Twarc

Hot on the heels of v2.1.0 this bugfix release corrects the `twarc2 flatten` sub-command to ensure that data is both flattened and output as line oriented JSON.

twarc2 search blacklivesmatter | twarc2 flatten | jq .text

The `twarc.expansions.flatten()` function has been updated to always return a list of tweets, and `twarc.expansions.ensure_flattened()` can be used to make sure data has been flattened already when processing tweet data. Since it is designed for use in twarc plugins and other pieces of software that need to work with tweets it is also available for import from twarc:

from twarc import ensure_flattened

In addition this release also includes `twarc conversation` for retrieving tweets from a a particular conversation thread.

This bugfix release irons out some wrinkles that have been discovered during usage:

* Fix for handling search responses with missing `data` stanza but that contain a next token for another page of results. 464
* Stream diagnostics now to go stderr to not interfere with JSON being written to stdout. 456
* twarc can be run from the command line as a module now: e.g. `python -m twarc2 search barackobama` 455
* twarc2 command help text indicates times are UTC

Bugfix for using `--since-id`.

Added the HTTP status code when emitting a message about unparseable JSON.

This release adds the `--max-results` option to the search command to override the default of 100 for search/recent and 500 for search/all. See 449 for the details on why this is sometimes needed.