rdf4h is a library for working with RDF in Haskell.
For details see the GitHub project page:
http://robstewart57.github.io/rdf4h/
Supports GHC versions from 9.2.5 (stackage lts-20.11).
To enter a development environment, you can use Nix and direnv which install all required software, also allowing you to use your preferred shell. Once installed, just run
$ direnv allow
and you'll have a working development environment for now and the future whenever you enter this directory.
This development environment allows to use either Stack or Cabal for building the software.
The coverage of the W3C RDF standards are:
Format | Parsing | Serialising |
---|---|---|
NTriples | complete | complete |
Turtle | complete | complete |
RDF/XML | complete | not supported |
These results are produced with version 4.0 of this library.
These tests are run on the W3C unit tests for RDF formats: https://github.com/w3c/rdf-tests.
-
The parsers in this library parse large files/strings contents entirely before generating RDF triples. This doesn't scale for very large files. Implementing stream based RDF parsers would overcome this problem, e.g. by creating input streams enabling output streams in the io-streams library to consume triples on-the-fly during parsing. This is discussed here: #56 (comment) and #44 (comment)
-
RDF/XML serialisation of RDF graphs.
To run all the tests (parsers and the library API):
$ git submodule update --init --recursive
$ git submodule foreach git pull origin gh-pages
$ stack test --test-arguments="--quickcheck-tests 1000"
To run specific parser tests when bug fixing:
$ stack test --test-arguments="--pattern /parser-w3c-tests-ntriples/"
$ stack test --test-arguments="--pattern /parser-w3c-tests-turtle/"
$ stack test --test-arguments="--pattern /parser-w3c-tests-xml/"
To run the bencharks:
$ wget https://www.dropbox.com/s/z1it340emcreowj/bills.099.actions.rdf
$ gzip -d bills.099.actions.rdf.gz
$ stack bench