W3c rdf sparql conformance#7
Conversation
…gnore to exclude Emacs backup files.
….1 SELECT/ASK tests with UPDATE/DELETE operations as unsupported with focus on read-only HDT storage backend. The changes include commits to upstream forks of Oxigraph and HDT. Performed aggressive code review and fix using OpenAI gpt-5.3-codex xhigh with a focus on reliability and user data protection.
|
related HDT changes: DeciSym/hdt#1 |
| /// - `canonical_path` exists and is canonicalized. | ||
| /// - `graph_iri` is syntactically valid. | ||
| #[derive(Debug, Clone)] | ||
| pub(crate) struct ResolvedGraphPath { |
There was a problem hiding this comment.
in general, I have probably been overusing pub vs pub(crate) wrt best practices, as it makes it easier to reuse code in our downstream dependencies. May need to put a TODO on one of our boards to redefine any of the new code pub(crate)'s as we attempt to import these changes
There was a problem hiding this comment.
Can be done as a follow-up, but we should probably add metadata mutator for the file_iri that is generated and stored in the header. For instance:
<file:///tmp/.tmpxLjDTC/.tmpm9dGd8.nt> <http://purl.org/HDT/hdt#statisticalInformation> _:publicationInformation .
since this is generally a file that is either converted and/or combined from a series of sources, it's pointing to a random tempfile iri. Ideally we could set this to an IRI in the actual dataset as a means of potential query optimization or data linkage via metadata
| entailment: query::EntailmentMode, | ||
| #[clap(long, default_value_t = false)] | ||
| /// Print query plan JSON to stderr for each query | ||
| debug_query_plan: bool, |
There was a problem hiding this comment.
I wonder if we can map this to the top-level verbosity flag rather than tracking two levels of debug log messages
… to pass checks for dead code accurately.
…ve to Debian-based Linux distributions.
…d to minimize triples loaded into memory at once
2 participants
No description provided.