Open ionathan opened 3 weeks ago
When I wrote this, I was not aware of the fact that DataVerse uses different checksum implementations. I agree this should be fixed, but in order to do it properly, we should first find out the full picture of how DataVerse handles checksums.
Apparently, DataVerse can be configured to work with one of four hashing algorithms: MD5, SHA-1, SHA-256, and SHA-512 Source. There is an API route to check which one is in use, but it is only intended for upload, it does not give a guarantee about what checksums might be present on existing data. I therefore think our best bet is to iterate through a hard-coded list of keys until we find one that is present in the API response.
Description of the problem:
While trying to load a registry from a DOI of dataverse.nl, I realized that they use SHA1. In pooch the hash algorithm is "fixed" to md5.
Full code that generated the error
Full error message