Hysen Labs
Library / SDK
adityajo/sarcasmbot avatar
adityajo

sarcasmbot

SarcasmBot: An open-source sarcasm-generation module for chatbots

50 stars8 forksJavaGPL-3.0
01
DEEP OPEN-SOURCE ANALYSIS

sarcasmbot, a chatbot that answers with sarcasm

An open source Java module that replies sarcastically to user input, presented as a KDD workshop paper with authors and step by step setup.

02
DEEP OPEN-SOURCE ANALYSIS

The module

sarcasmbot is a chat generation module that replies sarcastically to user input. It is open source, and the README invites you to use the code and report bugs at the author's IIT Bombay email addresses. The behavior is a single defining trick: take whatever the user says and answer with sarcasm. That narrow focus is what makes it easy to slot into a chatbot pipeline.

03
DEEP OPEN-SOURCE ANALYSIS

The paper and authors

The project has a paper behind it. The authors are listed as Aditya Joshi, Anoop Kunchukuttan, Pushpak Bhattacharyya, and Mark J Carman, and the work was presented at the WISDOM workshop at KDD 2015 in Sydney. So the README doubles as the metadata for published research, with the citation right in the repository rather than buried in a references section.

04
DEEP OPEN-SOURCE ANALYSIS

Getting it running

The setup steps are concrete. Download a sentiment word list, set up an Eclipse or similar project and import the code from the SarcasmTalker directory, then download and uncompress the Stanford postagger. The jars go into the project, and you locate the tagger model file. The model file path is argument one and the sentiment list path is argument two when you call Main.sarcasticChatter. The software is distributed under the GNU General Public License, version 3, with the standard no warranty disclaimer.

05
DEEP OPEN-SOURCE ANALYSIS

Editorial conclusion

The module generates sarcastic replies to user input and is published work, cited at the WISDOM workshop. Setup needs a sentiment lexicon, a Java project import, and the Stanford tagger, and the code is GPL licensed free software.

06
DEEP OPEN-SOURCE ANALYSIS

Official sources

07
Community notes

Community notes