on.Watch

Service for aggregating output from multiple time-lapsed webcams

still a work in progress.

Client Setup

The listening server will accept image posts from any number of clients. Clients must make all requests using a valid api-token and a valid camera id-name, both created on the listening server. Client will check if the listening server is accepting images for the given camera before capturing an image and uploading it. No images will be captured if the listening server says collection for the given camera is disabled.

• client code requires python3^
• create a virtualenv under watch/cam and pip install -r -client-requirements.txt
• install fswebcam, sudo apt-get fswebcam
• copy .env_copy to .env and update with your host/token/cam-id-name
• copy cam/capture_copy.sh to cam/capture.sh and update PROJ_PATHs
• put the above in a cron job to run as frequently as you desire

Server Setup

This service is setup as a master server listening for image posts on registered cameras.

• provision a server to listen
• install python3.6^
• install postgres, nginx
• in a virtaulenv pip install -r requirements.txt
• copy gunicorn_copy.conf to /etc/init/gunicorn.conf and update USER and PATH_TO_PROJ_BASE
• initctl reload-configuration to pull in our gunicorn commands (see notes in comments for upstart installation)
• copy nginx_copy.conf to /etc/nginx/sites-available/watch and update DOMAIN, PATH_TO_PROJ_BASE, PROJ_APP_NAME
• symlink /etc/nginx/sites-available/watch to /etc/nginx/sites-enabled/watch, may need to remove sites-enabled/default
• setup your ssl certs with letsencrypt, see letsencrypt.info
• sudo service nginx restart, sudo service gunicorn start
• add cameras and api tokens using either the django-admin interface or the admin management command ./manage.py admin --help