This is code for running a survey that discovers how different people group word contexts into senses.
If you are going to use it in your own work, please change text and email in
./templates/start.html, and email in ./templates/group.html.
Also please site our (submitted) paper :)
Also note that this survey does not work in mobile browsers due to drag-n-drop issues (should be possible to fix). Desktop IE is supported.
There is an alternative design with jars instead of baskets
in the new-variant branch.
Check out user instructions in ./templates/start.html.
You can add ?source=some-value argument to experiment link, this will
be tracked in the source field.
You can download results for respondents who completed all tasks
at /export URL, and results for all respondents at /export?all=.
Partisipants will be in _participants.csv, and there will be a csv file
for each participant.
Install (short version):
pip install -r requirements.txt
In more detail, assuming Ubuntu 20.04:
sudo apt install python3-venv python3-dev git clone git@github.com:lopuhin/sense-grouping-survey.git cd sense-grouping-survey python3 -m venv venv . venv/bin/activate pip install pip -U pip install -r requirements.txt
Data is stored in an SQLite database in ./db.sqlite3.
Run migrations and load initial data:
./manage.py migrate ./manage.py loaddata survey/fixtures/initial_data.yaml
Next you can load dummy contexts for testing:
./manage.py loaddata survey/fixtures/contexts.yaml
or load real contexts from a more convenient format
(see example-contexts.csv):
./manage.py load_contexts "example-contexts.csv" # or ./manage.py load_contexts_with_fillers "example-contexts-with-fillers.csv"
Create superuser:
./manage.py createsuperuser
Start development server:
./manage.py runserver
Fist, collect static:
./manage.py collectstatic
Then there are many options for production deployment, see https://docs.djangoproject.com/en/1.10/howto/deployment/
One easy but not optimal (it's better to put nginx in front)
way is to use just gunicorn and supervisord:
sudo apt install supervisor
Then create sgs.supervisor.conf file along this lines
(again, this is far from ideal in many cases, just an example):
[program:sgs] environment=SECRET_KEY="generate some secret key, dont share it!" command=/root/sense-grouping-survey/venv/bin/gunicorn sgs.wsgi -b 127.0.0.1:8000 -w 4 --timeout 60 --access-logfile=/root/sense-grouping-survey/access.log --access-logformat '%%(h)s %%(l)s %%(u)s %%(t)s "%%(r)s" %%(s)s %%(b)s "%%(f)s" "%%(a)s" %%(f)s' directory=/root/sense-grouping-survey/ autostart=true autorestart=true redirect_stderr=true stdout_logfile=/root/sense-grouping-survey/error.log
and then do:
supervisorctl start sgs
and then set up proxy via nginx.