refactor: split _markitdown.py into modular components

[t3tra-dev]

Description

This PR addresses the growing complexity of _markitdown.py by splitting it into smaller, more focused modules. The changes improve code organization and maintainability while preserving all existing functionality.

Changes

• Created a new converters/ package to house different converter implementations
• Split converters into logical groups (document, web, media, text, archive)
• Moved core MarkItDown class functionality to core.py
• Separated exception classes into exceptions.py
• Updated imports and tests to reflect new structure

Testing

• All existing tests pass without modification
• Verified no functionality changes

Implementation Details

The refactoring follows these principles:

1. Single Responsibility: Each module handles a specific type of conversion
2. Open/Closed: New converters can be added without modifying existing code
3. Interface Segregation: Clear base class and consistent converter interface
4. Dependency Inversion: Core MarkItDown class depends on abstractions

Migration Notes

This is a non-breaking change as all public APIs remain unchanged. Internal imports are updated to reflect the new structure.

[t3tra-dev]

@microsoft-github-policy-service agree