Convert cluster environment methods to properties

[awaelchli]

Proposed refactor

Convert the getter methods in the cluster environments to properties (read-only).

Motivation

The cluster environment is meant to be a look-up table, read-only

Pitch

Convert the getter methods in the cluster environments to properties (read-only).

PRO:

• Consistent with the main_address and main_port methods which are already properties.
• More pythonic?

CON:

• Implementing a cluster environment requires overriding all methods/properties in the base abstract class. It can easily be forgotten by to add the @property decorator, leading to an error message a beginner might not understand.
• Backward compatibility is harder to get right. On the metaclass level, we will have to wrap all non-property methods into properties when a developer uses an old environment implementation or implements one but doesn't add the decorator.

Additional context

Continuation from #6303

---

If you enjoy Lightning, check out our other projects! ⚡

• Metrics: Machine learning metrics for distributed, scalable PyTorch applications.

• Lite: enables pure PyTorch users to scale their existing code on any kind of device while retaining full control over their own loops and optimization logic.

• Flash: The fastest way to get a Lightning baseline! A collection of tasks for fast prototyping, baselining, fine-tuning, and solving problems with deep learning.

• Bolts: Pretrained SOTA Deep Learning models, callbacks, and more for research and production with PyTorch Lightning and PyTorch.

• Lightning Transformers: Flexible interface for high-performance research using SOTA Transformers leveraging Pytorch Lightning, Transformers, and Hydra.

cc @justusschock @awaelchli @rohitgr7 @ananthsub

[stale]

This issue has been automatically marked as stale because it hasn't had any recent activity. This issue will be closed in 7 days if no further activity occurs. Thank you for your contributions, Pytorch Lightning Team!