Pynonymizer

Added
- Added `fake_update` Strategy feature: `sql_type`: Choose an sql datatype to cast to when anonymizing values. The new value will be cast to the chosen datatype.
Fixed
- Fixed PostgreSQL column escaping. Keywords like `SELECT`, `FROM` or `WHERE` can now be used in strategy files.

Added
- Added new table strategy `DELETE`, which should delete _with_ checks (e.g. foreign keys) on most providers.

Added
- Added ability to pipe output to/from pynonymizer from stdout, using `-` in place of the input/output arguments. This functionality is available for mysql and postgres providers.

This means you can now use pynonymizer as part of a pipeline with other tools, e.g.

mysqldump [...] | pynonymizer -i - -o - | aws s3 cp - s3://bucket/aws-test.tar.gz

Changed
- Changed default logging output to stderr. This is to facilitate stdin/out being used for data.
Removed
- Removed production logging feature in favour of stderr/out logging. Logging to files will no longer by considered pynonymizer's concern.

Changed
- Changed package metadata to improve PyPI presence.

Fixed
- Fixed an incorrectly labelled version string: `1.11.10`. Re-releasing under this version.

Changed
- Changed mysql provider to include an arbitrary delay after anonymize_db, to prevent
interference with transactional dump `mysqldump` calls.
Removed
- Removed the database connection pre-test as its use suggests conditions about the database
that are not present in all circumstances with all providers. Pynonymizer will no longer test a connection as part of a dry-run.