saved ontology. The ontology persists across browser sessions.Users can upload/download ontologies onto elastic search, this helps sharing ontologies across computers and browsers. The application requires a copy of a working ontology in local storage to function. When using the application for the first time users are required to download the latest ontology from elastic search and save it in local storage.
###Concept Aliases Concept aliases need to be defined while concepts are defined. (eg) CO2 is an alias for the concept Carbon Dioxide. The application maps instances of extracted entities to concepts via aliases. Since the application is aware of the various concepts and their aliases, When a user queries for Carbon dioxide it automatically fetches documents with extracted entities containing 'carbon dioxide' or 'CO2'.
###Exporting ontologies
Once the user is satisfied with the defined ontology he can export the defined concepts. This opens up a modal window containing a list of all the defined concepts. The user needs to copy the contents of this modal and paste it in ./insight-generator/data/entities.txt
. This file is used by the insight generator python module for entity extraction.
Information Retrieval and Data Science (IRDS) research group, University of Southern California.