Add a migration to generate the "results" table

Possible columns:

* id
* chain_id
* round_id
* alghoritm (string or enum for `cluster-matching`, `linear-qf`, and otherw in the future)
* file_url
* created_at
* started_at // to have stats on calculations performance
* finished_at // to have stats on calculations performance + can be null in case it's running now
* generated_at // with the last time it was generated. if all the jobs are finished it will have the same value of finished_at
* last_requested_at

(think about the indexes needed for the query we will use in the next tasks)