Ormar

🐛 Fixes

* Fix deepcopy issues introduced in pydantic 1.9 [685](https://github.com/collerek/ormar/issues/685)

✨ Breaking Changes

* Dropped support for python 3.6
* `Queryset.get_or_create` returns now a tuple with model and bool value indicating if the model was created (by MojixCoder - thanks!) [554](https://github.com/collerek/ormar/pull/554)
* `Queryset.count()` now counts the number of distinct parent model rows by default, counting all rows is possible by setting `distinct=False` (by erichaydel - thanks) [588](https://github.com/collerek/ormar/pull/588)

✨ Features

* Added support for python 3.10

🐛 Fixes

* Fix inconsistent `JSON` fields behaviour in `save` and `bulk_create` [584](https://github.com/collerek/ormar/issues/584)
* Fix maximum recursion error [580](https://github.com/collerek/ormar/pull/580)

✨ Features

* Add `queryset_class` option to `Model.Meta` that allows you to easily swap `QuerySet` for your Model (by ponytailer - thanks!) [538](https://github.com/collerek/ormar/pull/538)
* Allow passing extra `kwargs` to `IndexColumns` that will be passed to sqlalchemy `Index` (by zevisert - thanks) [575](https://github.com/collerek/ormar/pull/538)

🐛 Fixes

* Fix nullable setting on `JSON` fields [529](https://github.com/collerek/ormar/issues/529)
* Fix bytes/str mismatch in bulk operations when using orjson instead of json (by ponytailer - thanks!) [538](https://github.com/collerek/ormar/pull/538)

✨ Features

* Add `post_bulk_update` signal (by ponytailer - thanks!) [524](https://github.com/collerek/ormar/pull/524)

🐛 Fixes

* Fix support for `pydantic==1.9.0` [502](https://github.com/collerek/ormar/issues/502)
* Fix timezone issues with datetime [504](https://github.com/collerek/ormar/issues/504)
* Remove literal binds in query generation to unblock postgres arrays [/tophat/ormar-postgres-extensions/9](https://github.com/tophat/ormar-postgres-extensions/pull/9)
* Fix bulk update for `JSON` fields [519](https://github.com/collerek/ormar/issues/519)

💬 Other

* Improve performance of `bulk_create` by bypassing `databases` `execute_many` suboptimal implementation. (by Mng-dev-ai thanks!) [520](https://github.com/collerek/ormar/pull/520)
* Bump min. required `databases` version for sqlalchemy 1.4 to `>=5.4`.

✨ Features

* Add ability to pass `comment` to sqlalchemy when creating a column [485](https://github.com/collerek/ormar/issues/485)

🐛 Fixes

* Fix `LargeBinary` fields that can be nullable [409](https://github.com/collerek/ormar/issues/409)
* Make `ormar.Model` pickable [413](https://github.com/collerek/ormar/issues/413)
* Make `first()` and `get()` without arguments respect ordering of main model set by user, fallback to primary key (asc, and desc respectively) [453](https://github.com/collerek/ormar/issues/453)
* Fix improper quoting of non-aliased join `on` clauses in postgress [455](https://github.com/collerek/ormar/issues/455)

🐛 Fixes

* Hot fix for validators not being inherited when parent `ormar` model was set [365](https://github.com/collerek/ormar/issues/365)