Images
- Images are served using a IIIF Image API 2.0-compliant server, Cantaloupe, fed JPEG 2000 images that have been compressed from source TIFFs + JPEGS.
- Image metdata (which comes from a variety of semi-structured sources) is fetched at build time from an S3 bucket where PMA staff has collected hi-resolution TIFF files, and derived from HTTP URLs where necessary in the case of the JSON-based Navigart and PMA TMS APIs.
- Compressed images are discovered with a SPARQL query against the integrated linked data and then compressed in a batch process.
Special considerations
- Unexpected TIFF compression profiles
All TIFF files are forced uncompressed during normalization and bit depths are normalized to 8bpp.
Colorspaces that are not sRGB are converted during normalization.
Multipage TIFF files should be pre-processed using the data pipeline's included script (accessible from pipeline workflows stage-images
).
References