add a kafka_stream_to_value helper function in serializers.py

similar to kafka_to_value but takes a stream as input and return a msgpack.Unpacker object on which the user can iterate.

also make the stream journal client accept a string as output_stream config entry and support the "-" value (for stdout).

This is useful to be able to dump a storage content is a journal-compat file, eg. using a replayer with the following config file:

storage: cls: memory journal_writer: cls: stream output_stream: dump.msgpack journal_client: xxx

or using a backfiller with something like:

storage: cls: postgresql db: xxx journal_writer: cls: stream output_stream: dump.msgpack

---

Migrated from D7933 (view on Phabricator)