README
---
Code & data to reproduce the analyses in our ACL 2014 paper: 

    Humans Require Context to Infer Ironic Intent (so Computers Probably do, too)
        Byron C Wallace, Do Kook Choe, Laura Kertz, and Eugene Charniak

Made possible by support from the Army Research Office (ARO), grant 64481-MA / W9111F-13-1-0406
"Sociolinguistically Informed Natural Language Processing: Automating Irony Detection"

Contact: Byron Wallace (byron.wallace@gmail.com)

We make the data available in several ways; if you just want a simple CSV, see below! 

* The actual database - a flat sqlite file - is ironate.db.zip, it needs to be unzipped, of course. 
* The database-schema.txt file (in this directory) contains information regarding the database.
* See irony_stats.py for instructions on how to reproduce our analyses. This also gives examples on working with the database in python (and in SQL, since we issue queries directly). Note that this requires sklearn, numpy, & statsmodels modules to be installed.


**** 
We are also making the data available as a simple CSV, with each row corresponding to a comment and a summary label provided. This label is 1 if any annotator (of the three) labeled any segment in the corresponding comment as 1. This is therefore a very liberal definition of 'irony', and may or may not be appropriate, depending on the aims and task.