Skip to content

Commit

Permalink
Fix XML root minting (see #386)
Browse files Browse the repository at this point in the history 
Delegate root minting to FacadeXNodeBuilder (see #380)
Some code cleanup #373
  • Loading branch information
@luigi-asprino
luigi-asprino committed Jun 21, 2023
1 parent f486f82 commit d382c0c
Show file tree
Hide file tree
Showing 27 changed files with 233 additions and 274 deletions.
5 changes: 2 additions & 3 deletions sparql-anything-archive/src/main/java/io/github/sparqlanything/zip/FolderTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -44,21 +44,20 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
logger.warn("No location provided");
return;
}
String root = Triplifier.getRootArgument(properties);
String dataSourceId = "";
String matches = properties.getProperty(ZipTriplifier.MATCHES, ".*");

logger.trace("Matches {}", matches);

builder.addRoot(dataSourceId, root);
builder.addRoot(dataSourceId);

try {
Path path = Paths.get(url.toURI());
AtomicInteger i = new AtomicInteger(1);
Files.walk(path).forEach(p -> {
logger.trace("{} matches? {}", p.toString(), path.toString().matches(matches));
if (p.toString().matches(matches)) {
builder.addValue(dataSourceId, root, i.getAndIncrement(), p.toUri().toString());
builder.addValue(dataSourceId, builder.getRoot(dataSourceId), i.getAndIncrement(), p.toUri().toString());
}
});

Expand Down
5 changes: 2 additions & 3 deletions sparql-anything-archive/src/main/java/io/github/sparqlanything/zip/TarTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -44,14 +44,13 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
return;
}
Charset charset = Triplifier.getCharsetArgument(properties);
String root = Triplifier.getRootArgument(properties);
String dataSourceId = "";
String matches = properties.getProperty(ZipTriplifier.MATCHES, ".*");

logger.trace("Matches {}", matches);

// Graph g = GraphFactory.createDefaultGraph();
builder.addRoot(dataSourceId, root);
builder.addRoot(dataSourceId);

try {
TarArchiveInputStream debInputStream = (TarArchiveInputStream) new ArchiveStreamFactory()
Expand All @@ -61,7 +60,7 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
while ((entry = (TarArchiveEntry) debInputStream.getNextEntry()) != null) {

if (entry.getName().matches(matches)) {
builder.addValue(dataSourceId, root, i, entry.getName());
builder.addValue(dataSourceId, builder.getRoot(dataSourceId), i, entry.getName());
i++;
}

Expand Down
5 changes: 2 additions & 3 deletions sparql-anything-archive/src/main/java/io/github/sparqlanything/zip/ZipTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -44,19 +44,18 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
logger.warn("No location provided");
return;
}
String root = Triplifier.getRootArgument(properties);
String dataSourceId = "";
Charset charset = Triplifier.getCharsetArgument(properties);
String matches = properties.getProperty(MATCHES, ".*");

builder.addRoot(dataSourceId, root);
builder.addRoot(dataSourceId);

ZipInputStream zis = new ZipInputStream(url.openStream(), charset);
ZipEntry ze;
int i = 1;
while ((ze = zis.getNextEntry()) != null) {
if (ze.getName().matches(matches)) {
builder.addValue(dataSourceId, root, i, NodeFactory.createLiteral(ze.getName()));
builder.addValue(dataSourceId, builder.getRoot(dataSourceId), i, NodeFactory.createLiteral(ze.getName()));
i++;
}
}
Expand Down
7 changes: 3 additions & 4 deletions sparql-anything-bib/src/main/java/io/github/sparqlanything/bib/BibtexTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -44,11 +44,10 @@ public BibtexTriplifier() {
@Override
public void triplify(Properties properties, FacadeXGraphBuilder builder) throws IOException {

String root = Triplifier.getRootArgument(properties);
String dataSourceId = "";
String namespace = PropertyUtils.getStringProperty(properties, IRIArgument.NAMESPACE);

builder.addRoot(dataSourceId, root);
builder.addRoot(dataSourceId);

try (InputStream is = Triplifier.getInputStream(properties)){

Expand All @@ -57,13 +56,13 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
BibTeXDatabase bibDB = bibtexParser.parse(reader);
AtomicInteger count = new AtomicInteger();
bibDB.getEntries().forEach((key, entry) -> {
String containerIdChild = root + key;
String containerIdChild = builder.getRoot(dataSourceId).concat(key.toString());
try {
builder.addType(dataSourceId, containerIdChild, new URI(namespace + entry.getType().toString()));
} catch (URISyntaxException e) {
logger.error("",e);
}
builder.addContainer(dataSourceId, root, count.incrementAndGet(), containerIdChild);
builder.addContainer(dataSourceId, builder.getRoot(dataSourceId), count.incrementAndGet(), containerIdChild);
entry.getFields().forEach((keyField, valueField) -> {
builder.addValue(dataSourceId, containerIdChild, keyField.toString(), valueField.toUserString());
});
Expand Down
5 changes: 2 additions & 3 deletions sparql-anything-binary/src/main/java/io/github/sparqlanything/binary/BinaryTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -63,7 +63,6 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
logger.warn("Using default encoding (Base64)");
}

String root = Triplifier.getRootArgument(properties);
String dataSourceId = "";
// Charset charset = getCharsetArgument(properties);
// boolean blank_nodes = Triplifier.getBlankNodeArgument(properties);
Expand All @@ -81,9 +80,9 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
break;
}
// Add root
builder.addRoot(dataSourceId, root);
builder.addRoot(dataSourceId);
// Add content
builder.addValue(dataSourceId, root, 1, NodeFactory.createLiteralByValue(value, XSDDatatype.XSDbase64Binary));
builder.addValue(dataSourceId, builder.getRoot(dataSourceId), 1, NodeFactory.createLiteralByValue(value, XSDDatatype.XSDbase64Binary));
}

private byte[] downloadUrl(URL toDownload) {
Expand Down
8 changes: 1 addition & 7 deletions sparql-anything-csv/src/main/java/io/github/sparqlanything/csv/CSVSlice.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -25,13 +25,11 @@ public class CSVSlice implements Slice<CSVRecord> {
private CSVRecord record;
private int iteration;
private String dataSourceId;
private String rootId;
LinkedHashMap<Integer, String> headers;

public static CSVSlice makeSlice(CSVRecord record, int iteration, String dataSourceId, String rootId, LinkedHashMap<Integer, String> headers){
public static CSVSlice makeSlice(CSVRecord record, int iteration, String dataSourceId, LinkedHashMap<Integer, String> headers){
CSVSlice r = new CSVSlice();
r.dataSourceId = dataSourceId;
r.rootId = rootId;
r.iteration = iteration;
r.record = record;
r.headers = headers;
Expand All @@ -53,10 +51,6 @@ public String getDatasourceId() {
return dataSourceId;
}

@Override
public String getRootId() {
return rootId;
}

public LinkedHashMap<Integer, String> getHeaders(){
return this.headers;
Expand Down
11 changes: 5 additions & 6 deletions sparql-anything-csv/src/main/java/io/github/sparqlanything/csv/CSVTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -119,13 +119,12 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
// return;

CSVFormat format = buildFormat(properties);
String root = Triplifier.getRootArgument(properties);
Charset charset = Triplifier.getCharsetArgument(properties);

String dataSourceId = ""; // there is always 1 data source id

// Add type Root
builder.addRoot(dataSourceId, root);
builder.addRoot(dataSourceId);

try (InputStream is = Triplifier.getInputStream(properties);){
Reader in = new InputStreamReader(new BOMInputStream(is), charset);
Expand All @@ -145,7 +144,7 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
log.debug("current row num: {}", rown);
}
CSVRecord record = recordIterator.next();
processRow(rown, dataSourceId, root, record, headers_map, builder);
processRow(rown, dataSourceId, builder.getRoot(dataSourceId), record, headers_map, builder);
}
log.debug("{} records", rown);
} catch (IllegalArgumentException e) {
Expand Down Expand Up @@ -230,7 +229,7 @@ public boolean hasNext() {
public Slice next() {
rown++;
log.trace("next slice: {}", rown);
return CSVSlice.makeSlice(recordIterator.next(), rown, dataSourceId, root, headers_map);
return CSVSlice.makeSlice(recordIterator.next(), rown, dataSourceId, headers_map);
}
};
}
Expand All @@ -240,7 +239,7 @@ public Slice next() {
@Override
public void triplify(Slice slice, Properties p, FacadeXGraphBuilder builder) {
CSVSlice csvo = (CSVSlice) slice;
builder.addRoot(csvo.getDatasourceId(), csvo.getRootId());
processRow(csvo.iteration(), csvo.getDatasourceId(), csvo.getRootId(), csvo.get(), csvo.getHeaders(), builder);
builder.addRoot(csvo.getDatasourceId());
processRow(csvo.iteration(), csvo.getDatasourceId(), builder.getRoot(csvo.getDatasourceId()), csvo.get(), csvo.getHeaders(), builder);
}
}
11 changes: 5 additions & 6 deletions sparql-anything-docs/src/main/java/io/github/sparqlanything/docs/DocxTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -54,19 +54,18 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
if (url == null)
return;

String root = Triplifier.getRootArgument(properties);
String dataSourceId = "";
String namespace = PropertyUtils.getStringProperty(properties, IRIArgument.NAMESPACE);
boolean mergeParagraphs = Boolean.parseBoolean(properties.getProperty(MERGE_PARAGRAPHS, "false"));
boolean headers = Boolean.parseBoolean(properties.getProperty(TABLE_HEADERS, "false"));

builder.addRoot(dataSourceId, root);
builder.addRoot(dataSourceId);

InputStream is = url.openStream();
try (XWPFDocument document = new XWPFDocument(is)) {
List<XWPFParagraph> paragraphs = document.getParagraphs();

builder.addType(dataSourceId, root, namespace + "Document");
builder.addType(dataSourceId, builder.getRoot(dataSourceId), namespace + "Document");

int count = 1;
if (!mergeParagraphs) {
Expand All @@ -83,7 +82,7 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
}


builder.addContainer(dataSourceId, root, count, paragraphURI);
builder.addContainer(dataSourceId, builder.getRoot(dataSourceId), count, paragraphURI);
builder.addValue(dataSourceId, paragraphURI, 1, para.getText());

count++;
Expand All @@ -95,7 +94,7 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
sb.append(para.getText());
sb.append("\n");
}
builder.addValue(dataSourceId, root, count,
builder.addValue(dataSourceId, builder.getRoot(dataSourceId), count,
NodeFactory.createLiteral(sb.toString(), XSDDatatype.XSDstring));
count++;
}
Expand All @@ -105,7 +104,7 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
XWPFTable xwpfTable = (XWPFTable) it.next();

String tableId = namespace + "Table_" + count;
builder.addContainer(dataSourceId, root, count, tableId);
builder.addContainer(dataSourceId, builder.getRoot(dataSourceId), count, tableId);

LinkedHashMap<Integer, String> headers_map = new LinkedHashMap<Integer, String>();
int rown = 0;
Expand Down
30 changes: 15 additions & 15 deletions sparql-anything-html/src/main/java/io/github/sparqlanything/html/HTMLTriplifier.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,6 @@ public class HTMLTriplifier implements Triplifier {
private static final String HTML_NS = "http://www.w3.org/1999/xhtml#";
private static final String DOM_NS = "https://html.spec.whatwg.org/#";

private String root;

private static String localName(Element element) {
String tagName = element.tagName().replace(':', '|');
Expand All @@ -79,11 +78,12 @@ private static String localName(Element element) {
@Override
public void triplify(Properties properties, FacadeXGraphBuilder builder) throws IOException, TriplifierHTTPException {

this.root = Triplifier.getRootArgument(properties);
// this.root = Triplifier.getRootArgument(properties);
Charset charset = Triplifier.getCharsetArgument(properties);
boolean blank_nodes = PropertyUtils.getBooleanProperty(properties, IRIArgument.BLANK_NODES);
String namespace = PropertyUtils.getStringProperty(properties, IRIArgument.NAMESPACE);
String selector = properties.getProperty(PROPERTY_SELECTOR, ":root");
String dataSourceId = "";

log.trace(properties.toString());
if (properties.containsKey(PROPERTY_METADATA) && Boolean.parseBoolean(properties.getProperty(PROPERTY_METADATA))) {
Expand All @@ -96,7 +96,7 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws
}
}

log.trace("namespace {}\n root {}\ncharset {}\nselector {}", namespace, root, charset, selector);
log.trace("namespace {}\n root {}\ncharset {}\nselector {}", namespace, builder.getRoot(dataSourceId), charset, selector);

Document doc;
// If location is a http or https, raise exception if status is not 200
Expand All @@ -114,25 +114,25 @@ public void triplify(Properties properties, FacadeXGraphBuilder builder) throws

Elements elements = doc.select(selector);
String rootResourceId = null;
String dataSourceId = "";

if (elements.size() > 1) {
// Create a root container
rootResourceId = root;
builder.addRoot(dataSourceId, rootResourceId);
rootResourceId = builder.getRoot(dataSourceId);
builder.addRoot(dataSourceId);
}

int counter = 0;
for (Element element : elements) {
counter++;
String resourceId = toResourceId(element, blank_nodes);
String resourceId = toResourceId(element, blank_nodes, builder, dataSourceId);
if (elements.size() > 1) {
// link to root container
builder.addContainer(dataSourceId, rootResourceId, counter, toResourceId(element, blank_nodes));
builder.addContainer(dataSourceId, rootResourceId, counter, toResourceId(element, blank_nodes, builder, dataSourceId));
} else {
// Is root container
rootResourceId = root;
resourceId = root;
builder.addRoot(dataSourceId, rootResourceId);
rootResourceId = builder.getRoot(dataSourceId);
resourceId = builder.getRoot(dataSourceId);
builder.addRoot(dataSourceId);
}
try {
populate(builder, dataSourceId, element, blank_nodes, resourceId);
Expand Down Expand Up @@ -188,22 +188,22 @@ private void populate(FacadeXGraphBuilder builder, String dataSourceId, Element
if (child.outerHtml().trim().equals("")) continue;
counter++;
if (child instanceof Element) {
builder.addContainer(dataSourceId, resourceId, counter, toResourceId((Element) child, blank_nodes));
populate(builder, dataSourceId, (Element) child, blank_nodes, toResourceId((Element) child, blank_nodes));
builder.addContainer(dataSourceId, resourceId, counter, toResourceId((Element) child, blank_nodes, builder, dataSourceId));
populate(builder, dataSourceId, (Element) child, blank_nodes, toResourceId((Element) child, blank_nodes, builder, dataSourceId));
} else {
builder.addValue(dataSourceId, resourceId, counter, child.outerHtml());
}
}

}

private String toResourceId(Element element, boolean blankNodes) {
private String toResourceId(Element element, boolean blankNodes, FacadeXGraphBuilder builder, String dataSourceId) {
if (blankNodes) {
return Integer.toHexString(element.hashCode());
} else {
String ln = localName(element);
log.debug(ln);
return this.root.concat("/").concat(ln);
return builder.getRoot(dataSourceId).concat("/").concat(ln);
}
}

Expand Down
17 changes: 17 additions & 0 deletions sparql-anything-it/src/test/java/io/github/sparqlanything/it/IssuesTest.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -746,6 +746,23 @@ public void testIssue351() throws URISyntaxException, IOException {

}

/**
* See <a href="https://github.com/SPARQL-Anything/sparql.anything/issues/386">...</a>
*/
@Test
public void testIssue386() throws URISyntaxException, IOException {
Dataset ds = DatasetFactory.createGeneral();
QC.setFactory(ARQ.getContext(), FacadeX.ExecutorFactory);
Query query;
String queryStr = IOUtils.toString(Objects.requireNonNull(getClass().getClassLoader().getResource("issues/issue386.sparql")).toURI(), StandardCharsets.UTF_8);
query = QueryFactory.create(queryStr);
QueryExecution qExec = QueryExecutionFactory.create(query, ds);
ResultSet rs = qExec.execSelect();
assertTrue(rs.hasNext());
assertEquals("http://example.org/document",rs.next().get("root").asResource().getURI());
assertFalse(rs.hasNext());
}

/**
* See <a href="https://github.com/SPARQL-Anything/sparql.anything/issues/371">...</a>
*/
Expand Down
13 changes: 13 additions & 0 deletions sparql-anything-it/src/test/resources/issues/issue386.sparql
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
PREFIX fx: <http://sparql.xyz/facade-x/ns/>
SELECT ?root {

SERVICE<x-sparql-anything:> {
fx:properties fx:content "<root><child>child1</child><child>child2</child></root>" ;
fx:media-type "application/xml" ;
fx:blank-nodes false ;
fx:root "http://example.org/document" .

?root a fx:root .

}
}
8 changes: 1 addition & 7 deletions sparql-anything-json/src/main/java/io/github/sparqlanything/json/JSONPathSlice.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,6 @@ public class JSONPathSlice implements Slice<Object> {
private Object object;
private int iteration;
private String dataSourceId;
private String rootId;

private JSONPathSlice(){}

Expand All @@ -44,17 +43,12 @@ public String getDatasourceId() {
return dataSourceId;
}

@Override
public String getRootId() {
return rootId;
}

public static JSONPathSlice makeSlice(Object object, int iteration, String rootId, String dataSourceId){
public static JSONPathSlice makeSlice(Object object, int iteration, String dataSourceId){
JSONPathSlice r = new JSONPathSlice();
r.object = object;
r.iteration = iteration;
r.dataSourceId = dataSourceId;
r.rootId = rootId;
return r;
}
}
Loading

0 comments on commit d382c0c

Please sign in to comment.