A seachengine to retrieve informations from the homepage of the university leipzig. Created during the information retrieval internship in winter semester 2017.
Like in google, the tool should be able to serve query recommendations right when you type in your query.
Please note that we could change the scope of the ticket in any way. The things that I have mentioned here is only what first comes to my mind and should serve as a discussion base for determining the real scope of what is to be done in the ticket!
The following stuff needs to be done:
Define a persistence structure where:
all executed queries could be stored
a fuzzy search for a given user query is possible
user interactions with recommendations could be stored (like "for query xy we recommendet "a", "b", "c" and the user choosed "c").
Maybe we could use some fancy tools (e.g. not only a single rdbms) like event sourcing (apache kafka, rabbit mq) or stuff like that (that would be really interesting :) )
For a given user query:
find and rank in a fuzzy style recommendations for the stuff that a user types
No Logging of user interactions needs to be done, since such stuff will be done in #1
Like in google, the tool should be able to serve query recommendations right when you type in your query.
Please note that we could change the scope of the ticket in any way. The things that I have mentioned here is only what first comes to my mind and should serve as a discussion base for determining the real scope of what is to be done in the ticket!
The following stuff needs to be done:
No Logging of user interactions needs to be done, since such stuff will be done in #1