metafacture-html/ (main): Fix Checkstyle violations.

blackwinter · blackwinter · commit 781261530c6e · 2021-09-06T10:47:12.000+02:00
diff --git a/metafacture-html/src/main/java/org/metafacture/html/ElementExtractor.java b/metafacture-html/src/main/java/org/metafacture/html/ElementExtractor.java
@@ -13,22 +13,24 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
-package org.metafacture.html;
 
-import java.io.IOException;
-import java.io.Reader;
+package org.metafacture.html;
 
-import org.apache.commons.io.IOUtils;
-import org.jsoup.Jsoup;
-import org.jsoup.nodes.Document;
-import org.jsoup.nodes.Element;
 import org.metafacture.framework.FluxCommand;
 import org.metafacture.framework.ObjectReceiver;
 import org.metafacture.framework.annotations.Description;
 import org.metafacture.framework.annotations.In;
 import org.metafacture.framework.annotations.Out;
 import org.metafacture.framework.helpers.DefaultObjectPipe;
 
+import org.apache.commons.io.IOUtils;
+import org.jsoup.Jsoup;
+import org.jsoup.nodes.Document;
+import org.jsoup.nodes.Element;
+
+import java.io.IOException;
+import java.io.Reader;
+
 /**
  * Extracts the the specified element from an HTML document
  *
@@ -51,10 +53,11 @@ public ElementExtractor(final String selector) {
     @Override
     public void process(final Reader reader) {
         try {
-            Document document = Jsoup.parse(IOUtils.toString(reader));
-            Element firstElement = document.select(selector).first();
+            final Document document = Jsoup.parse(IOUtils.toString(reader));
+            final Element firstElement = document.select(selector).first();
             getReceiver().process(firstElement.data());
-        } catch (IOException e) {
+        }
+        catch (final IOException e) {
             e.printStackTrace();
         }
     }
diff --git a/metafacture-html/src/main/java/org/metafacture/html/HtmlDecoder.java b/metafacture-html/src/main/java/org/metafacture/html/HtmlDecoder.java
@@ -13,57 +13,60 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
+
 package org.metafacture.html;
 
-import java.io.IOException;
-import java.io.Reader;
-import java.io.UnsupportedEncodingException;
-import java.net.URLDecoder;
-import java.nio.charset.StandardCharsets;
-import java.util.HashMap;
-import java.util.Map;
-import java.util.UUID;
+import org.metafacture.framework.FluxCommand;
+import org.metafacture.framework.StreamReceiver;
+import org.metafacture.framework.annotations.Description;
+import org.metafacture.framework.annotations.In;
+import org.metafacture.framework.annotations.Out;
+import org.metafacture.framework.helpers.DefaultObjectPipe;
 
 import org.apache.commons.io.IOUtils;
 import org.jsoup.Jsoup;
 import org.jsoup.nodes.Attribute;
 import org.jsoup.nodes.Attributes;
 import org.jsoup.nodes.Document;
 import org.jsoup.nodes.Element;
-import org.metafacture.framework.FluxCommand;
-import org.metafacture.framework.StreamReceiver;
-import org.metafacture.framework.annotations.Description;
-import org.metafacture.framework.annotations.In;
-import org.metafacture.framework.annotations.Out;
-import org.metafacture.framework.helpers.DefaultObjectPipe;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
+import java.io.IOException;
+import java.io.Reader;
+import java.io.UnsupportedEncodingException;
+import java.net.URLDecoder;
+import java.nio.charset.StandardCharsets;
+import java.util.HashMap;
+import java.util.Map;
+import java.util.UUID;
+
 /**
  * Decode HTML to metadata events. Each input document represents one record.
  *
  * @author Fabian Steeg (fsteeg)
  *
  */
-@Description("Decode HTML to metadata events. The attrValsAsSubfields option can be used to override "
-        + "the default attribute values to be used as subfields (e.g. by default "
-        + "`link rel=\"canonical\" href=\"http://example.org\"` becomes `link.canonical`). "
-        + "It expects an HTTP-style query string specifying as key the attributes whose value should "
-        + "be used as a subfield, and as value the attribute whose value should be the subfield value, "
-        + "e.g. the default contains `link.rel=href`. To use the HTML element text as the value "
-        + "(instead of another attribute), omit the value of the query-string key-value pair, "
-        + "e.g. `title.lang`. To add to the defaults, instead of replacing them, start with an `&`, "
-        + "e.g. `&h3.class`")
+@Description("Decode HTML to metadata events. The attrValsAsSubfields option can be used to override " +
+        "the default attribute values to be used as subfields (e.g. by default " +
+        "`link rel=\"canonical\" href=\"http://example.org\"` becomes `link.canonical`). " +
+        "It expects an HTTP-style query string specifying as key the attributes whose value should " +
+        "be used as a subfield, and as value the attribute whose value should be the subfield value, " +
+        "e.g. the default contains `link.rel=href`. To use the HTML element text as the value " +
+        "(instead of another attribute), omit the value of the query-string key-value pair, " +
+        "e.g. `title.lang`. To add to the defaults, instead of replacing them, start with an `&`, " +
+        "e.g. `&h3.class`")
 @In(Reader.class)
 @Out(StreamReceiver.class)
 @FluxCommand("decode-html")
 public class HtmlDecoder extends DefaultObjectPipe<Reader, StreamReceiver> {
 
+    private static final Logger LOG = LoggerFactory.getLogger(HtmlDecoder.class);
+
     private static final String DEFAULT_ATTR_VALS_AS_SUBFIELDS = //
             "meta.name=content&meta.property=content&link.rel=href&a.rel=href";
+
     private Map<String, String> attrValsAsSubfields;
-    private static final Logger LOG =
-            LoggerFactory.getLogger(HtmlDecoder.class);
 
     public HtmlDecoder() {
         setAttrValsAsSubfields(DEFAULT_ATTR_VALS_AS_SUBFIELDS);
@@ -72,28 +75,29 @@ public HtmlDecoder() {
     @Override
     public void process(final Reader reader) {
         try {
-            StreamReceiver receiver = getReceiver();
+            final StreamReceiver receiver = getReceiver();
             receiver.startRecord(UUID.randomUUID().toString());
-            Document document = Jsoup.parse(IOUtils.toString(reader));
+            final Document document = Jsoup.parse(IOUtils.toString(reader));
             process(document, receiver);
             receiver.endRecord();
-        } catch (IOException e) {
+        }
+        catch (final IOException e) {
             LOG.error(e.getMessage(), e);
         }
     }
 
-    private void process(Element parent, StreamReceiver receiver) {
-        for (Element element : parent.children()) {
+    private void process(final Element parent, final StreamReceiver receiver) {
+        for (final Element element : parent.children()) {
             receiver.startEntity(element.nodeName());
-            Attributes attributes = element.attributes();
+            final Attributes attributes = element.attributes();
             boolean addedValueAsSubfield = false;
-            for (Attribute attribute : attributes) {
+            for (final Attribute attribute : attributes) {
                 addedValueAsSubfield = handleAttributeValuesAsSubfields(receiver, element, attributes, attribute);
                 receiver.literal(attribute.getKey(), attribute.getValue());
             }
             if (element.children().isEmpty()) {
-                String text = element.text().trim();
-                String value = text.isEmpty() ? element.data() : text;
+                final String text = element.text().trim();
+                final String value = text.isEmpty() ? element.data() : text;
                 if (!value.isEmpty() && !addedValueAsSubfield) {
                     receiver.literal("value", value);
                 }
@@ -103,34 +107,34 @@ private void process(Element parent, StreamReceiver receiver) {
         }
     }
 
-    private boolean handleAttributeValuesAsSubfields(StreamReceiver receiver, Element element,
-            Attributes attributes, Attribute attribute) {
-        String fullFieldKey = element.nodeName() + "." + attribute.getKey();
+    private boolean handleAttributeValuesAsSubfields(final StreamReceiver receiver, final Element element, final Attributes attributes, final Attribute attribute) {
+        final String fullFieldKey = element.nodeName() + "." + attribute.getKey();
         if (attrValsAsSubfields.containsKey(fullFieldKey)) {
-            String configValue = attrValsAsSubfields.get(fullFieldKey);
+            final String configValue = attrValsAsSubfields.get(fullFieldKey);
             if (configValue.trim().isEmpty()) {
                 receiver.literal(attribute.getValue(), element.text().trim());
                 return true;
-            } else {
-                String value = attributes.get(configValue);
+            }
+            else {
+                final String value = attributes.get(configValue);
                 receiver.literal(attribute.getValue(), value);
             }
         }
         return false;
     }
 
-    public void setAttrValsAsSubfields(String mapString) {
+    public void setAttrValsAsSubfields(final String mapString) {
         this.attrValsAsSubfields = new HashMap<String, String>();
-        String input = mapString.startsWith("&") ? DEFAULT_ATTR_VALS_AS_SUBFIELDS + mapString
-                : mapString;
-        for (String nameValuePair : input.split("&")) {
-            String[] nameValue = nameValuePair.split("=");
+        final String input = mapString.startsWith("&") ? DEFAULT_ATTR_VALS_AS_SUBFIELDS + mapString : mapString;
+        for (final String nameValuePair : input.split("&")) {
+            final String[] nameValue = nameValuePair.split("=");
             try {
-                String utf8 = StandardCharsets.UTF_8.name();
-                String key = URLDecoder.decode(nameValue[0], utf8);
-                String val = nameValue.length > 1 ? URLDecoder.decode(nameValue[1], utf8) : "";
+                final String utf8 = StandardCharsets.UTF_8.name();
+                final String key = URLDecoder.decode(nameValue[0], utf8);
+                final String val = nameValue.length > 1 ? URLDecoder.decode(nameValue[1], utf8) : "";
                 attrValsAsSubfields.put(key, val);
-            } catch (UnsupportedEncodingException e) {
+            }
+            catch (final UnsupportedEncodingException e) {
                 LOG.error(e.getMessage(), e);
             }
         }
diff --git a/metafacture-html/src/test/java/org/metafacture/html/ElementExtractorTest.java b/metafacture-html/src/test/java/org/metafacture/html/ElementExtractorTest.java
@@ -13,6 +13,7 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
+
 package org.metafacture.html;
 
 import static org.mockito.Mockito.verify;
diff --git a/metafacture-html/src/test/java/org/metafacture/html/HtmlDecoderTest.java b/metafacture-html/src/test/java/org/metafacture/html/HtmlDecoderTest.java
@@ -13,6 +13,7 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
+
 package org.metafacture.html;
 
 import static org.mockito.Mockito.inOrder;
