Use sha1 hash instead of md5 for cache filenames

[jflemer-ndp]

The MD5 hash is not allowed by FIPS 140-2. Allow running with FIPS enforcement enabled.

[chris-johnson-nrl]

Hi,

Thanks for the suggestion!

Hashing in GeoIPS is not carried out to positively identify items that have been identified as no or low security threats. Hashing algorithm selection in GeoIPS is not security related.

Several geostationary platforms (GOES-16, GOES-17, HIMAWARI, MET8, MET11) have sensors that require geolocation files (look up tables really) to properly place their data on the projection of the globe that is required for any particular product. In theory there could be one geolocation file per sensor. In practice that file is so large that accessing it slows down processing a lot when you're processing at large scales. Instead, GEOIPS makes dozens of much smaller geolocation files (corresponding to areas of interest) that are just indices to the location of the spatial information in the larger geolocation file. Each of these smaller geolocation files have names that include a two part unique identifier. Part one of the unique identifier is based on metadata specific to a sensor. Part two of the unique identifier is based on the definition of the area of interest (called 'sector file' in GeoIPS jargon). Hashing is used to create those unique identifiers so that these geolocation files can be found from various parts of the application.

Several files in GeoIPS use the python 2.7 built in hash method (abi reader, ahi reader, and a few others). Any change to the selection of hashing algorithm anywhere would need to accompanied by changes to the selection of the hashing algorithm everywhere.