MP3 uses a pyschoacoustic model in its lossy compression so that it doesn't bother to reproduce sounds that can't be heard. This is the same, except for text: it replaces long words with shorter ones, sometimes even the shortest, most expressive 'words' available in Unicode: emojis.
- You'll need to obtain the file 'full-emoji-list' from the Unicode Consortium.
- Then run 'python extract_emoji.py' to create emoji file
Run tw.py, and navigate to localhost:5000