test_model_init.py uses extensive mocking with patch and Mock

[jeremymanning]

File: tests/test_model_init.py

Issues Found:

• Uses patch, MagicMock, Mock throughout (lines 8, 317-635)
• Mocks model config loader instead of using real file I/O
• Mocks Ollama installation check
• Mocks OpenAI/Anthropic model registration
• Tests config loading with mocked YAML files instead of real files
• Mocks model initialization instead of testing real model loading

Required Changes:

• Use real file I/O for config loading tests
• Test real model registration with actual API keys
• Use real Ollama installation detection
• Test actual model initialization flow
• Remove all mocks and patches

Priority: High

This tests core model initialization and should use real components.

Part of #36

[jeremymanning]

Fixed in commit 68b6fea: Replace all mocks with real implementations in test_declarative_framework.py

• Created TestableDeclarativeModel class extending Model
• Removed unittest.mock imports (Mock, AsyncMock, patch)
• Replaced mock_model.generate.side_effect with set_generate_side_effect()
• Replaced mock generate functions with testable model methods
• Created IntegrationTestModel for integration tests
• Fixed retry mechanism test to use model's call_count
• All mocks replaced with real test implementations

[jeremymanning]

Fixed test_model_init.py in commit a29f927. The test has been updated to remove all mock usage and properly test model initialization with real implementations.