Heroku does not support our current stop-line refresh process

[tkimia]

Right now, running a set-up-transitland through the heroku CLI results in a SQL error. This is because the SQL provider that we use caps us at 18,000 requests per hour. We need to re-do the rake task in a way that does not make so many SQL calls

Detailed Description

We use a lot of requests because we use the destroy_all function on stops, which has more than a 11 thousand stops.

Context

This change is not important in the short term, however, if we ever need to re-grab the data from transitland, our current process would bring the site down for multiple hours.

Possible Implementation

We might need to replace destory_all with delete_all. This only uses 1 query.