[🐋 narwhals] create stub modules, classes, and functions for narwhals integration

[cosmicBboy]

This PR adds stub modules, classes, and functions for the narwhals integration #1894.

Claude code summary:

Created Files and Structure:

1. Entry Point Module (pandera/narwhals.py)

• Main public API entry point for narwhals integration
• Imports and exports all key components (Column, DataFrameSchema, DataFrameModel, etc.)
• Registers narwhals backends

2. API Components (pandera/api/narwhals/)

• types.py: Core type definitions (NarwhalsData, CheckResult, type aliases)
• components.py: Column schema validation component
• container.py: DataFrameSchema for full DataFrame validation
• model.py: DataFrameModel for class-based schema definition
• utils.py: Utility functions for schema extraction and validation depth
• model_config.py: Configuration for DataFrameModel

3. Backend Implementation (pandera/backends/narwhals/)

• base.py: Base backend functionality with NarwhalsSchemaBackend
• register.py: Backend registration system
• checks.py: Check backend for validation functions
• components.py: Column validation backend
• container.py: DataFrame validation backend
• builtin_checks.py: Built-in validation checks (equal_to, greater_than, etc.)

4. Engine Layer (pandera/engines/narwhals_engine.py)

• Core narwhals engine and data types
• Type conversion utilities
• DataType classes for common types (Int64, Float64, String, etc.)

5. Typing Support (pandera/typing/narwhals.py)

• Generic type annotations for DataFrame[T], LazyFrame[T], Series[T]
• Pydantic integration for schema validation
• Format conversion utilities

Key Features:

• Import Entry Point: import pandera.narwhals provides the full API
• Modular Architecture: Follows the same patterns as polars integration
• Type Safety: Strong typing with narwhals generics
• Backend Registration: Automatic registration of validation backends
• Extensible Checks: Built-in validation functions with decorator system
• Pydantic Integration: Support for Pydantic v2 validation
• Format Support: Placeholder infrastructure for various data formats

The implementation maintains the same API surface as the polars integration while adapting to
Narwhals' DataFrame-agnostic approach. All modules contain placeholder implementations that follow
the correct patterns and interfaces, ready for full implementation.