Java Servlet-based interface to the Ashurbanipal data.
Exploring Project Gutenberg with natural language processing.
This project supplies the processing capacity to the ashurbanipal.web.ui front-end. The project uses Apache Wink for no very good reason, but includes four resources supplying five URLs:
-
FileMetadataLookup.java
-
/lookup: Query the Project Gutenberg metadata to provide a list of matching texts. Query processing, idiotic as it is, is based loosely on the book Introduction to Information Retrieval, by Christopher D. Manning, Prabhakar Raghavan, and Hinrich Schütze.
-
/lookup/{etext_no}: Return the metadata associated with a single Project Gutenberg text. This is needed when using a URL similar to http://dpg.crsr.net/#773 or the forward/back history support.
-
-
FileCombinationRecommendations.java, FileStyleRecommendations.java, and FileTopicRecommendations.java: Provide lists of recommendations for a supplied etext number.
-
/combination
-
/style
-
/topic
-
Data for the application is stored in tab-separated text files, for easy reading by the application on start-up. While the application is running, it is all stored in memory to improve response time.
-
ashurbanipal.web.ui: Javascript client UI to the ashurbanipal.web interfaces.
-
ashurbanipal: Applications to generate the data set on which recommendations are based.
Tommy M. McGuire wrote this.
GNU GPLv2 or later.