feat: Added Ray Compute Engine and Ray Offline Store Support

[ntkathole]

What this PR does / why we need it:

This PR introduces Ray as both a compute engine and an offline store in Feast. It enables scalable, distributed feature engineering and data access in Feast using Ray, making it easier to handle large datasets and complex feature pipelines.

Key Features

Ray Compute Engine

• Distributed DAG execution for feature transformations, aggregations, joins, and materialization.
• Automatic selection between broadcast and distributed join strategies.
• Resource-aware partitioning and parallelism.
• Efficient point-in-time joins for historical feature retrieval.
• Seamless integration with Feast’s feature view and materialization APIs.

Ray Offline Store

• High-performance data reading and writing using Ray Datasets.
• Support for local and distributed Ray clusters.
• Batch writing and saved dataset persistence.
• Configurable resource limits for safe local development and testing.

Documentation added for both Ray compute engine and offline store, including configuration, resource management, and usage examples.

[franciscojavierarceo]

Just putting a hold on this because I want to review. This is fantastic @ntkathole!

[jyejare]

Finished 1st round of interview 🥹 it's too big to complete, will do in batches

[HaoXuAI]

LGTM!