Emotion Twenty Questions in Chinese

Emotion Twenty Questions (emo20q) is a game for collecting language about how people talk about emotions and an artificial intelligence task to imitate this verbal behavior.

The Emotion Twenty Questions in Chinese (emo20q_cn) is a project to play emo20q in Chinese.

Getting Started

pip3 install -r requirements.txt

Past Work

English EMO20Q
- This is not very well organized. I made it when I was in grad school before I was following good programming practices. It was written in Python 2.7 originally. I've updated it so it runs with Python 3 but haven't had rigorous testing (the old code had doctest but it should have a unit testing component)

Current Work

Human-human emo20q in Chinese collected from WeChat (raw files not available publicly due to private data)
```
 ./wechat_pilot/groupchat*txt
```

Annotations/glosses

 ./wechat_pilot/emo20q_glossing_final.txt # not available publically
 ./wechat_pilot/emo20q_glossing_final_anonymous.txt

some scripts:
- split dialogs into turns:
```
./split_wechat.py 
```
- extract the list of emotions that players chose:
```
./emo20q/data.py -e
```
- generate html ./docs directory from the annotation/gloss file
```
./emo20q/data.py -w docs
```

To Do's

Refactoring old emo20q code
- separate dialog engine from emo20q implementation
- add better testing. I like doctest but probably better to use pytest
Figure out Windows newline difference
- using unix2dos wechat_pilot/emo20q_glossing_final.txt worked but it would be good to have a more convenient solution.
IDEA: translate README to Chinese
IDEA: study use of WeChat emojis
...

Contributors and Maintainers

Because this repository is a public version of a more active private repository, many commits have been removed/squashed and thus we acknowledge contributions here.

Shanshan Kong @sskong85 collected and annotated the data on WeChat
Abe Kazemzadeh @abecode maintains the repository and is the point of contact with the University of St. Thomas IRB