rasterio.cache module

Rasterio caches HTTP responses using GDAL’s VSI CURL cache.

A global LRU cache of 16 MB shared among all downloaded content is enabled by default, and content in it may be reused after a dataset has been closed and reopened. Responses from FTP and HTTP servers, including cloud storage like AWS S3 and Microsoft Azure, are stored in this cache. Responses from HTTP servers used by Rasterio’s Python openers are not.

The size of the cache and which responses are stored can be controlled by GDAL’s CPL_VSIL_CURL_CACHE_SIZE and CPL_VSIL_CURL_NON_CACHED configuration options. For details See https://gdal.org/en/latest/user/configoptions.html.

rasterio.cache.invalidate(pattern)

Invalidate responses in Rasterio’s HTTP cache

Parameters:

pattern (str) – All responses beginning with this pattern will be invalidated. Responses served from a particular website can be invalidated using a pattern like “https://example.com”. Responses served from an S3 bucket can be invalidated using a pattern like “s3://example.com”, where “example.com” is the bucket name. Invalidation can be made more selective by appending path segments to the pattern. “s3://example.com/prefix” will invalidate only responses served for requests for objects in the “example.com” bucket that have a key beginning with “prefix”.

Return type:

None

rasterio.cache.invalidate_all()

Invalidate all responses in Rasterio’s HTTP cache

Return type:

None