public virtual int CompareTo(RDFDataset.Quad o)
{
if (o == null)
{
return(1);
}
int rval = GetGraph().CompareTo(o.GetGraph());
if (rval != 0)
{
return(rval);
}
rval = GetSubject().CompareTo(o.GetSubject());
if (rval != 0)
{
return(rval);
}
rval = GetPredicate().CompareTo(o.GetPredicate());
if (rval != 0)
{
return(rval);
}
return(GetObject().CompareTo(o.GetObject()));
}
// generates unique and duplicate hashes for bnodes
/// <exception cref="JsonLD.Core.JsonLdError"></exception>
public virtual object HashBlankNodes(IEnumerable <string> unnamed_)
{
#if !PORTABLE
IList <string> unnamed = new List <string>(unnamed_);
IList <string> nextUnnamed = new List <string>();
IDictionary <string, IList <string> > duplicates = new Dictionary <string, IList <string
> >();
IDictionary <string, string> unique = new Dictionary <string, string>();
// NOTE: not using the same structure as javascript here to avoid
// possible stack overflows
// hash quads for each unnamed bnode
for (int hui = 0; ; hui++)
{
if (hui == unnamed.Count)
{
// done, name blank nodes
bool named = false;
IList <string> hashes = new List <string>(unique.Keys);
hashes.SortInPlace();
foreach (string hash in hashes)
{
string bnode = unique[hash];
namer.GetName(bnode);
named = true;
}
// continue to hash bnodes if a bnode was assigned a name
if (named)
{
// this resets the initial variables, so it seems like it
// has to go on the stack
// but since this is the end of the function either way, it
// might not have to
// hashBlankNodes(unnamed);
hui = -1;
unnamed = nextUnnamed;
nextUnnamed = new List <string>();
duplicates = new Dictionary <string, IList <string> >();
unique = new Dictionary <string, string>();
continue;
}
else
{
// name the duplicate hash bnods
// names duplicate hash bnodes
// enumerate duplicate hash groups in sorted order
hashes = new List <string>(duplicates.Keys);
hashes.SortInPlace();
// process each group
for (int pgi = 0; ; pgi++)
{
if (pgi == hashes.Count)
{
// done, create JSON-LD array
// return createArray();
IList <string> normalized = new List <string>();
// Note: At this point all bnodes in the set of RDF
// quads have been
// assigned canonical names, which have been stored
// in the 'namer' object.
// Here each quad is updated by assigning each of
// its bnodes its new name
// via the 'namer' object
// update bnode names in each quad and serialize
for (int cai = 0; cai < quads.Count; ++cai)
{
RDFDataset.Quad quad = quads[cai];
foreach (string attr in new string[] { "subject", "object", "name" })
{
if (quad.ContainsKey(attr))
{
IDictionary <string, object> qa = (IDictionary <string, object>)quad[attr];
if (qa != null && (string)qa["type"] == "blank node" && ((string)qa["value"]).IndexOf
("_:c14n") != 0)
{
qa["value"] = namer.GetName((string)qa["value"]);
}
}
}
normalized.Add(RDFDatasetUtils.ToNQuad(quad, quad.ContainsKey("name"
) && !(quad["name"] == null) ? (string)((IDictionary <string, object>)((IDictionary <string, object>)quad)["name"])["value"] : null));
}
// sort normalized output
normalized.SortInPlace();
// handle output format
if (options.format != null)
{
if ("application/nquads".Equals(options.format))
{
string rval = string.Empty;
foreach (string n in normalized)
{
rval += n;
}
return(rval);
}
else
{
throw new JsonLdError(JsonLdError.Error.UnknownFormat, options.format);
}
}
string rval_1 = string.Empty;
foreach (string n_1 in normalized)
{
rval_1 += n_1;
}
return(RDFDatasetUtils.ParseNQuads(rval_1));
}
// name each group member
IList <string> group = duplicates[hashes[pgi]];
IList <NormalizeUtils.HashResult> results = new List <NormalizeUtils.HashResult>();
for (int n_2 = 0; ; n_2++)
{
if (n_2 == group.Count)
{
// name bnodes in hash order
results.SortInPlace(new _IComparer_145());
foreach (NormalizeUtils.HashResult r in results)
{
// name all bnodes in path namer in
// key-entry order
// Note: key-order is preserved in
// javascript
foreach (string key in r.pathNamer.Existing().GetKeys())
{
namer.GetName(key);
}
}
// processGroup(i+1);
break;
}
else
{
// skip already-named bnodes
string bnode = group[n_2];
if (namer.IsNamed(bnode))
{
continue;
}
// hash bnode paths
UniqueNamer pathNamer = new UniqueNamer("_:b");
pathNamer.GetName(bnode);
NormalizeUtils.HashResult result = HashPaths(bnode, bnodes, namer, pathNamer);
results.Add(result);
}
}
}
}
}
// hash unnamed bnode
string bnode_1 = unnamed[hui];
string hash_1 = HashQuads(bnode_1, bnodes, namer);
// store hash as unique or a duplicate
if (duplicates.ContainsKey(hash_1))
{
duplicates[hash_1].Add(bnode_1);
nextUnnamed.Add(bnode_1);
}
else
{
if (unique.ContainsKey(hash_1))
{
IList <string> tmp = new List <string>();
tmp.Add(unique[hash_1]);
tmp.Add(bnode_1);
duplicates[hash_1] = tmp;
nextUnnamed.Add(unique[hash_1]);
nextUnnamed.Add(bnode_1);
JsonLD.Collections.Remove(unique, hash_1);
}
else
{
unique[hash_1] = bnode_1;
}
}
}
#else
throw new PlatformNotSupportedException();
#endif
}
/// <summary>Parses RDF in the form of N-Quads.</summary>
/// <remarks>Parses RDF in the form of N-Quads.</remarks>
/// <param name="input">the N-Quads input to parse.</param>
/// <returns>an RDF dataset.</returns>
/// <exception cref="JsonLD.Core.JsonLdError"></exception>
public static RDFDataset ParseNQuads(string input)
{
// build RDF dataset
RDFDataset dataset = new RDFDataset();
// split N-Quad input into lines
string[] lines = RDFDatasetUtils.Regex.Eoln.Split(input);
int lineNumber = 0;
foreach (string line in lines)
{
lineNumber++;
// skip empty lines
if (RDFDatasetUtils.Regex.EmptyOrComment.Matcher(line).Matches())
{
continue;
}
// parse quad
Matcher match = RDFDatasetUtils.Regex.Quad.Matcher(line);
if (!match.Matches())
{
throw new JsonLdError(JsonLdError.Error.SyntaxError, "Error while parsing N-Quads; invalid quad. line:"
+ lineNumber);
}
// get subject
RDFDataset.Node subject;
if (match.Group(1) != null)
{
var subjectIri = Unescape(match.Group(1));
AssertAbsoluteIri(subjectIri);
subject = new RDFDataset.IRI(subjectIri);
}
else
{
subject = new RDFDataset.BlankNode(Unescape(match.Group(2)));
}
// get predicate
var predicateIri = Unescape(match.Group(3));
AssertAbsoluteIri(predicateIri);
RDFDataset.Node predicate = new RDFDataset.IRI(predicateIri);
// get object
RDFDataset.Node @object;
if (match.Group(4) != null)
{
var objectIri = Unescape(match.Group(4));
AssertAbsoluteIri(objectIri);
@object = new RDFDataset.IRI(objectIri);
}
else
{
if (match.Group(5) != null)
{
@object = new RDFDataset.BlankNode(Unescape(match.Group(5)));
}
else
{
string language = Unescape(match.Group(8));
string datatype = match.Group(7) != null?Unescape(match.Group(7)) : match.Group
(8) != null ? JSONLDConsts.RdfLangstring : JSONLDConsts.XsdString;
AssertAbsoluteIri(datatype);
string unescaped = Unescape(match.Group(6));
@object = new RDFDataset.Literal(unescaped, datatype, language);
}
}
// get graph name ('@default' is used for the default graph)
string name = "@default";
if (match.Group(9) != null)
{
name = Unescape(match.Group(9));
AssertAbsoluteIri(name);
}
else
{
if (match.Group(10) != null)
{
name = Unescape(match.Group(10));
}
}
RDFDataset.Quad triple = new RDFDataset.Quad(subject, predicate, @object, name);
// initialise graph in dataset
if (!dataset.ContainsKey(name))
{
IList <RDFDataset.Quad> tmp = new List <RDFDataset.Quad>();
tmp.Add(triple);
dataset[name] = tmp;
}
else
{
// add triple if unique to its graph
IList <RDFDataset.Quad> triples = (IList <RDFDataset.Quad>)dataset[name];
if (!triples.Contains(triple))
{
triples.Add(triple);
}
}
}
return(dataset);
}
internal static string ToNQuad(RDFDataset.Quad triple, string graphName)
{
return(ToNQuad(triple, graphName, null));
}
internal static string ToNQuad(RDFDataset.Quad triple, string graphName, string bnode
)
{
RDFDataset.Node s = triple.GetSubject();
RDFDataset.Node p = triple.GetPredicate();
RDFDataset.Node o = triple.GetObject();
string quad = string.Empty;
// subject is an IRI or bnode
if (s.IsIRI())
{
quad += "<" + Escape(s.GetValue()) + ">";
}
else
{
// normalization mode
if (bnode != null)
{
quad += bnode.Equals(s.GetValue()) ? "_:a" : "_:z";
}
else
{
// normal mode
quad += s.GetValue();
}
}
if (p.IsIRI())
{
quad += " <" + Escape(p.GetValue()) + "> ";
}
else
{
// otherwise it must be a bnode (TODO: can we only allow this if the
// flag is set in options?)
quad += " " + Escape(p.GetValue()) + " ";
}
// object is IRI, bnode or literal
if (o.IsIRI())
{
quad += "<" + Escape(o.GetValue()) + ">";
}
else
{
if (o.IsBlankNode())
{
// normalization mode
if (bnode != null)
{
quad += bnode.Equals(o.GetValue()) ? "_:a" : "_:z";
}
else
{
// normal mode
quad += o.GetValue();
}
}
else
{
string escaped = Escape(o.GetValue());
quad += "\"" + escaped + "\"";
if (JSONLDConsts.RdfLangstring.Equals(o.GetDatatype()))
{
quad += "@" + o.GetLanguage();
}
else
{
if (!JSONLDConsts.XsdString.Equals(o.GetDatatype()))
{
quad += "^^<" + Escape(o.GetDatatype()) + ">";
}
}
}
}
// graph
if (graphName != null)
{
if (graphName.IndexOf("_:") != 0)
{
quad += " <" + Escape(graphName) + ">";
}
else
{
if (bnode != null)
{
quad += " _:g";
}
else
{
quad += " " + graphName;
}
}
}
quad += " .\n";
return(quad);
}
public Object Normalize()
{
this.quads = new List <IDictionary <string, IDictionary <string, string> > >();
this.blankNodeInfo = new Dictionary <string, IDictionary <string, IList <object> > >();
this.hashToBlankNodes = new Dictionary <string, IList <string> >();
this.canonicalIssuer = new IdentifierIssuer("_:c14n");
/*
* 2) For every quad in input dataset:
* STATUS : step 2 is good!
*/
foreach (string graphName in this.dataset.Keys)
{
IList <IDictionary <string, IDictionary <string, string> > > triples = (IList <IDictionary <string, IDictionary <string, string> > >) this.dataset[graphName];
if (graphName.Equals("@default"))
{
graphName.Replace("@default", null);
}
foreach (IDictionary <string, IDictionary <string, string> > quad in triples)
{
if (!string.ReferenceEquals(graphName, null))
{
if (graphName.StartsWith("_:", StringComparison.Ordinal))
{
IDictionary <string, string> tmp = new Dictionary <string, string>();
tmp["type"] = "blank node";
quad["name"] = tmp;
}
else
{
IDictionary <string, string> tmp = new Dictionary <string, string>();
tmp["type"] = "IRI";
quad["name"] = tmp;
}
quad["name"]["value"] = graphName;
}
this.quads.Add(quad);
/* 2.1) For each blank node that occurs in the quad, add a
* reference to the quad using the blank node identifier in the
* blank node to quads map, creating a new entry if necessary.
* */
foreach (string key in quad.Keys)
{
Dictionary <string, string> component = (Dictionary <string, string>)quad[key];
if (key.Equals("predicate") || !component["type"].Equals("blank node"))
{
continue;
}
string id = component["value"];
if (this.blankNodeInfo[id] == null)
{
IDictionary <string, IList <Object> > quadList = new Dictionary <string, IList <Object> >();
quadList["quads"] = new List <Object>();
quadList["quads"].Add(quad);
this.blankNodeInfo[id] = quadList;
}
else
{
this.blankNodeInfo[id]["quads"].Add(quad);
}
}
}
List <string> nonNormalized = new List <string>();
nonNormalized.AddRange(blankNodeInfo.Keys);
//Collections.sort(nonNormalized);
/* 4) Initialize simple, a boolean flag, to true.
* STATUS : if this does not work we have a serious problem
*/
bool simple = true;
/*
* 5) While simple is true, issue canonical identifiers for blank nodes:
*/
while (simple)
{
// 5.1) Set simple to false.
simple = false;
// 5.2) Clear hash to blank nodes map.
this.hashToBlankNodes.Clear();
/*
* 5.3) For each blank node identifier identifier in non-normalized
* identifiers:
* STATUS : working on it
*/
foreach (string id in nonNormalized)
{
string hash = hashFirstDegreeQuads(id);
if (this.hashToBlankNodes.ContainsKey(hash))
{
this.hashToBlankNodes[hash].Add(id);
}
else
{
List <string> idList = new List <string>();
idList.Add(id);
this.hashToBlankNodes.Add(hash, idList);
}
}
/*
* 5.4) For each hash to identifier list mapping in hash to blank
* nodes map, lexicographically-sorted by hash:
*/
foreach (string hash in sortMapKeys(this.hashToBlankNodes))
{
IList <string> idList = this.hashToBlankNodes[hash];
if (idList.Count() > 1)
{
continue;
}
/* 5.4.2) Use the Issue Identifier algorithm, passing canonical
* issuer and the single blank node identifier in identifier
* list, identifier, to issue a canonical replacement identifier
* for identifier.
*/
string id = idList[0];
this.canonicalIssuer.getId(id);
// 5.4.3) Remove identifier from non-normalized identifiers.
nonNormalized.Remove(id);
// 5.4.4) Remove hash from the hash to blank nodes map.
this.hashToBlankNodes.Remove(hash);
// 5.4.5) Set simple to true.
simple = true;
}
}
/*
* 6) For each hash to identifier list mapping in hash to blank nodes
* map, lexicographically-sorted by hash:
* STATUS: does not loop through it
*/
foreach (string hash in sortMapKeys(this.hashToBlankNodes))
{
IList <string> idList = this.hashToBlankNodes[hash];
/*
* 6.1) Create hash path list where each item will be a result of
* running the Hash N-Degree Quads algorithm.
*/
var hashPathList = new List <IDictionary <string, object> >();
/*
* 6.2) For each blank node identifier identifier in identifier
* list:
*/
foreach (string id in idList)
{
/*
* 6.2.1) If a canonical identifier has already been issued for
* identifier, continue to the next identifier.
*/
if (this.canonicalIssuer.hasID(id))
{
continue;
}
/*
* 6.2.2) Create temporary issuer, an identifier issuer
* initialized with the prefix _:b.
*/
IdentifierIssuer issuer = new IdentifierIssuer("_:b");
/*
* 6.2.3) Use the Issue Identifier algorithm, passing temporary
* issuer and identifier, to issue a new temporary blank node
* identifier for identifier.
*/
issuer.getId(id);
/*
* 6.2.4) Run the Hash N-Degree Quads algorithm, passing
* temporary issuer, and append the result to the hash path
* list.
*/
hashPathList.Add(hashNDegreeQuads(issuer, id));
}
/*
* 6.3) For each result in the hash path list,
* lexicographically-sorted by the hash in result:
*/
sortMapList(hashPathList);
foreach (var result in hashPathList)
{
if (result["issuer"] != null)
{
foreach (var existing in ((IdentifierIssuer)result["issuer"]).getOrder())
{
this.canonicalIssuer.getId(existing);
}
}
}
}
/*
* Note: At this point all blank nodes in the set of RDF quads have been
* assigned canonical identifiers, which have been stored in the
* canonical issuer. Here each quad is updated by assigning each of its
* blank nodes its new identifier.
*/
// 7) For each quad, quad, in input dataset:
List <string> normalized = new List <string>();
foreach (var quadMap in this.quads)
{
/*
* Create a copy, quad copy, of quad and replace any existing
* blank node identifiers using the canonical identifiers previously
* issued by canonical issuer. Note: We optimize away the copy here.
* STATUS : currently working on it
*/
foreach (var key in quadMap.Keys)
{
if (key.Equals("predicate"))
{
continue;
}
else
{
var component = quadMap[key];
if (component["type"].Equals("blank node") && !component["value"].StartsWith(this.canonicalIssuer.getPrefix()))
{
component.Add("value", this.canonicalIssuer.getId(component["value"]));
}
}
}
// 7.2) Add quad copy to the normalized dataset.
RDFDataset.Quad quad = new RDFDataset.Quad(quadMap, quadMap.ContainsKey("name") && quadMap["name"] != null
? (quadMap["name"])["value"] : null);
normalized.Add(RDFDatasetUtils.ToNQuad(quad, quadMap.ContainsKey("name") && quadMap["name"] != null
? (quadMap["name"])["value"] : null));
}
// 8) Return the normalized dataset.
Collections.SortInPlace(normalized);
if (this.options.format != null)
{
if ("applications/nquads".Equals(this.options.format))
{
StringBuilder rval = new StringBuilder();
foreach (var n in normalized)
{
rval.Append(n);
}
return(rval.ToString());
}
else
{
// will need to implement error handling
return(null);
}
}
else
{
StringBuilder rval = new StringBuilder();
foreach (var n in normalized)
{
rval.Append(n);
}
try
{
return(RDFDatasetUtils.ParseNQuads(rval.ToString()));
}
catch (Exception ex)
{
Console.Out.WriteLine(ex);
return(ex);
}
}
}
return(null);
}
/*
* STATUS : working on it
*/
private string hashFirstDegreeQuads(string id)
{
IDictionary <string, IList <Object> > info = this.blankNodeInfo[id];
if (info.ContainsKey("hash"))
{
return(info["hash"].ToString());
}
// 1) Initialize nquads to an empty list. It will be used to store quads
// in N-Quads format.
IList <string> nquads = new List <string>();
// 2) Get the list of quads quads associated with the reference blank
// node identifier in the blank node to quads map.
IList <Object> quads = info["quads"];
// 3) For each quad quad in quads:
foreach (var quad in quads)
{
// 3.1) Serialize the quad in N-Quads format with the following
// special rule:
// 3.1.1) If any component in quad is an blank node, then serialize
// it using a special identifier as follows:
// copy = {}
IDictionary <string, IDictionary <string, string> > copy = new Dictionary <string, IDictionary <string, string> >();
/* 3.1.2) If the blank node's existing blank node identifier
* matches the reference blank node identifier then use the
* blank node identifier _:a, otherwise, use the blank node
* identifier _:z.
* STATUS: working
*/
RDFDataset.Quad quadMap = (RDFDataset.Quad)quad;
foreach (var key in quadMap)
{
IDictionary <string, string> component = new Dictionary <string, string>();
component.Add(key.Key, key.Value.ToString());
if (key.Equals("predicate"))
{
copy.Add(key.Key, component);
continue;
}
copy.Add(key.Key, modifyFirstDegreeComponent(component, id));
}
RDFDataset.Quad copyQuad = new RDFDataset.Quad(copy, copy.ContainsKey("name") && copy["name"] != null
? (copy["name"])["value"] : null);
nquads.Add(RDFDatasetUtils.ToNQuad(copyQuad, copyQuad.ContainsKey("name") && copyQuad["name"] != null
? (string)((IDictionary <string, object>)copyQuad["name"])["value"] : null));
// 4) Sort nquads in lexicographical order.
}
Collections.SortInPlace(nquads);
// 5) Return the hash that results from passing the sorted, joined
// nquads through the hash algorithm.
return(NormalizeUtils.sha256HashnQuads(nquads));
}
