Add meta agent option and add continuous action space compatibility to async

[Mark2000]

Description

Closes #296

Allows for meta-agents to be created that concatenate the observation and action spaces of the satellites.

Also fixes some bugs in the async connectors for continuous action spaces.

⚠️Note to reviewer⚠️ Primarily concerned with changes to gym.py. RLlib utilities are not robustly tested/supported.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)

How should this pull request be reviewed?

• By commit
• All changes at once

How Has This Been Tested?

Added tests for meta-agents.

Future Work

None.

Checklist

• I have performed a self-review of my code
• I have commented my code in hard-to-understand areas
• I have made corresponding changes to the documentation and release notes
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works

[Copilot]

Pull Request Overview

This PR adds meta-agent support to the ConstellationTasking environment, allowing multiple satellites to be grouped under a single agent that controls them with concatenated observation and action spaces. Additionally, it fixes bugs related to continuous action space handling in async connectors and modernizes the PyTorch policy loading functionality.

• Added meta-agent groupings functionality with observation/action space concatenation
• Fixed continuous action space compatibility issues in async connectors
• Updated PyTorch policy loading to use modern RLModule checkpoint format

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

File	Description
tests/integration/test_int_full_environments.py	Added test environment with meta-agent groupings and updated test to cover both regular and meta-agent scenarios
src/bsk_rl/utils/rllib/discounting.py	Fixed continuous action space handling with new utility functions and improved NO_ACTION detection
src/bsk_rl/utils/rllib/init.py	Modernized PyTorch policy loading to use RLModule checkpoint format instead of legacy approach
src/bsk_rl/gym.py	Added core meta-agent implementation with observation/action space concatenation and satellite grouping logic

Comments suppressed due to low confidence (1)