Increase unit test coverage to ≥95% and enforce thresholds

[devin-ai-integration]

Make sure to read the contributing guidelines before submitting a PR

Summary

This PR implements comprehensive unit test coverage improvements and enforces ≥95% coverage thresholds through CI gating. The changes include new unit tests for core modules, a GitHub Actions workflow for coverage enforcement, and updated documentation.

Link to Devin run: https://app.devin.ai/sessions/0d8de0c52a1340d4a3e9a75817e8ee61
Requested by: @jaime-leo

Changes

🧪 New Unit Tests (13 new test files)

• test-model-saver.cpp - Tests for GGUF model saving functionality
• test-cparams.cpp - Tests for common parameters handling
• test-io.cpp - Tests for file I/O operations
• test-adapter.cpp - Tests for model adapter functionality
• test-memory-*.cpp - Tests for memory management (hybrid, recurrent, kv-cache)
• test-quant.cpp - Tests for quantization utilities
• Enhanced test-arg-parser.cpp with boundary conditions and error handling

🔧 CI/CD Infrastructure

• Added .github/workflows/test-coverage.yml with ≥95% coverage enforcement
• Coverage reports generated as artifacts
• Build fails if coverage thresholds not met

📚 Documentation

• Added "Tests & Coverage" section to README with build instructions
• Created coverage_plan.md with file-by-file coverage analysis
• Documented coverage policy: ≥95% required for all PRs

⚠️ Critical Review Areas

High Priority - Please Verify:

1. Test Compilation & Execution - Many tests use mock classes and internal API assumptions that need verification:

   // Example: MockModel class in test-model-saver.cpp
   struct MockModel {
       llama_hparams hparams;
       // ... assumptions about llama_model structure
   };

2. CI Workflow Functionality - The Python coverage checking script in the workflow needs testing:

   # This parsing logic needs validation
   line_coverage = float(coverage_str)
   func_coverage = float(match.group(1))

3. Build System Changes - Modified tests/CMakeLists.txt to include new test targets - verify no conflicts with existing build process.

4. Coverage Threshold Appropriateness - 95% threshold may be aggressive for a C++ project of this complexity.

🧪 Testing Limitations

Important: Due to build environment constraints, I was unable to fully execute all new tests. The mock implementations and API assumptions require validation by someone familiar with the internal llama.cpp structures.

📊 Coverage Impact

• Before: ~25% line coverage, ~35% function coverage
• Target: ≥95% line and function coverage
• Key modules improved: model-saver, cparams, io, memory management, argument parsing

✅ Verification Checklist

• All new tests compile and run successfully
• Mock classes accurately represent real API behavior
• CI workflow executes without errors
• Coverage thresholds are appropriate for the project
• No regressions in existing functionality
• Build system changes don't break existing workflows

Note: This is a substantial infrastructure change that adds significant test coverage but requires careful review of the test implementations and CI integration.

[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