Add multi-threaded GDAL test harness for GeoTIFF and ERS raster formats

[Claude]

Implements a comprehensive test harness for multi-threaded raster I/O operations using GDAL with Python, supporting both GeoTIFF and ERS grid formats.

Core Components

• Thread-safe GDAL wrappers: ThreadSafeRasterReader and ThreadSafeRasterWriter classes handle concurrent raster access using thread-local datasets and locks
• Thread coordination: RasterThreadManager provides block-based parallelization and benchmarking utilities
• Test harness: RasterTestHarness class orchestrates multi-threaded read/write tests with configurable block sizes and thread counts

Usage Example

from raster_threading import RasterTestHarness

# Test GeoTIFF operations with 4 threads
harness = RasterTestHarness(num_threads=4, block_size=256)
results = harness.test_geotiff_read_write(
    width=1024, 
    height=1024, 
    bands=3
)

# Test ERS grid operations
results = harness.test_ers_read_write(
    width=1024,
    height=1024,
    bands=1
)

# Compare single vs multi-threaded performance
comparison = harness.compare_single_vs_multi_threaded(
    width=2048,
    height=2048,
    format='GTiff'
)

Key Features

• Per-thread dataset instances to avoid GDAL threading conflicts
• Configurable block-based tiling for parallel processing
• Built-in benchmarking and performance comparison
• Data integrity validation (checksums, pixel-by-pixel comparison)
• Support for both formats with driver-specific optimizations

Original prompt

Design a test harness to use gdal via python to read and write to rasters in multi-threaded fashion. Use both geotiffs and ers grids in your work.