Create w3ids for sqlite downloads

[cmungall]

We use S3 URLs a lot in OAK code and docs:

https://github.com/INCATools/ontology-access-kit/search?q=https%3A%2F%2Fs3.amazonaws.com%2Fbbop-sqlite%2F

We should create w3id.org/semsql/ to have a PURL for the downloads

OR we include this with OBO PURLs

[jamesaoverton]

I've often thought about doing something like this, but I always got stuck with worries about maintenance, stability, format/compression, versioning the ontologies, versioning the schema/software used to generate the SQL. Happy to discuss.

[cmungall]

Definitely good to reason through these:

• for versioning, currently there is no guarantees that specific versionIRIs are available at the moment, but they are versioned by import date
• we release both compressed and uncompressed but plan to switch to the former, see syncing an upstream gzip file with an expanded local version cthoyt/pystow#45
• I would like to have a core set of views that become part of a 1.0.0 standard - there can always be convenience views later
• the core version for the base statements table I propose to be just be subject, predicate, object, value, language, datatype, with stanza being optional