- Fixes flush and compaction on the GSPO index. It might improve Oxigraph performances and storage space.
- SPARQL: fixes some optimizations in presence quoted triples with nested variables.
- SPARQL profiler: adds EXISTS operation to the explanation and profiling tree.
- Upgrades RocksDB to 8.1.1.
- Rust: adds
Store.explain_query_opt
method that allows to get an explanation of the evaluation with profiling statistics. - Server: adds explanation and profiling to the
query
action (--explain
,--explain-file
and--stats
options). - Python: implements the
__match_args__
attribute on terms (literals, triples, quads...) to allowmatch
operator usage. - Server: adds the
--cors
option to theserve
actions to allow cross-origin resource sharing.
- SPARQL: fixes evaluation of empty aggregation without GROUP BY: aggregators now return their default value (0 for COUNT...) and not an empty row.
- SPARQL: fixes parsing of decimal with more than 19 digits in the fractional part.
- Server docker image: switch to the smaller distroless base images.
- Bulk loader: by default only uses 2 concurrent threads and around 2GB of RAM.
- Server load: progress is now printed to stderr much more regularly.
- Read only and secondary RocksDB storage. Allows to open the data as read-only or to follow a primary instance.
- Server: adds multiple commands:
serve-secondary
andserve-read-only
to serve the HTTP server in secondary and read-only modes.dump
to dump the database content to a file.query
andupdate
to execute SPARQL query and updates.backup
to do a database backup.optimize
to optimize the database storage.
- Server: adds
format
andgraph
options to theload
command in order to select the loaded data format and the target graph. - Python:
Store.extend
andStore.bulk_extend
methods. - Python: allows to pickle basic data model classes.
- Upgrades RocksDB to 8.0.0.
- Server: listening to localhost now properly listen to both IPv4 and IPv6.
- RDF/XML and XML parser results: avoid an ever growing buffer.
- JSON SPARQL results: allows the "head" key to be at the end of the document.
- TSV SPARQL results: properly quote
\t
and print trailing line jumps. Term::from_str
: fixes parsing of blank nodes in object position of quoted triples.QuerySolution
: implementsEq
andDebug
.- JS: Reduces WASM build size.
- OxRDF: fixes
Graph
andDataset
serialization to output proper N-Triples and N-Quads.
- Fixes serialization of times with a decimal number of seconds lower than 10 (now
01:01:01.01
and not01:01:1.01
). - Turtle and TriG: fixes parsing for prefixes named after keywords (
prefix
,base
,graph
,true
andfalse
).
From<NamedOrBlankNode>
forGraphName
(and similarly for*Ref
).- Prebuilt Python wheels for Linux with musl libc.
- Fixes TriG serialization.
QueryDataset
is now properly exposed in the public Rust API.
- SPARQL: Implementation of the
LATERAL
operator following SPARQL SEP-0006. Support is behind thesep-0006
feature inspargebra
and enabled by default in Oxigraph. - SPARQL: Implementation of the
ADJUST
function following SPARQL SEP-0002. Support is behind thesep-0002
feature inspargebra
and enabled by default in Oxigraph. - Rust: There is a new stand-alone crate
oxsdatatypes
implementing Rust structs for the common XML schema datatypes. It was part of the Oxigraph crate and it might be useful for other SPARQL or XPath implementations. - Rust: The
oxigraph
crate can now be compiled forwasm32-wasi
with the basic in-memory backend but without RocksDB.
- SPARQL: The property path evaluator was sometime emitting duplicates when evaluating disjunctive patterns (
a|b
). It is now fixed. - SPARQL: If written explicitly in the SPARQL query, the regular expressions are now compiled once and not for each row.
- SPARQL: Property path evaluation with both start and end variables bound has been optimized.
- SPARQL: Casts to
xsd:decimal
fromxsd:float
andxsd:double
now properly fails on overflow instead of overflowing the internal 128-bits representation. This follows XPath casting rules. - Rust: The minimal supported Rust version is set at 1.60 and enforced using the CI.
- Python: Local builds will now target the specific Python version instead of abi3. abi3 wheels are still released on Pypi alongside new Python-version specific builds for Linux+GNU.
- SPARQL: Fixes a panic when the estimated upper bound of the results size was overflowing a
usize
. - Python: Uses
typing.IO
in Python stubs instead of narrower interfaces. - Upgrades RocksDB to 7.9.2,
quick-xml
to 0.27 andpyo3
to 0.18.
- SPARQL: Property path with unbound graph are now evaluated.
For example,
SELECT * WHERE { GRAPH ?g { ?s ex:p+ ?o } }
now works. - SPARQL: The query optimizer is now fuzzed for better testing.
- SPARQL: Evaluation of zero-length paths with both start and end unbounded now return only terms defined in the current graph but not terms only defined in the query.
For example, in
SELECT * WHERE { VALUES ?s { 1 } { ?s ex:p? ?o } }
the literal1
won't be returned anymore if not in the queried graph. - Python: type stubs are now compatible with Python 3.7 and Mypy strict mode.
- RDF/XML: allows entities declared using other entities.
- Upgrades
quick-xml
to 0.26.
- Server: The
/store
endpoints now has ano_transaction
HTTP option forPOST
andPUT
request to get better performances at the cost of transactional guarantees. - Server: The
/store
endpoints now has alenient
HTTP option forPOST
andPUT
request to ignore syntax errors (requires theno_transaction
option). - Server: allows path that are not valid UTF-8 in file path CLI arguments.
- Rust:
From<spargebra::Query>
tooxigraph::Query
(thanks to @hobofan).
- SPARQL:
NOW()
function properly returns the current time and not 1970-01-01 - SPARQL: fixes serialization of SPARQL queries (property path and STRSTARTS function).
- SPARQL: slightly optimize aggregates by avoiding an unneeded projection.
- SPARQL: the parser now cleanly fails if invalid
VALUES
clauses are provided. - SPARQL: In DELETE/INSERT UPDATE the currently written values can't be read anymore ("Halloween problem").
oxrdf
: makes Clippy run without warnings whenrdf-star
is disable.- Python: makes type annotations compatible with Python 3.7.
- Python: makes sure the parameter default value is always included in the type annotation.
- Python: avoid uploading debug wheels to pypi.
- Python type stubs are now generated and distributed alongside wheels.
- SPARQL: Fixes evaluation of sub queries in which some variables might not be always bound.
- Python: Uses https://pyoxigraph.readthedocs.io instead of https://oxigraph.org/pyoxigraph/ to store pyoxigraph documentation.
- Server: Migration to Clap v4. Slight changes to the console textual outputs.
- Server: Makes sure body is always sent for all HTTP response codes that are not explicitly to not do that according to the HTTP specification.
- Python: Adds support of reading from
io.TextIOBase
.
- Fixes a bug in the SPARQL parser with nested anonymous blank nodes (
[]
). - Migrates the Python documentation skin to Furo.
- Allows multiple value to be given to
oxigraph_server
--file
parameter. - Limits parallelism of
oxigraph_server
file loader to avoid stack overflows.
QueryResults::with_http_redirection_limit
that allows to set a limit on the number of HTTP redirection followed (0
by default).
- Fixes a bug in SPARQL parser that was always enabling the
SILENT
option of theSERVICE
operator. - Allows unknown keys in the objects present in the SPARQL JSON query results. It allows to parse e.g. Virtuoso query results.
- Allows
"typed-literal"
as an alias of"literal"
in SPARQL JSON query results. - Improves the HTTP client error reporting: no silent failures on 4XX and 5XX responses.
- Upgrades RocksDB to 7.3.1.
- Upgrades quick-xml to 0.23.
- Fixes a bug in the server bulk loader that crashed on very small files.
- Upgrades RocksDB to v7.1.
- Removes some dead code in the bulk loader.
- The default git branch is now
main
and notmaster
(thanks to @nyurik). - Upgrades RocksDB to v7.0.4.
- Limits the number of bulk loader threads to at most 4 (allows bigger BTree chunks and a better data layout).
- Limits the number of files opened by RocksDB to the soft file descriptor limit minus 48.
- Fixes compilation on ARM.
- Moves some lints from errors to warnings in order to avoid compilation failures on different Rust versions.
- The bulk loader now checks available memory and tries to increase its batch size to make use of it.
- The Bulk loader provides now a
--lenient
option to keep loading a file even if invalid data is found (works only with N-Triples and N-Quads). This behavior can be customised in the Rust API using theBulkLoader::on_parse_error
method.
- Rocksdb has been upgrade to 7.0.2. It now requires a C++17 compatible compiler. This required dropping support of macOS 10.9 to 10.13.
- JS: Oxigraph NPM package is now also supporting web browsers and WebPack.
- JS: RDF term related classes now overrides the
toString
method. - Python: It is now possible to directly give a file path to the
parse
,serialize
,Store.load
,Store.bulk_load
andStore.dump
functions. - Python: New
Store.clear_graph
,Store.clear
,Store.optimize
andStore.flush
methods.
sophia_api
traits implementation following a request of Sophia maintainer.
- SPARQL: fixes evaluation of SPARQL queries with no results but an
ORDER BY
clause. There should be no group in the output instead of one empty group. This behavior has been changed following this discussion. - SPARQL: fixes SPARQL-star evaluation of nested triples with both variables and constants.
- SPARQL: if results are sorted, literals are now ordered by value, then datatype, then language tag. This ordering is considered as "implementation defined" by the SPARQL specification and is very likely to change in the future.
- Python: all costly methods now release the python GIL allowing multithreaded usages of pyoxigraph.
- Rust: SPARQL results writer now flushes the buffer at the end of the results writes. This makes their API less error-prone.
- Rust: the bulk loader API has been rewritten to allow hooking a progress indicator and set parallelism limit.
- Server: it is now possible to bulk load gzipped files.
- Fixes a bug in the
bulk_load_dataset
method that was creating an invalid database. - Server: Takes into account also URL query parameters if the send SPARQL request body is using form-urlencoded.
- Upgrades RocksDB to v0.28.2.
- Generate clean python sdist files compiling Oxigraph from scratch with the proper
Cargo.lock
. - Do not push beta releases to homebrew and python stable documentation.
- Moves RocksDB binding directory to
oxrocksdb-sys
.
- Fixes release on crates.io of the RocksDB bindings.
- RDF-star support.
Triple
is now a possibleTerm
. Serialization formats and SPARQL support have been updated to match the latest version of the specification draft. - Fast data bulk load with the
Store
bulk_load_dataset
andbulk_load_graph
methods and a special command line option of the server. - It is now possible to quickly backup the database using the
backup
method. - Rust:
*Syntax::from_extension
to easy guess a graph/dataset/sparql result format from a file extension. - Rust: Custom SPARQL functions are now supported using
QueryOptions::with_custom_function
. - Rust: Simple in-memory graph (
Graph
) and dataset (Dataset
) data structures with canonicalization. - Nightly build of the server binary and docker image, and of pyoxigraph wheels.
Store
operations are now transactional using the "repeatable read" isolation level: the store only exposes changes that have been "committed" (i.e. no partial writes) and the exposed state does not change for the complete duration of a read operation (e.g. a SPARQL query) or a read/write operation (e.g. a SPARQL update). theStore
transaction
method now allows to do read/write transactions. -RDF-star <https://w3c.github.io/rdf-star/cg-spec>
_ is now supported (including serialization formats and SPARQL-star). :py:class:.Triple
can now be used in :py:attr:.Triple.object
, :py:attr:.Triple.object
, :py:attr:.Quad.subject
and :py:attr:.Quad.object
.
- SPARQL: It is now possible to compare
rdf:langString
literals with the same language tag. - SPARQL: The parser now validates more carefully the inputs following the SPARQL specification and test suite.
- SPARQL: Variable scoping was buggy with "FILTER EXISTS". It is now fixed.
- Rust: RDF model, SPARQL parser and SPARQL result parsers have been moved to stand-alone reusable libraries.
- Rust: HTTPS is not supported by default with the
http_client
option. You need to enable thenative-tls
or therustls
feature of theoxhttp
crate to enable a TSL layer. - Rust: The error types have been cleaned.
Most of the
Store
methods now return aStorageError
that is more descriptive than the previousstd::io::Error
. The new error type all implementsInto<std::io::Error>
for easy conversion. - Rust: There is now a
Subject
struct that is the union ofNamedNode
,BlankNode
andTriple
. It is The used type of thesubject
field of theTriple
andQuad
structs. - Rust: The SPARQL algebra is not anymore publicly exposed in the
oxigraph
crate. The newoxalgebra
crate exposes it. - Rust:
UpdateOptions
API have been rewritten. It can now be built usingFrom<QueryOptions>
orDefault
. - Server: The command line API has been redesign. See the server README for more information.
- Server: The HTTP implementation is now provided by
oxhttp
. - Server: The HTTP response bodies are now generated on the fly instead of being buffered.
- Python: The
SledStore
andMemoryStore
classes have been removed in favor of theStore
class. - JS: The
MemoryStore
class has been renamed toStore
. - JS: The RDF/JS
DataFactory
interface is now implemented by theoxigraph
module itself and theMemoryStore.dataFactory
property has been removed. - The implementation of SPARQL evaluation has been improved for better performances (especially joins).
- The TLS implementation used in SPARQL HTTP calls is now rustls and not native-tls. The host system certificate registry is still used.
- Spargebra: The basic RDF terms are now the ones of the
oxrdf
crate.
SledStore
andMemoryStore
. There is only theRocksDbStore
anymore that is renamed toStore
.oxigraph_wikibase
is now stored in its own repository.- Rust:
From
implementations betweenoxigraph
terms andrio_api
terms.
Many thanks to Thad Guidry, James Overton and Jeremiah who sponsored the project during the development of this version.
- SPARQL 1.1 Query Results JSON Format parser.
- Python wheels for macOS are now universal2 binaries.
- The
Cargo.lock
file is now provided with releases to avoid compilation failures because of changes in dependencies. - Uses clap instead of argh for the server arguments parsing.
- Upgrades PyO3 to v0.14.
- The HTTP server allows to query the union of all graphs using the
union-default-graph
query parameter and to use the union graph for updateWHERE
clauses using theusing-union-graph
parameter. - Exposes Sled flush operation (useful for platforms without auto-flush like Windows or Android).
- Fixes a possible out of bound panic in SPARQL query evaluation.
- Upgrades RocksDB to 6.17.3.
- Server: Fixes HTTP content negotiation (charset constraints, failure to properly handle
*/*
...). - Makes Clippy 1.51 happy.
- Support of XML entities to the RDF/XML parser
- Serve: Allows unsupported query parameters in HTTP SPARQL requests.
- Fixes WASM compilation bug and optimises WASM release packages.
- Fixes named graph creation inside of a SledStore transaction.
- Fixes
pyoxigraph
build by enforcing a givenmaturin
version. - Adds code to build Python wheels for MacOS and Windows.
- SPARQL 1.1 Update support for Rust, Python and JavaScript. All store-like classes now provide an
update
method. - SPARQL 1.1 Query Results CSV and TSV Formats serializers and TSV format parser.
- SPARQL 1.1 Graph Store HTTP Protocol partial support in
oxigraph_server
. This protocol is accessible under the/store
path. - The SPARQL Query and Update algebra is now public.
- The stores are now "graph aware" i.e. it is possible to create and keep empty named graphs.
- A simple built-in HTTP client. In the Rust library, is disabled by default behind the
http_client
feature. It powers SPARQL federation and SPARQL UPDATELOAD
operations. std::str::FromStr
implementations toNamedNode
,BlankNode
,Literal
,Term
andVariable
allowing to easily parse Turtle/SPARQL serialization of these terms.- Optional Sled storage for
oxigraph_server
.
- The
default_graph_uris
andnamed_graph_uris
parameters frompyoxigraph
query
methods. - Python 3.5 support.
(Memory|RocksDB|Sled)Store::prepare_query
methods. It is possible to cache SPARQL query parsing using theQuery::parse
function and give the parsed query to thequery
method.
- Loading data into
oxigraph_server
is now possible using/store
and not anymore using/
. For example, you should use nowcurl -f -X POST -H 'Content-Type:application/n-quads' --data-binary "@MY_FILE.nq" http://localhost:7878/store
to add the N-Quads file MY_FILE.nt to the server dataset. - Fixes evaluation of
MONTH()
andDAY()
functions on thexsd:date
values. Variable::new
now validates the variable name.(Memory|RocksDB|Sled)Store::query
does not have an option parameter anymore. There is now a newquery_opt
method that allows giving options.xsd:boolean
SPARQL function now properly follows XPath specification.- Fixes SPARQL
DESCRIBE
evaluation.
The disk data format has been changed between Oxigraph 0.1 (version 0) and Oxigraph 0.2 (version 1). Data is automatically migrated from the version 0 format to the version 1 format when opened with Oxigraph 0.2.
- The
"sophia"
feature implementing thesophia_api
traits on Oxigraph terms and stores. - Explicit types for quads iterators returned by stores.
QueryOptions::with_default_graph
now takes animpl Into<GraphName>
instead of animpl Into<NamedNode>
.QueryOptions::with_named_graph
now takes animpl Into<NamedOrBlankNode>
instead of animpl Into<NamedNode>
.pyoxigraph
query
methods now takes two new parameters,default_graph
andnamed_graphs
.default_graph_uris
andnamed_graph_uris
parameters are deprecated.- Fixes a bug in
xsd:gYear
parsing.
QueryOptions
now allows settings the query dataset graph URIs (the SPARQL protocoldefault-graph-uri
andnamed-graph-uri
parameters).pyoxigraph
storequery
methods allows to provide the dataset graph URIs. It also provides an option to use all graph names as the default graph.- "default graph as union option" now works with FROM NAMED.
pyoxigraph
now exposes and documentsVariable
,QuerySolution
,QuerySolutions
andQueryTriples
oxigraph
Rust library with SPARQL 1.1 query support and memory, Sled and RocksDB stores.oxigraph_server
standalone SPARQL server.oxigraph_wikibase
standalone SPARQL server loading data from a Wikibase instance.pyoxigraph
Python library based on Oxigraph.oxigraph
NodeJS library based on Oxigraph.