Add YAML configuration file support to CLI tools

[devin-ai-integration]

Make sure to read the contributing guidelines before submitting a PR

Summary

This PR adds YAML configuration file support to llama.cpp CLI tools (llama-cli and llama-server) while maintaining full backward compatibility with existing command-line arguments.

Link to Devin run: https://app.devin.ai/sessions/09992668d325454881f7da8b8be9ba3e
Requested by: @jaime-leo

Key Features

• New --config argument: Accept YAML configuration files for all CLI tools
• Complete parameter coverage: Support for model, context, sampling, threading, GPU, and I/O parameters
• Proper precedence: CLI arguments override config file values
• Robust error handling: Graceful handling of invalid YAML files and missing files
• Optional dependency: YAML support is optional and automatically disabled if yaml-cpp is not found
• Comprehensive testing: New tests verify YAML functionality and backward compatibility

Changes

Core Implementation

• CMake integration: Added LLAMA_YAML_CPP option with automatic yaml-cpp detection
• YAML parsing: New common_params_load_from_yaml() function with comprehensive parameter mapping
• Argument integration: Added --config argument that loads YAML when processed
• Parameter structure: Added config_file field to common_params

Testing & Examples

• Test coverage: Added YAML-specific tests in test-arg-parser.cpp
• Example configs: Created sample YAML files for main CLI and server tools
• Error testing: Verified handling of invalid YAML and missing files

Usage Example

# config.yaml
model: "models/llama-2-7b-chat.gguf"
ctx_size: 4096
threads: 8
temperature: 0.7
top_k: 40
conversation: true
antiprompt:
  - "User:"
  - "Human:"

# Use config file
./llama-cli --config config.yaml

# CLI args override config values
./llama-cli --config config.yaml --threads 16 --temperature 0.9

Review Focus Areas

⚠️ Critical areas for review:

1. CMake dependency detection (CMakeLists.txt lines 194-216): The yaml-cpp finding logic has multiple fallback paths - verify it works across different platforms and installation methods

2. YAML parameter mapping (common/arg.cpp lines 82-239): Extensive manual mapping between YAML keys and parameter fields - check for accuracy and completeness against CLI argument equivalents

3. Build system impact: Verify the optional dependency doesn't break builds on systems without yaml-cpp

4. Precedence correctness: Confirm CLI arguments properly override config file values in all scenarios tested

5. Error handling robustness: Review exception handling for YAML parsing errors and file I/O issues

Testing

• ✅ All existing tests pass (backward compatibility maintained)
• ✅ New YAML functionality tests added and passing
• ✅ CLI override precedence verified
• ✅ Error handling for invalid/missing files tested
• ✅ Manual testing with both llama-cli and llama-server

Backward Compatibility

All existing CLI usage patterns continue to work unchanged. The --config argument is purely additive and does not affect any existing functionality.

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring