Hikaru

A maintenance release that formalizes Hikaru's compatibility with black 21.7b0.

This release adds support for the newest K8s Python client, 1.17. You can now select which version of the client to use when creating Hikaru objects yourself, when when Hikaru creates them for you when parsing YAML.

NOTE: this is not a drop in replacement, as some import changes have been necessary to keep existing behaviours in IDEs while adding support for multiple releases. See the ReleaseNotes for the change you need to make. Happily, you can make the changes in code that currently works with Hikaru 0.5x and then simply upgrade Hikaru. If you have already import specific releases of models then you won't have any changes to make at all.

Additionally, the release allows the use of the newest version of the 'black' code formatter.

This is a patch release that supports the new version of the 'black' code formatter and fixes a bug in processing DaemonEndpoint down inside NodeStatus. This patch does update the name of that class's 'port' attribute to 'Port' which matches the official swagger spec.

The main feature of this release is the addition of CRUD-style methods to many of the HikaruDocumentBase subclass such as Pod, ReplicaSet, and so on. This provides for simpler access to the corresponding underlying Kubernetes operations while still making available the full power of those operations. Classes that support update() on their instances are now also context managers, and when the context scope closes the update happens automatically. You can also establish a rollback to the initial state of the object if an error arises during the with block. Further requested features and bug fixes are part of this release; please see the release notes for details.

* Integrated the official Kubernetes Python client; now you can make Kubernetes calls through Hikaru classes and methods.
* Added a subclass registration mechanism that allows you to create subclasses of Hikaru classes and have Hikaru use them when needed (some limitations apply).
* Reorganized the model subpackage to pave the way for later support of multiple versions of the Kubernetes client.
* Improved the output of dff() to give more detailed reporting of what changed between two versions of a Hikaru object.
* Changed code annotation generation approach so cyclic structures are now supported directly; no more workarounds for items like JSONSchemaProps.
* Added a 'documents' module to every model version package; this contains only references to top-level Hikaru classes.
* Fixed known bugs.
* Updated all docs.

- Added support a new two new styles of generated code from `get_python_source()`: the 'black' style, using the 'black' formatter, and None, which outputs syntactically correct Python but with no formatting at all (this is the fastest generation option and is good if the code is going to be dynamically executed).
- New `get_type_warnings()` method on HikaruBase objects; compares actual values with the types currently populating an instance, and looks for required values that are missing. Generates a list of warning records for any problems found.
- New `diff()` method of HikaruBase; compares to object hierarchies and generates difference records indicating where they are different.
- Removed dead code.
- Improved and documented all exceptions that are raised.
- Added support for round-tripping between YAML, Python objects, Python source, JSON, and Python dicts. You can now start with any of these, move between them, and get back the original representation.
- Raised testing coverage to 99% overall.
- Documentation updates; includes a section on patterns and recipes.