Fork merge: Diagram-Text tracelinks, Improvements to the MME detection capabilities

.editorconfig

@@ -15,7 +15,6 @@ ij_visual_guides = none
 ij_wrap_on_typing = false
 
 [*.{kt,kts}]
-insert_final_newline = true
 ij_kotlin_allow_trailing_comma = false
 ij_kotlin_allow_trailing_comma_on_call_site = false
 ij_kotlin_name_count_to_use_star_import = 999
@@ -118,7 +117,7 @@ ij_java_for_statement_wrap = off
 ij_java_generate_final_locals = false
 ij_java_generate_final_parameters = false
 ij_java_if_brace_force = never
-ij_java_imports_layout = $*,|,java.**,javax.**,|,org.**,com.**,|,*
+ij_java_imports_layout = $*, |, java.**, javax.**, |, org.**, com.**, |, *
 ij_java_indent_case_from_switch = false
 ij_java_insert_inner_class_imports = false
 ij_java_insert_override_annotation = true
@@ -155,7 +154,7 @@ ij_java_names_count_to_use_import_on_demand = 3
 ij_java_new_line_after_lparen_in_annotation = false
 ij_java_new_line_after_lparen_in_deconstruction_pattern = true
 ij_java_new_line_after_lparen_in_record_header = false
-ij_java_packages_to_use_import_on_demand = java.awt.*,javax.swing.*
+ij_java_packages_to_use_import_on_demand = java.awt.*, javax.swing.*
 ij_java_parameter_annotation_wrap = off
 ij_java_parentheses_expression_new_line_after_left_paren = false
 ij_java_parentheses_expression_right_paren_on_new_line = false
@@ -269,3 +268,4 @@ ij_java_while_on_new_line = false
 ij_java_wrap_comments = true
 ij_java_wrap_first_method_in_call_chain = false
 ij_java_wrap_long_lines = false
+

framework/common/pom.xml

@@ -32,6 +32,11 @@
       <groupId>org.apache.commons</groupId>
       <artifactId>commons-text</artifactId>
     </dependency>
+    <dependency>
+      <groupId>org.apache.jena</groupId>
+      <artifactId>jena-arq</artifactId>
+      <version>4.8.0</version>
+    </dependency>
     <dependency>
       <groupId>org.apache.opennlp</groupId>
       <artifactId>opennlp-tools</artifactId>
@@ -50,6 +55,11 @@
       <artifactId>jgrapht-core</artifactId>
       <version>${jgrapht.version}</version>
     </dependency>
+    <dependency>
+      <groupId>org.jsoup</groupId>
+      <artifactId>jsoup</artifactId>
+      <version>1.16.1</version>
+    </dependency>
     <dependency>
       <groupId>org.junit.jupiter</groupId>
       <artifactId>junit-jupiter-engine</artifactId>

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/Disambiguation.java

@@ -0,0 +1,153 @@
+/* Licensed under MIT 2023-2024. */
+package edu.kit.kastel.mcse.ardoco.core.api;
+
+import java.io.IOException;
+import java.io.Serializable;
+import java.util.*;
+
+import com.fasterxml.jackson.annotation.JsonCreator;
+import com.fasterxml.jackson.annotation.JsonProperty;
+import com.fasterxml.jackson.core.JsonGenerator;
+import com.fasterxml.jackson.databind.JsonSerializer;
+import com.fasterxml.jackson.databind.SerializerProvider;
+import com.fasterxml.jackson.databind.annotation.JsonSerialize;
+
+import edu.kit.kastel.mcse.ardoco.core.architecture.Deterministic;
+
+/**
+ * This class represents an abbreviation with a known set of meanings. An abbreviation is a string such as "ArDoCo" and has the meaning "Architecture
+ * Documentation Consistency". The abbreviation that is disambiguated by the meanings of this class is final, but the meanings can be changed. An instance of
+ * this class can be serialized and deserialized into JSON using Jackson.
+ */
+@Deterministic
+@JsonSerialize(using = Disambiguation.DisambiguationSerializer.class)
+public class Disambiguation implements Comparable<Disambiguation>, Serializable {
+    private final String abbreviation;
+    private final SortedSet<String> meanings;
+
+    public String getAbbreviation() {
+        return abbreviation;
+    }
+
+    public SortedSet<String> getMeanings() {
+        return new TreeSet<>(meanings);
+    }
+
+    @Override
+    public int compareTo(Disambiguation o) {
+        return abbreviation.compareTo(o.abbreviation);
+    }
+
+    @Override
+    public boolean equals(Object obj) {
+        if (this == obj)
+            return true;
+        if (obj instanceof Disambiguation other) {
+            return getAbbreviation().equals(other.getAbbreviation());
+        }
+        return false;
+    }
+
+    @Override
+    public int hashCode() {
+        return Objects.hash(getAbbreviation());
+    }
+
+    /**
+     * Used by the Jackson library to serialize a disambiguation into JSON format.
+     */
+    public static class DisambiguationSerializer extends JsonSerializer<Disambiguation> {
+        @Override
+        public void serialize(Disambiguation abbreviation, JsonGenerator jsonGenerator, SerializerProvider serializerProvider) throws IOException {
+            jsonGenerator.writeStartObject();
+            jsonGenerator.writeStringField("abbreviation", abbreviation.abbreviation);
+            jsonGenerator.writeArrayFieldStart("meanings");
+            var meanings = abbreviation.meanings;
+            for (var meaning : meanings) {
+                jsonGenerator.writeString(meaning);
+            }
+            jsonGenerator.writeEndArray();
+            jsonGenerator.writeEndObject();
+        }
+    }
+
+    /**
+     * Creates a new disambiguation of the provided abbreviation using the given array of meanings.
+     *
+     * @param abbreviation the abbreviation that is disambiguated by this instance
+     * @param meanings     an array of meanings for the abbreviation, may be empty
+     */
+    @JsonCreator
+    public Disambiguation(@JsonProperty("abbreviation") String abbreviation, @JsonProperty("meanings") String[] meanings) {
+        this(abbreviation, new TreeSet<>(Arrays.stream(meanings).toList()));
+    }
+
+    /**
+     * Creates a new disambiguation of the provided abbreviation using the given set of meanings.
+     *
+     * @param abbreviation the abbreviation that is disambiguated by this instance
+     * @param meanings     a set of meanings for the abbreviation, may be empty
+     */
+    public Disambiguation(String abbreviation, SortedSet<String> meanings) {
+        this.abbreviation = abbreviation;
+        this.meanings = new TreeSet<>(meanings);
+    }
+
+    /**
+     * Adds all meanings from another disambiguation to this disambiguation. Be careful, this does not perform any checks regarding the abbreviations.
+     *
+     * @param other the other disambiguation
+     * @return this
+     */
+    public Disambiguation addMeanings(Disambiguation other) {
+        meanings.addAll(other.meanings);
+        return this;
+    }
+
+    /**
+     * Searches the text for meanings contained by this disambiguation and replaces them with the abbreviation.
+     *
+     * @param text       the text to search
+     * @param ignoreCase whether letter case should be ignored when searching for the meanings
+     * @return the abbreviated text
+     */
+    public String replaceMeaningWithAbbreviation(String text, boolean ignoreCase) {
+        var abbreviatedText = text;
+        for (String meaning : meanings) {
+            String pattern = ignoreCase ? "(?i)" : "";
+            pattern += meaning;
+            abbreviatedText = abbreviatedText.replaceAll(pattern, abbreviation);
+        }
+        return abbreviatedText;
+    }
+
+    /**
+     * Merges the first map with the second map in a new map. If a key already exists, the disambiguations are merged non-destructively.
+     *
+     * @param a the first map
+     * @param b the second map
+     * @return a mutable map
+     */
+
+    public static SortedMap<String, Disambiguation> merge(SortedMap<String, Disambiguation> a, SortedMap<String, Disambiguation> b) {
+        var mergedMap = new TreeMap<>(a);
+        for (var entry : b.entrySet()) {
+            mergedMap.merge(entry.getKey(), entry.getValue(), Disambiguation::merge);
+        }
+        return mergedMap;
+    }
+
+    /**
+     * Merges the first disambiguation with the second disambiguation into a new instance.
+     *
+     * @param a first ambiguation
+     * @param b second ambiguation
+     * @return new merged disambiguation
+     */
+
+    public static Disambiguation merge(Disambiguation a, Disambiguation b) {
+        var temp = new TreeSet<>(a.meanings);
+        temp.addAll(b.meanings);
+        return new Disambiguation(a.abbreviation, temp);
+    }
+}

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/InputDiagramData.java

@@ -1,4 +1,4 @@
-/* Licensed under MIT 2023. */
+/* Licensed under MIT 2023-2024. */
 package edu.kit.kastel.mcse.ardoco.core.api;
 
 import java.io.File;
@@ -7,6 +7,7 @@
 import java.util.List;
 import java.util.Objects;
 
+import edu.kit.kastel.mcse.ardoco.core.common.tuple.Pair;
 import edu.kit.kastel.mcse.ardoco.core.data.PipelineStepData;
 
 /**
@@ -16,7 +17,9 @@ public class InputDiagramData implements PipelineStepData {
     public static final String ID = "InputDiagramData";
     private static final List<String> ALLOWED_FILE_TYPES = List.of("jpg", "png", "jpeg");
 
-    private final transient String pathToDiagrams;
+    private String pathToDiagrams;
+
+    private List<Pair<String, File>> diagramFiles;
 
     /**
      * Create the data by a directory of image files.
@@ -36,12 +39,24 @@ public InputDiagramData(File diagramDirectory) {
         this(Objects.requireNonNull(diagramDirectory).getAbsolutePath());
     }
 
+    /**
+     * Create the data by a list of image files.
+     *
+     * @param diagramFiles the diagram files
+     */
+    public InputDiagramData(List<Pair<String, File>> diagramFiles) {
+        Objects.requireNonNull(diagramFiles);
+        this.diagramFiles = diagramFiles;
+    }
+
     /**
      * Get all image files of the given directory (not recursive).
      *
      * @return a list of image files ordered by name
      */
-    public List<File> getFiles() {
+    public List<Pair<String, File>> getDiagramData() {
+        if (diagramFiles != null)
+            return diagramFiles;
         if (pathToDiagrams == null)
             return List.of();
         File directoryOfDiagrams = new File(pathToDiagrams);
@@ -51,10 +66,11 @@ public List<File> getFiles() {
         File[] allFiles = directoryOfDiagrams.listFiles();
         if (allFiles == null)
             return List.of();
-        List<File> diagrams = Arrays.stream(allFiles)
+        List<Pair<String, File>> diagrams = Arrays.stream(allFiles)
                 .filter(File::isFile)
                 .filter(f -> ALLOWED_FILE_TYPES.stream().anyMatch(t -> f.getName().toLowerCase().endsWith("." + t)))
                 .sorted(Comparator.comparing(File::getName))
+                .map(f -> new Pair<>(f.getName(), f))
                 .toList();
         logger.info("Found {} diagrams to consider.", diagrams.size());
         return diagrams;

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/InputTextData.java

@@ -1,4 +1,4 @@
-/* Licensed under MIT 2023. */
+/* Licensed under MIT 2023-2024. */
 package edu.kit.kastel.mcse.ardoco.core.api;
 
 import edu.kit.kastel.mcse.ardoco.core.data.PipelineStepData;
@@ -7,7 +7,7 @@ public class InputTextData implements PipelineStepData {
 
     public static final String ID = "InputTextData";
 
-    private transient String text;
+    private String text;
 
     public InputTextData(String text) {
         super();

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/PreprocessingData.java

@@ -1,17 +1,17 @@
-/* Licensed under MIT 2022-2023. */
+/* Licensed under MIT 2022-2024. */
 package edu.kit.kastel.mcse.ardoco.core.api;
 
 import edu.kit.kastel.mcse.ardoco.core.api.text.Text;
 import edu.kit.kastel.mcse.ardoco.core.data.PipelineStepData;
 
 /**
- * This class serves as container for different data after preprocessing to store as {@link PipelineStepData}. Right
- * now, this includes the preprocessed {@link Text} only.
+ * This class serves as container for different data after preprocessing to store as {@link PipelineStepData}. Right now, this includes the preprocessed
+ * {@link Text} only.
  */
 public class PreprocessingData implements PipelineStepData {
     public static final String ID = "PreprocessingData";
 
-    private transient Text preprocessedText;
+    private final Text preprocessedText;
 
     public PreprocessingData(Text preprocessedText) {
         super();

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/diagramconsistency/common/DiagramUtility.java

@@ -1,4 +1,4 @@
-/* Licensed under MIT 2023. */
+/* Licensed under MIT 2023-2024. */
 package edu.kit.kastel.mcse.ardoco.core.api.diagramconsistency.common;
 
 import java.util.List;
@@ -108,7 +108,7 @@ public static List<Box> getContainedBoxes(Box box, SortedMap<String, Box> boxes)
      */
     public static Box addBox(Diagram diagram, String text) {
         TextBox textBox = new TextBox(0, 0, 0, 0, 1.0, text, null);
-        Box box = new Box(String.valueOf(diagram.getBoxes().size()), new int[] { 0, 0, 0, 0 }, 1.0, null, List.of(textBox), null);
+        Box box = new Box(diagram, String.valueOf(diagram.getBoxes().size()), new int[] { 0, 0, 0, 0 }, 1.0, null, List.of(textBox), null);
 
         diagram.addBox(box);
         return box;