STI UI demo

This application provides a UI interface to interact with STI and visualise its results. It is implemented based on Node.js.

Installation

• Download STI
• Install Node.js on your computer (see link)
• You need to ensure you have the following Node modules: express and socket.io. By default, these are included in the download (/node_modules)
• Build STI. cd into the parent folder of this folder, run mvn clean install. If build is successful, you should see a sti-[version]-jar-with-dependencies.jar file in the /sti-jar-builder/target folder
• cd ui into the root folder. Edit config.json following descriptions below.
• Start the serer by node main.js
• Open a browser and point your address to http://localhost:3000/

Configuration config.json

This file configures the UI application. It links to other files used to configure TableMiner+. The following fields are defined. If you do not need email notification when a task completes, you do not need to change anything. Otherwise, as minimum, you must change emaillogin, emailpass, emailhost, emailport, emailfrom to configure your own email account.

• lib: should point to the jars required to run TableMiner+. By default, this is '../target/*', where your maven build will place a sti-[version]-jar-with-dependencies.jar
• cache: should point to the folder containing the Solr index caches. By default this is '../resources/cache'
• tablePrevClass: a class used to download the webpage specified by user and preview its content before annotating the contained tables. This should be uk.ac.shef.dcs.sti.ui.InputfilePreview
• stiMainClass: the main class to run the TableMiner+ task. This should be uk.ac.shef.dcs.sti.ui.TableMinerPlusSingle
• tmp: a system required folder to keep intermediary files
• error: an error webpage to display when things go wrong
• output: a system required folder to contain output files
• log4j: the log4j configuration file. You must set this correctly to see the progress logs. By default this is '../config/log4j.properties'
• stiprofile: the configuration file for running TableMiner+. By default this is '../config/sti.properties'. You will be able to chanage the settings through the UI. Detailed instructions are included in the file.
• stiwebprofile: the configuration file for Web search used by TableMiner+. By default this is '../config/websearch.properties'. You will be able to chanage the settings through the UI. Detailed instructions are included in the file.
• stikbprofile: the configuration file for knowledge base search used by TableMiner+. By default this is '../config/kbsearch.properties'. You will be able to chanage the settings through the UI. Detailed instructions are included in the file.
• server: the server address.
• emaillogin: TableMiner+ emails you when a task is complete. To do this it uses an email account. This is the login email address for that account. If not provided, you will not receive email notification.
• emailpass: The password to log into the above email account.
• emailhost: The host of the email account
• emailport: Port number of the email host
• emailfrom: Sender of the email
• tableParsers: input files (e.g., Webpages) must be preprocessed to extract the HTML <table> elements from the Webpage and these <table> must be represented as an internal Java object. A table parser is responsible for this job. Depending on different Webpages, you may encounter different <table> templates that require different processing. Hence Website-specific table parsers may have to be implemented from time to time. Currently there are four table parsers implemented, that can be found in the uk.ac.shef.dcs.sti.parser.table package. To support new Webpages, you need to create a new implementation of TableParser, which must implement Browsable if you want it to support Webpage preview in the UI. Then add that class into this JSON property.