eclipse-rdf4jGH-5058: additional parser code (WIP)

Fedict · Jul 15, 2024 · f40a74a · f40a74a
1 parent ed1e748
commit f40a74a
Show file tree

Hide file tree

Showing 3 changed files with 50 additions and 27 deletions.
diff --git a/core/rio/csvw/src/main/java/org/eclipse/rdf4j/rio/csvw/CSVWParser.java b/core/rio/csvw/src/main/java/org/eclipse/rdf4j/rio/csvw/CSVWParser.java
@@ -218,7 +218,6 @@ private List<Value> getColumns(Model metadata, Resource tableSchema) throws RDFP
 		return RDFCollections.asValues(metadata, head.get(), new ArrayList<>());
 	}
 
-
 	/**
 	 * Get "about" URL template, to be used to create the subject of the triples
 	 *
@@ -392,7 +391,8 @@ private void parseCSV(Model metadata, RDFHandler handler, URI csvFile, CellParse
 					if (doReplace) {
 						values.put("{_col}", Long.toString(i));
 					}
-					handleStatement(handler, cellParsers[i], null, aboutSubject, values);
+					System.err.println("column: " + i);
+					handleStatement(handler, cellParsers[i], aboutSubject, values, needReplacement[i]);
 				}
 				line++;
 			}
@@ -415,10 +415,7 @@ private void handleStatement(RDFHandler handler, CellParser cellParser, String c
 		IRI predicate = cellParser.getPropertyIRI();
 		Resource o = cellParser.getValueUrl(cell);
 
-		Statement stmt = Statements.statement((s != null) ? s : aboutSubject,
-				predicate,
-				(o != null) ? o : val,
-				null);
+		Statement stmt = Statements.statement((s != null) ? s : aboutSubject, predicate, (o != null) ? o : val, null);
 		handler.handleStatement(stmt);
 	}
 
@@ -432,15 +429,32 @@ private void handleStatement(RDFHandler handler, CellParser cellParser, String c
 	 */
 	private void handleStatement(RDFHandler handler, CellParser cellParser, String cell, Resource aboutSubject,
 			Map<String, String> values) {
-		Resource s = cellParser.getAboutUrl(cell);
+		Resource s = cellParser.getAboutUrl(values);
 		IRI predicate = cellParser.getPropertyIRI();
-		Resource o = cellParser.getValueUrl(cell);
-		Value val = cellParser.parse(cell);
+		Value o = cellParser.getValueUrl(values, cell);
+		if (o == null) {
+			o = cellParser.parse(cell);
+		}
+
+		Statement stmt = Statements.statement((s != null) ? s : aboutSubject, predicate, o, null);
+		handler.handleStatement(stmt);
+	}
+
+	/**
+	 * Generate statement
+	 *
+	 * @param handler
+	 * @param cellParser
+	 * @param cells
+	 * @param aboutSubject
+	 */
+	private void handleStatement(RDFHandler handler, CellParser cellParser, Resource aboutSubject,
+			Map<String, String> values, boolean needsReplacement) {
+		Resource s = cellParser.getAboutUrl(values);
+		IRI predicate = cellParser.getPropertyIRI();
+		Resource o = (needsReplacement) ? cellParser.getValueUrl(values, null) : cellParser.getValueUrl(null);
 
-		Statement stmt = Statements.statement((s != null) ? s : aboutSubject,
-				predicate,
-				(o != null) ? o : val,
-				null);
+		Statement stmt = Statements.statement((s != null) ? s : aboutSubject, predicate, o, null);
 		handler.handleStatement(stmt);
 	}
 

diff --git a/core/rio/csvw/src/main/java/org/eclipse/rdf4j/rio/csvw/CSVWUtil.java b/core/rio/csvw/src/main/java/org/eclipse/rdf4j/rio/csvw/CSVWUtil.java
@@ -10,13 +10,11 @@
  *******************************************************************************/
 package org.eclipse.rdf4j.rio.csvw;
 
-import com.opencsv.CSVParserBuilder;
-import com.opencsv.CSVReader;
-import com.opencsv.CSVReaderBuilder;
 import java.io.Reader;
 import java.nio.charset.Charset;
 import java.nio.charset.StandardCharsets;
 import java.util.Optional;
+
 import org.eclipse.rdf4j.model.IRI;
 import org.eclipse.rdf4j.model.Model;
 import org.eclipse.rdf4j.model.Resource;
@@ -28,6 +26,9 @@
 import org.eclipse.rdf4j.rio.csvw.parsers.CellParser;
 import org.eclipse.rdf4j.rio.csvw.parsers.CellParserFactory;
 
+import com.opencsv.CSVParserBuilder;
+import com.opencsv.CSVReader;
+import com.opencsv.CSVReaderBuilder;
 
 /**
  * Utility class, mostly about configuring the reader based on the JSON-LD metadata
@@ -112,7 +113,6 @@ private static IRI getDatatypeIRI(Model metadata, Resource column) {
 		return CoreDatatype.XSD.valueOf(datatype.stringValue().toUpperCase()).getIri();
 	}
 
-
 	/**
 	 * Get format string, e.g date format
 	 *

diff --git a/core/rio/csvw/src/main/java/org/eclipse/rdf4j/rio/csvw/parsers/CellParser.java b/core/rio/csvw/src/main/java/org/eclipse/rdf4j/rio/csvw/parsers/CellParser.java
@@ -16,6 +16,7 @@
 import java.util.HashSet;
 import java.util.Map;
 import java.util.Set;
+import java.util.regex.MatchResult;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 import java.util.stream.Collectors;
@@ -193,14 +194,19 @@ private String[] getPlaceholders(String template) {
 		Matcher matcher = PLACEHOLDERS.matcher(template);
 		String ownPlaceholder = getOwnPlaceholder(template);
 
-		if (matcher.find()) {
-			Set<String> placeholders = matcher.results()
-					.map(m -> m.group())
-					.filter(m -> !m.equals(ownPlaceholder))
-					.collect(Collectors.toSet());
-			return placeholders.toArray(new String[placeholders.size()]);
+		Set<String> placeholders = matcher.results()
+				.map(MatchResult::group)
+				.filter(m -> !m.equals(ownPlaceholder))
+				.collect(Collectors.toSet());
+		System.err.println("placeholders " + placeholders);
+		System.err.println("own placeholders " + ownPlaceholder);
+
+		if (placeholders.isEmpty()) {
+			System.err.println("no placeholder for " + template);
+			return null;
 		}
-		return null;
+		return placeholders.toArray(new String[placeholders.size()]);
+
 	}
 
 	/**
@@ -215,7 +221,7 @@ public IRI getAboutUrl(String cell) {
 		}
 		String s = aboutUrl;
 		if (aboutPlaceholder != null && cell != null) {
-			s = aboutUrl.replace(encodedName, getValueOrDefault(cell));
+			s = aboutUrl.replace(aboutPlaceholder, getValueOrDefault(cell));
 		}
 		return Values.iri(s);
 	}
@@ -298,7 +304,8 @@ public IRI getValueUrl(String cell) {
 		}
 		String s = valueUrl;
 		if (valuePlaceholder != null && cell != null) {
-			s = valueUrl.replace(encodedName, getValueOrDefault(cell));
+			System.err.println("repace " + valuePlaceholder + " " + cell);
+			s = valueUrl.replace(valuePlaceholder, getValueOrDefault(cell));
 		}
 		return Values.iri(s);
 	}
@@ -315,7 +322,7 @@ public IRI getValueUrl(Map<String, String> values, String cell) {
 			return null;
 		}
 		String s = valueUrl;
-		if (valuePlaceholder != null) {
+		if (valuePlaceholder != null && cell != null) {
 			s = valueUrl.replace(encodedName, getValueOrDefault(cell));
 		}
 		for (String val : valuePlaceholders) {
@@ -331,7 +338,9 @@ public IRI getValueUrl(Map<String, String> values, String cell) {
 	 */
 	public void setValueUrl(String valueUrl) {
 		this.valueUrl = valueUrl;
+		System.err.println("valueurl " + valueUrl);
 		// check if this URL contains column placeholders
+		this.valuePlaceholder = getOwnPlaceholder(valueUrl);
 		this.valuePlaceholders = getPlaceholders(valueUrl);
 	}