Write an Elasticsearch import script

[starakaj]

The script should:

• Know the timestamp of the CSV with which the Elasticsearch db is currently synced
• Take the timestamp of a new CSV as a target
• Diff the previous CSV agains the target CSV, using the existing Python script
• If there is no previous CSV, diff against a blank CSV (right?)
• Update the Elasticsearch db to reflect the state described by the target CSV