v0.5.6

• SPARQL: DESCRIBE: do not describe values of blank node "variables".
• SPARQL: Fixes some bug in the parser related to spacing.
• SPARQL: Fixes evaluation of SERVICE clauses with unsupported custom functions.
• JSON-LD: fixes serialization of relative IRIs looking like keywords.
• RocksDB: reduce the number of copies in read operations.

v0.5.5

• JSON-LD: JSON-LD 1.1 is now implemented and enabled by default.
• JavaScript: new parse method that allows to parse a RDF file from a buffer, an iterator of buffers or and async iterator of buffers (including streams).
• oxrdf: new Dataset.quads_for_pattern method allowing to iterate only on quads matching a given pattern.

v0.5.4

• RDF/XML: support of the current 1.2 draft behind the rdf-12 cargo feature (directional language strings and triple terms).
• implementation of RDFC 1.0 canonicalization algorithm on the Graph and Dataset containers.

v0.5.3

Three SPARQL changes:

• support VERSION declaration.
• fixes parsing of HAVING when there are multiple conditions.
• compute ordering values for ORDER BY only once (allows ORDER BY RAND() to work properly).

v0.5.3-post.1

Retrigger the publishing CI after fixing an auth error

v0.5.2

• SPARQL: fixes evaluation of RDF 1.2 annotation syntax (e.g. SELECT * WHERE { ?s ?p ?o {| ?p2 ?o |}).
• Bump RocksDB to 10.7.5 (requires a C++ 20 compiler).
• Built Linux Python wheels now require Glibc 2.28.

v0.5.1

• spareval: Add QueryEvaluator.prepare method that allows to prepare a query then execute it.
  Injecting variable bindings and customizing the query dataset is possible on the prepared query.
  FROM and FROM NAMED clauses are taken into account by this API, opposite to the older QueryEvaluator.execute method.
• spargebra: Addstandard-unicode-escaping feature that unescape \u and \U sequences everywhere in the query, following to the SPARQL specification.
• CLI: oxigraph load now has a --non-atomic option to load data in a non-atomic way. Allows compaction to happen in the background during loading.
• NTriples/Turtle: fixes parsing of language tags with numbers like es-419.
• SPARQL: fixes evaluation of SELECT * FROM NAMED <ex:> WHERE { GRAPH ?g {} }: all named graphs of the underlying store returned and not only ex:.
• SPARQL: make parsing fail on duplicated variable names in VALUES like VALUES (?a ?a) {}.

v0.5.0

This release is dedicated to the memory of Gregg Kellogg, author of multiple specifications implemented by Oxigraph and great inspiration.

Major changes:

• Support for current W3C working drafts RDF 1.2 and SPARQL 1.2, hidden behind the rdf-12 or sparql-12 features.
  This replace the rdf-star feature.
  RDF 1.2 includes triple terms using a slightly different syntax as RDF-star. Support is disabled by default in the oxigraph crate but enabled by default in the Python and JS bindings and the CLI.
• oxigraph: the Store.start_transaction method that returns a new Transaction object.
  transactions can be committed with the commit method or rolled back by just dropping the Transaction object.
  Similarly, BulkLoader has a commit method to save the bulk loaded data.
  The Store::query method is now deprecated in favor of the SparqlEvaluator API

Read the changelog for all changes in this release.

v0.5.0-beta.5

• SPARQL query evaluation: adds a CancellationToken to easily cancel a running SPARQL query and use it in the new --timeout-s option of the CLI query command to set a timeout for the query.
• Adds split_file_for_parallel_parsing method to enable easier NTriples and NQuads parallel parsing from a file. Make use of it in a bulk loader parallel_load_from_file method that is also used by the CLI load command and the Python bulk_load function.
• The bulk loader Rust API has now a commit method that must be called to save the loaded data

v0.5.0-beta.4

• oxigraph: Transaction has now stronger lifetime bounds to ensure operations on a transactions are not running longer than the transaction itself. It also affects types like QuadIter or GraphNameIter that now carry a lifetime.
• spareval: expose QueryEvaluator::evaluate_expression to evaluate expressions