Support CSV in `_bulk`

[xeraa]

Description

Importing CSV programmatically (without the use of Kibana) is surprisingly complicated; and often requires a fallback to miller, awk, or similar.

A feature that could support curl -H "Content-Type: text/csv" -XPOST "localhost:9200/myindex/_bulk" --data-binary @- would be very useful. And it could be limited in scope by design: only target a single index (myindex in this example) and require the first row to define the field names. But it should be possible to target an ingest pipeline like any other _bulk request to make changes to the structure as needed.

For reference, this is quite a common feature in other datastores:

• https://clickhouse.com/docs/en/integrations/data-formats/csv-tsv
• https://www.postgresql.org/docs/current/sql-copy.html
• https://duckdb.org/docs/data/csv/overview.html

[elasticsearchmachine]

Pinging @elastic/es-distributed (Team:Distributed)