emotionannotate

For the SMILE project

Set up

Python 2.7 with pip 7.1.2

Use 'pip install -r requirement.txt' to install all relevant Python dependancies.

Run web app

cd src

python app.py

Run batch classification

If you have a list of tweets that you want to run our emotion classifier on, you can:

cd src

python Classification.py --inputdir --outputdir

e.g. 'python Classification.py --inputdir ../input --outputdir ../output'

More info in src/readme.txt

Input/output data format

Input and output are both in json format. For example,

{"tweetid":"614912375288893440", "text":"@britishmuseum awesome museum"}

as an input entry,

{"tweetid":"614912375288893440", "text":"@britishmuseum awesome museum", "emotions":{"anger":"no","disgust":"no","happy":"yes","sad":"no","surprise":"no"} }

as its output.

Building the Development Environment

Requirements

• Ansible
• Vagrant
• VirtualBox

Starting up

From the project directory:

vagrant up

This will download the relevant Vagrant box, create the development environment and install the required packages. Once finished, you can use:

vagrant ssh

to connect to the box. To start the development server:

cd /vagrant/src/
python app.py

The system will be running on your machine. You can load the homepage in a web browser: http://127.0.0.1:9000