feat: implement LM Studio runtime adapter (#17)

[himanshu231204]

Summary

Implement full LMStudioRuntime adapter that integrates with LM Studio's local server API. This PR closes issue #17.

Changes

• src/modeldock/adapters/runtimes/lmstudio.py - Full implementation of LMStudioRuntime adapter
• tests/unit/test_lmstudio_runtime.py - Comprehensive unit tests (31 tests)

Implementation Details

The adapter supports:

• Model discovery via /v1/models endpoint
• Interactive sessions using OpenAI-compatible API
• Model download via native API or UI guidance
• Configurable host and API key
• Comprehensive error handling with typed exceptions

Key Features

1. RuntimePort Compliance: Fully implements all RuntimePort methods
2. OpenAI Compatibility: Uses OpenAI SDK for inference via LM Studio's compatible API
3. Native API Support: Uses LM Studio's native API for model management
4. Configurable: Supports custom host and API key via constructor or environment variables
5. Graceful Error Handling: Typed exceptions with actionable hints

Testing

All tests pass:

• ✅ 183 unit tests passed (6 skipped)
• ✅ 31 LM Studio tests passed (100% coverage)
• ✅ 2 integration tests passed (2 skipped)
• ✅ 16 E2E tests passed

Static analysis:

• ✅ ruff - All checks passed
• ✅ mypy - No type errors
• ✅ bandit - No security issues

Usage Example

import modeldock as md

# List models from LM Studio
md.installed()

# Load a model
client = md.load("qwen/qwen3.5-9b", backend="lmstudio")

# Or use LM Studio explicitly
mgr = md.Manager(backend="lmstudio")
client = mgr.load("qwen/qwen3.5-9b")

Checklist

• Code follows project standards (Clean Architecture, SOLID principles)
• Comprehensive unit tests with >80% coverage
• All static analysis checks pass (ruff, mypy, bandit)
• No changes to core logic or other adapters
• Documentation updated in code comments
• Cross-platform compatibility maintained

[github-actions]

🎉 Congratulations @himanshu231204!

Your pull request has been successfully merged into main. 🚀

Thank you for contributing to OpenAgentHQ and helping improve the project.

We truly appreciate your contribution and hope to see you back with more amazing PRs!

Happy Open Sourcing! ❤️