callidon/joseki

Pure Go library for working with RDF, a powerful framework for representing informations as graphs.

joseki

Joseki is a pure Go library for working with RDF, a powerful framework for representing informations as graphs.

For more informations about RDF itself, please see https://www.w3.org/TR/rdf11-concepts

Features

Joseki provides the following features to work with RDF :

• Structures to represent and manipulate the RDF model (URIs, Literals, Blank Nodes, Triples, etc)
• RDF Graphs to store data, with several implentations provided.
• A Low level API to query data stored in graphs.
• A High level API to query data using the SPARQL 1.1 query language. (WIP - Unstable)
• Query processing using modern techniques such as join ordering or optimized query execution plans.
• Load RDF data stored in files in various formats (N-Triples, Turtle, etc) into any graph.

Getting Started

This package aims to work with RDF graphs, which are composed of RDF Triple {Subject Object Predicate}. Using joseki, you can represent an RDF Triple as followed :

import (
    "github.com/Callidon/joseki/rdf"
    "fmt"
)
subject := rdf.NewURI("http://example.org/book/book1")
predicate := rdf.NewURI("http://purl.org/dc/terms/title")
object := rdf.NewLiteral("Harry Potter and the Order of the Phoenix")
triple := rdf.NewTriple(subject, predicate, object)
fmt.Println(triple)
// Output : {<http://example.org/book/book1> <http://purl.org/dc/terms/title> "Harry Potter and the Order of the Phoenix"}

You can also store your RDF Triples in a RDF Graph, using various type of graphs. Here, we use a Tree Graph to store our triple :

import (
    "github.com/Callidon/joseki/rdf"
    "github.com/Callidon/joseki/graph"
)
subject := rdf.NewURI("http://example.org/book/book1")
predicate := rdf.NewURI("http://purl.org/dc/terms/title")
object := rdf.NewLiteral("Harry Potter and the Order of the Phoenix")
graph := graph.NewTreeGraph()
graph.Add(rdf.NewTriple(subject, predicate, object))

You can also query any triple from a RDF Graph, using a low level API or a SPARQL query.

import (
    "github.com/Callidon/joseki/rdf"
    "github.com/Callidon/joseki/graph"
    "fmt"
)
graph := graph.NewTreeGraph()
// Datas stored in a file can be easily loaded into a graph
graph.LoadFromFile("datas/awesome-books.ttl", "turtle")
// Let's fetch the titles of all the books in our graph !
subject := rdf.NewVariable("title")
predicate := rdf.NewURI("http://www.w3.org/1999/02/22-rdf-syntax-ns#type")
object := rdf.NewURI("https://schema.org/Book")
for bindings := range graph.Filter(subject, predicate, object) {
    fmt.Println(bindings)
}

For more informations about specific features, see the documentation

joseki's Issues

Convert RDF to a Golang struct

Does this project enables to convert a rdf turtle (TTL file format) to a Golang struct?
I.e. marshal & unmarshal to/from TTL/struct.

Nested select queries return 400 error

Posting the following query to the http://data.linkedmdb.org/sparql endpoint returns a 400 error.

select ?label ?n {
  {
    select ?p (count(?o) as ?n) {
      ?s ?p ?o .
    }
    group by ?s
  }
  bind (replace(str(?s), ".*\\/", "") as ?label)
  filter (?n > 1000)
}
order by ?p

A curl request would look like this:

curl http://data.linkedmdb.org/sparql -X POST --data 'query=select+%3Flabel+%3Fn+%7B%0A++%7B%0A++++select+%3Fp+(count(%3Fs)+as+%3Fn)+%7B%0A++++++%3Fs+%3Fp+%3Fo+.%0A++++%7D%0A++++group+by+%3Fp%0A++%7D%0A++bind+(replace(str(%3Fp)%2C+%22.*%5C%5C%2F%22%2C+%22%22)+as+%3Flabel)%0A++filter+(%3Fn+%3E+1000)%0A%7D%0Aorder+by+%3Fp' -H 'Accept: application/sparql-results+json'

Expected outcome:

The endpoint returns a histogram of predicates, filtered to show only predicates that occur more than 1000 times.

Actual outcome:

The endpoint returns a 400 error with the following headers.

> POST /sparql HTTP/1.1
> Host: data.linkedmdb.org
> User-Agent: curl/7.55.1
> Accept: application/sparql-results+json
> Content-Length: 267
> Content-Type: application/x-www-form-urlencoded
>
} [267 bytes data]
* upload completely sent off: 267 out of 267 bytes
100   267    0     0  100   267      0    267  0:00:01 --:--:--  0:00:01  1265< HTTP/1.1 400 Parse_error__select_label_n_________select_p_counts_as_n________s_p_o___________group_by_p______bind_replacestrp___as_label___filter_n__1000__order_by_p__Encountered_select_at_line_3_column_5_Was_expecting_one_of_____IRIref______PNAME_NS______PNAME_LN______BLANK_NODE_LABEL______VAR1______VAR2______graph______optional______service______filter______true______false______INTEGER______DECIMAL______DOUBLE______INTEGER_POSITIVE______DECIMAL_POSITIVE______DOUBLE_POSITIVE______INTEGER_NEGATIVE______DECIMAL_NEGATIVE______DOUBLE_NEGATIVE______STRING_LITERAL1______STRING_LITERAL2______STRING_LITERAL_LONG1______STRING_LITERAL_LONG2____________NIL________________________ANON______
< Date: Tue, 17 Apr 2018 11:55:02 GMT
< Server: Jetty(6.1.4)
< Pragma: no-cache
< X-Joseki-Server: Joseki-3.0-dev
< Content-Type: text/html; charset=iso-8859-1
< Content-Length: 2387
< Connection: close

To the best of my knowledge, the query is correct.