knakk/rdf

RDF library for Go

rdf

This package introduces data structures for representing RDF resources, and includes functions for parsing and serialization of RDF data.

For complete documentation see godoc.

rdf's Issues

fails decoding a turtle rdf with just namespace and no id

in this particular case

@prefix ns7:    <https://domain.com/> .
ns1:Account\_0x41e55600548    ns4:seeAlso    ns7: .
ns7:    ns4:label    "blog" .

the decoder fail with unexpected character \t on ns7: line
the snippet passes validation in a turtle validator

Doesn't serialize to N-Quads

Although in the readme, it's not implemented.

Question: how to convert a TTL file to Golang structs?

I have a TTL file format which I want to convert to its relevant structs.
For example:
The following TTL file should be converted to 2 structs Host and Account

@prefix ns0:    <http://example.org/prop#> .

ns0:Host_48d491d1-7998-59d6-b3ed-07a0868d7536_531
    a                                 ns0:Host ;
    ns0:Exploitable                   "XXE", "PEC";
    ns0:Account                       ns0:Account_0abb0c18-9fb7-57f2-8bcd-2cb07ffbe865_161 ;
    ns0:HostSignificance              "3" ;
    ns0:ID                            "48d491d1-7998-59d6-b3ed-07a0868d7536_531" .

ns0:Account_0abb0c18-9fb7-57f2-8bcd-2cb07ffbe865_161
    a              ns0:Account ;
    ns0:GroupName  "Administrators" ;
    ns0:ID         "0abb0c18-9fb7-57f2-8bcd-2cb07ffbe865_161" ;
    ns0:Privileges "SeRemoteInteractiveLogonRight" ;
    ns0:UserId     "161" ;
    ns0:Username   "administrator" .

Decoders: better error handling on malformed input

Right now the decoders all give up parsing when there is some malformed RDF input.

I'd like to also make it possible to ignore errors and try to continue parsing, discarding only the malformed parts. This is easy in line-based formats N-Triples & N-Quads, but more difficult in Turtle & RDF/XML.

Start by adding an ParserOption ("Strict"?), which should indicate the desired behaviour. Default to true, but if set to false, it will try to continue, skipping the malformed input.

Maybe also add an option (an io.Writer perhaps, defaults to nil) so consumers can decide where to print errors & warnings.

Don't strip xsd type for xsd:integer in turtle serialization

In

rdf/rdf.go

Line 272 in 9659d91

case xsdInteger, xsdDecimal, xsdBoolean, xsdDouble:

xsd:integer is stripped from values. That's not correct. It was probably confused with xsd:int.

(see https://stackoverflow.com/questions/15336872/xsd-what-is-the-difference-between-xsinteger-and-xsint)

Lack of documentation + full example

The link to the documentation found in the .README file is not working.
Therefore, it's not trivial to understand how to create a .ttl file from an input data.
Is it possible to write kind of guide / supply "how to" guide.

allowing naming the prefixes/namespaces for turtle export

....so we can get nice prefixes instead of 'ns0' 'ns1' and so on

thanks for this, hugely useful project

JSON-LD support

First of all: thanks for this awesome library!

JSON-LD is one of the more popular serialization formats for RDF, but this library does not yet support it. Unfortunately, parsing JSON-LD is not that simple, but some go libraries exist that might help:

https://github.com/kazarena/json-gold
https://github.com/emersion/go-jsonld

Add support for more literal type formats

xs:integer...maybe xs:int

No license information

As far as I can tell there's no licensing information listed. That makes it difficult for others to use this code properly in other projects. Can you please specify a license?

Any plans for RDF/XML support in TripleEncoder?

I'm running into the TODO-panic message in https://github.com/knakk/rdf/blob/master/encoder.go#L114-L116 , when trying to encode triples into RDF/XML, and was wondering if there are any plans to implement it? :)

(I might have a go at it at some point, but not sure how well that will go)

'_' may appear in turtle local names

While the underscore is in the list of characters to be escaped (

rdf/encoder.go

Line 294 in 9659d91

case '_', '~', '.', '-', '!', '$', '&', '\'', '(', ')', '*', '+', ',', ';', '=', '/', '?', '#', '@', '%':

) they actually may appear in turtle local names (see rule 164s in https://www.w3.org/TR/turtle/#grammar-production-PN_CHARS_U)

There are some other characters that may also appear (like -,%)

see also: protegeproject/protege#89

So they should remain unescaped in local names. At the very least the user should get a choice.

@hydrarulz

Custom namespace functionality not working

My code is as follows:

Declaration of TripleEncoder:

out := rdf.NewTripleEncoder(os.Stdout, rdf.Turtle)
out.Namespaces = walls.Prefixes
out.GenerateNamespaces = false

Prefixes Declaration (in a different package):

var BNF = "http://wallscope.co.uk/ontology/bnf/"
var BNFRES = "http://wallscope.co.uk/resource/bnf/"

var DBO = "http://dbpedia.org/ontology/"
var DBP = "http://dbpedia.org/property/"
var DBR = "http://dbpedia.org/resource/"

var DC = "http://purl.org/dc/elements/1.1/"
var DCT = "http://purl.org/dc/terms/"
var FOAF = "http://xmlns.com/foaf/0.1/"
var GEO_POS = "http://www.w3.org/2003/01/geo/wgs84_pos#"
var RDF = "http://www.w3.org/1999/02/22-rdf-syntax-ns#"
var RDFS = "http://www.w3.org/2000/01/rdf-schema#"
var SCHEMA = "http://schema.org/"
var SKOS = "http://www.w3.org/2004/02/skos/core#"
var XSD = "http://www.w3.org/2001/XMLSchema#"

var WALLSONT = "http://wallscope.co.uk/ontology/"
var WALLSRES = "http://wallscope.co.uk/resource/"

var Prefixes = map[string]string{
	"dbo":       DBO,
	"dbp":       DBP,
	"dbr":       DBR,
	"dc":        DC,
	"dct":       DCT,
	"foaf":      FOAF,
	"geo-pos":   GEO_POS,
	"rdf":       RDF,
	"rdfs":      RDFS,
	"schema":    SCHEMA,
	"skos":      SKOS,
	"walls":     WALLSONT,
	"walls-res": WALLSRES,
	"xsd":       XSD,
}

log.Println(out.Namespaces) shows:

map[rdf:http://www.w3.org/1999/02/22-rdf-syntax-ns# dbp:http://dbpedia.org/property/ dbr:http://dbpedia.org/resource/ xsd:http://www.w3.org/2001/XMLSchema# dc:http://purl.org/dc/elements/1.1/ skos:http://www.w3.org/2004/02/skos/core# walls:http://wallscope.co.uk/ontology/ walls-res:http://wallscope.co.uk/resource/ dbo:http://dbpedia.org/ontology/ dct:http://purl.org/dc/terms/ foaf:http://xmlns.com/foaf/0.1/ geo-pos:http://www.w3.org/2003/01/geo/wgs84_pos# rdfs:http://www.w3.org/2000/01/rdf-schema# schema:http://schema.org/]

although the encoder outputs things like http://www.w3.org/2000/01/rdf-schema#label and http://www.w3.org/2004/02/skos/core#notation which should become rdfs:label and skos:notation respectively.

Am I doing something wrong? I can't find anything in the documentation about this issue

Question: How to get XML Schema data type of literal?

I have an n-triples dataset with both integer, string and float literal data types (So, an integer looks like:

"17601"^^<http://www.w3.org/2001/XMLSchema#integer>

).

I was wondering how I can determine what type a value has, in code?

License

From early coming year godoc will automatically forward to pkg.go.dev which doesn't display your docs due to an unrecognised license (see https://pkg.go.dev/github.com/knakk/rdf).

Do you mind using MIT or similar so that it will still be possible to read your docs?

How to customize prefixes?

Is there an option to custom the @prefix being created in the output of the TTL?

Add a real license?

Your license file says:

"Do what thou wilt shall be the whole of the Law."

this is not really a license. It is not explicit in a number of important points, to note: an explicit grant of permission to use, modify and redistribute the software in perpuity, and an explicit liability disclaimer. The former puts potential users of your software at risk, the latter may cause you problems if you ever get sued over this.

For background, see the following articles

http://www.linuxjournal.com/article/6225
https://news.ycombinator.com/item?id=5733245
https://opensource.google.com/docs/using/license/

Would you be willing to change your license to a standard open source license? For example, the MIT license is very permissive, and the Apache 2 license in addition grants you some protection against suits over patents.

Request to standardize license

My company makes it much easier to use OSS that has a standard license. Would you consider revising the license from "Do what thou wilt shall be the whole of the Law" to a Apache 2/MIT/etc?