Add comprehensive test suite for core modules and LLM clients

[koriyoshi2041]

Summary

This PR adds comprehensive unit tests for previously untested core modules and LLM client implementations, addressing the "Add better unit tests" item in CONTRIBUTING.md.

Changes

tests/test_comms_utils.py (34 tests)

• LMRequest/LMResponse dataclass serialization
• Socket send/recv protocol testing
• Batched request handling

tests/clients/test_openai.py (23 tests)

• Client initialization with various configurations
• Completion with string and message list prompts
• Usage tracking and accumulation
• Async completion support
• Integration tests (skipped without API key)

tests/clients/test_anthropic.py (27 tests)

• System message extraction and handling
• Message list format conversion
• Usage tracking per model
• Async completion support

tests/test_rlm_logger.py (18 tests)

• Logger initialization and file creation
• Metadata and iteration logging
• JSONL format validation
• Complex data serialization

tests/test_lm_handler.py (29 tests)

• Client routing by depth and model name
• Server lifecycle management
• Context manager protocol
• Full socket communication integration

Testing

uv run pytest tests/test_comms_utils.py tests/clients/test_openai.py \
    tests/clients/test_anthropic.py tests/test_rlm_logger.py \
    tests/test_lm_handler.py -v
# Result: 125 passed

Code Quality

• uv run ruff check: All checks passed
• uv run ruff format --check: All files formatted

Notes

• Integration tests are marked with @pytest.mark.skipif and require API keys
• Uses existing MockLM class for unit testing without API calls
• Follows existing test style from test_local_repl.py, test_parsing.py, test_gemini.py