chprice / anagram Goto Github PK

Set up the ORM code for the database for storing information on which anagrams have been generated, selected as best, what weights to put on the different metrics, and storing of individual metric data.

Add bonus words metric

Users are interested in topics. Allow phrases that feature certain words, or things related to certain topics, to be rated more highly.

Add verb noun agreement metric

Rate more highly verb-noun sets that logically make sense.

Add syntactic commonness metric.
Measures how likely the part of speech of the phrase is to naturally occur in English.
Adjective Noun is much more common than Adverb Noun, etc.

Set up Finatra and get it accepting requests.

Don't worry about CSS.

Also, don't forget to update the readme with references to the libraries I'm pulling from.

Add interesting POS combinations metric

If a user votes up a particular phrase, then that POS sequence is voted up. The most well voted POS sequences are given a higher weight based on how often they've been voted up.

Add a metric to measure the wordnet distance of the words in a phrase.
The idea is that related words are related topics, which are more interesting than random words.

Currently the results are filtered through a set so the redundancies are removed.
It would be far less intensive if the filtering occurs in the generation step.