857 large downloads stream

[entantoencuanto]

Related to #2774

✌️ What does this PR do?

• Tests different alternatives to download large query results, using streaming:
  • With an enumerator
  • Using ActionController::Live
• The streaming option is only available for json and csv formats

🔍 How should this be manually tested?

In addition to the normal endpoint there are two new endpoints that return the data in streaming:
For example:

• Normal request: /api/v1/data/data.{csv|json}?sql=select%20*%20from%20renfe_ave%20limit%202000
• Stream with enumerator request: /api/v1/data/stream_data.{csv|json}?sql=select%20*%20from%20renfe_ave%20limit%202000
• Stream using ActionController::Live: /api/v1/data/live_stream_data.{csv|json}?sql=select%20*%20from%20renfe_ave%20limit%202000

Does this PR changes any configuration file?

No

(Changes in these files might need to update the role in Ansible)

📖 Does this PR require updating the documentation?

No

[entantoencuanto]

Unfortunately the option of using PG::Result#stream_each seems to fail when another request is sent to the database because set_single_row_mode is disabled. I've chosen to use the normal query and iterate on the result using each or each_row (these functions don't support batches, to improve the response time we could explore a way to implement it)

With the use of each and each_row without streaming the response is slower (about 10%-20%)