Expose any GDAL recognized raster file as a HTTP accessible on-the-fly COG (Cloud Optimized GeoTIFF)
The on-the-fly COG file is not materialized to disk, can be of arbitrary size with little RAM consumption and can be accessed in a piecewise way with HTTP GET Range header.
Quality: proof-of-concept
What remains to be implemented:
- expose metadata in GDAL_METADATA tag
- allow compression
- expose overviews when they exist
Contact me if you need a different license than AGPL.
- Python 3
- GDAL native library
- GDAL Python bindings
Serve: cogserver.py my_gdal_raster --port 8080
Consume: gdalinfo /vsicurl/http://localhost:8080/my.tif