Do not reset the dev env when booting docker

The current web container's command destroys anything you might have in
your local DB from a previous session, assuming you always want start
from a clean environment. This is hardly the case and makes
`docker-compose up` take quite long. What if you just stopped containers
temporally while developing?

This changes the approach to not assume anything. If you need to install
a new gem or reset your DB just run the commands you would without
docker. You can run anything you want with `docker-compose run web bundle exec
<rails/rake command>` anyway.

For someone setting things for the first time, the `Dockerfile` process
still installs all dependencies.

docker-compose.yml

@@ -28,10 +28,7 @@ services:
       ADMIN_PASSWORD: ofn123
       OFN_DB_HOST: db
     command: >
-      bash -c "(bundle check || bundle install) &&
-               wait-for-it -t 30 db:5432 &&
-               bundle exec rake db:reset &&
-               bundle exec rake db:test:prepare ofn:sample_data || true &&
+      bash -c "wait-for-it -t 30 db:5432 &&
                rm -f tmp/pids/server.pid &&
                bundle exec rails s -p 3000 -b '0.0.0.0'"