Edison Search Interface

The Edison Web interface is a web application that allows one to search the description and keywords of feature extractors and preview their test code to see how they had been built. More details can be found in this paper.

Technologies being used within Web Application:

• jquery
• jade
• bootstrap
• jquery-searchable

Running the application

This application requires the following to generate its components and to run:

• node
• npm (node package manager)
• bash
• java (version > 1.6)
• maven
• regular web browser (Safari, FireFox, Google Chrome, etc..)

Steps to run the application:

1. Extract the javadoc from the online javadoc repository.

   • Replace the base_javadoc_url variable with the appropriate base_url (The url that produces the API overview page)
   • The current default is: http://cogcomp.cs.illinois.edu/software/doc/illinois-edison/apidocs/
   • Run ./extract_javadoc.sh from the javadoc_scraper directory

2. Use the jdparser maven package to parser and generate the data.json file that will be ingested and presented by the web interface.

   • Ensure that the base_url variable in the src/test/java/cz/zcu/validator/jdparser/JDParserTest.java file contains the base url with the root package path of the feature classes.
   • The current default is: http://cogcomp.cs.illinois.edu/software/doc/apidocs/edu/illinois/cs/cogcomp/edison/features/
   • Run mvn test from the javadoc_scraper directory, and this will generate the data.json file in the base directory of our web application

3. Start the Web Application

• Run npm install to install all the web application package dependencies
• Run node app
• In your browser open: http://localhost:3000