RED-9123: Improve performance of re-analysis (Spike)

layoutparser-service/layoutparser-service-internal-api/build.gradle.kts

@@ -7,4 +7,5 @@ description = "layoutparser-service-internal-api"
 
 dependencies {
     implementation("io.swagger.core.v3:swagger-annotations:2.2.15")
+    implementation("com.google.protobuf:protobuf-java-util:4.27.1")
 }

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentData.java

@@ -2,6 +2,11 @@ package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;
 
 import java.io.Serializable;
 
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPageProto.AllDocumentPages;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.AllDocumentPositionData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureProto.DocumentStructure;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextDataProto.AllDocumentTextData;
+
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -13,16 +18,26 @@ import lombok.experimental.FieldDefaults;
 @Builder
 @AllArgsConstructor
 @FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
-@Schema(description = "Object containing the complete document layout parsing information. It is split into 4 categories, structure, text, positions and pages: " + "The document tree structure of SemanticNodes such as Section, Paragraph, Headline, etc. " + "The text, which is stored as separate blocks of data. " + "The text positions, which are also stored as separate blocks. The Blocks are equal to the text blocks in length and order. " + "The page information.")
+@Schema(description = "Object containing the complete document layout parsing information. It is split into 4 categories, structure, text, positions and pages: "
+                      + "The document tree structure of SemanticNodes such as Section, Paragraph, Headline, etc. "
+                      + "The text, which is stored as separate blocks of data. "
+                      + "The text positions, which are also stored as separate blocks. The Blocks are equal to the text blocks in length and order. "
+                      + "The page information.")
 public class DocumentData implements Serializable {
 
     @Schema(description = "Contains information about the document's pages.")
-    DocumentPage[] documentPages;
+    AllDocumentPages documentPages;
     @Schema(description = "Contains information about the document's text.")
-    DocumentTextData[] documentTextData;
+    AllDocumentTextData documentTextData;
     @Schema(description = "Contains information about the document's text positions.")
-    DocumentPositionData[] documentPositions;
+    AllDocumentPositionData documentPositions;
     @Schema(description = "Contains information about the document's semantic structure.")
-    DocumentStructure documentStructure;
+    DocumentStructureWrapper documentStructureWrapper;
+
+
+    public DocumentStructure getDocumentStructure() {
+
+        return documentStructureWrapper.getDocumentStructure();
+    }
 
 }

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentPage.java

@@ -10,6 +10,7 @@ import lombok.Data;
 import lombok.NoArgsConstructor;
 import lombok.experimental.FieldDefaults;
 
+@Deprecated
 @Data
 @Builder
 @NoArgsConstructor

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentPositionData.java

@@ -10,6 +10,7 @@ import lombok.Data;
 import lombok.NoArgsConstructor;
 import lombok.experimental.FieldDefaults;
 
+@Deprecated
 @Data
 @Builder
 @NoArgsConstructor

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentStructure.java

@@ -16,6 +16,7 @@ import lombok.Data;
 import lombok.NoArgsConstructor;
 import lombok.experimental.FieldDefaults;
 
+@Deprecated
 @Data
 @Builder
 @NoArgsConstructor

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentStructureProto.java

@@ -0,0 +1,799 @@
+package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;
+
+// Generated by the protocol buffer compiler.  DO NOT EDIT!
+// NO CHECKED-IN PROTOBUF GENCODE
+// source: DocumentStructure.proto
+// Protobuf Java Version: 4.27.1
+@SuppressWarnings("all")
+public final class DocumentStructureProto {
+
+    private DocumentStructureProto() {}
+
+
+    static {
+        com.google.protobuf.RuntimeVersion.validateProtobufGencodeVersion(com.google.protobuf.RuntimeVersion.RuntimeDomain.PUBLIC,
+                /* major= */ 4,
+                /* minor= */ 27,
+                /* patch= */ 1,
+                /* suffix= */ "", DocumentStructureProto.class.getName());
+    }
+
+    public static void registerAllExtensions(com.google.protobuf.ExtensionRegistryLite registry) {
+
+    }
+
+
+    public static void registerAllExtensions(com.google.protobuf.ExtensionRegistry registry) {
+
+        registerAllExtensions((com.google.protobuf.ExtensionRegistryLite) registry);
+    }
+
+
+    public interface DocumentStructureOrBuilder extends
+            // @@protoc_insertion_point(interface_extends:DocumentStructure)
+            com.google.protobuf.MessageOrBuilder {
+
+        /**
+         * <pre>
+         * The root EntryData represents the Document.
+         * </pre>
+         *
+         * <code>.EntryData root = 1;</code>
+         *
+         * @return Whether the root field is set.
+         */
+        boolean hasRoot();
+
+
+        /**
+         * <pre>
+         * The root EntryData represents the Document.
+         * </pre>
+         *
+         * <code>.EntryData root = 1;</code>
+         *
+         * @return The root.
+         */
+        EntryDataProto.EntryData getRoot();
+
+
+        /**
+         * <pre>
+         * The root EntryData represents the Document.
+         * </pre>
+         *
+         * <code>.EntryData root = 1;</code>
+         */
+        EntryDataProto.EntryDataOrBuilder getRootOrBuilder();
+
+    }
+
+    /**
+     * Protobuf type {@code DocumentStructure}
+     */
+    public static final class DocumentStructure extends com.google.protobuf.GeneratedMessage implements
+            // @@protoc_insertion_point(message_implements:DocumentStructure)
+            DocumentStructureOrBuilder {
+
+        private static final long serialVersionUID = 0L;
+
+        static {
+            com.google.protobuf.RuntimeVersion.validateProtobufGencodeVersion(com.google.protobuf.RuntimeVersion.RuntimeDomain.PUBLIC,
+                    /* major= */ 4,
+                    /* minor= */ 27,
+                    /* patch= */ 1,
+                    /* suffix= */ "", DocumentStructure.class.getName());
+        }
+
+        // Use DocumentStructure.newBuilder() to construct.
+        private DocumentStructure(com.google.protobuf.GeneratedMessage.Builder<?> builder) {
+
+            super(builder);
+        }
+
+
+        private DocumentStructure() {
+
+        }
+
+
+        public static final com.google.protobuf.Descriptors.Descriptor getDescriptor() {
+
+            return DocumentStructureProto.internal_static_DocumentStructure_descriptor;
+        }
+
+
+        @Override
+        protected FieldAccessorTable internalGetFieldAccessorTable() {
+
+            return DocumentStructureProto.internal_static_DocumentStructure_fieldAccessorTable.ensureFieldAccessorsInitialized(DocumentStructure.class, Builder.class);
+        }
+
+
+        private int bitField0_;
+        public static final int ROOT_FIELD_NUMBER = 1;
+        private EntryDataProto.EntryData root_;
+
+
+        /**
+         * <pre>
+         * The root EntryData represents the Document.
+         * </pre>
+         *
+         * <code>.EntryData root = 1;</code>
+         *
+         * @return Whether the root field is set.
+         */
+        @Override
+        public boolean hasRoot() {
+
+            return ((bitField0_ & 0x00000001) != 0);
+        }
+
+
+        /**
+         * <pre>
+         * The root EntryData represents the Document.
+         * </pre>
+         *
+         * <code>.EntryData root = 1;</code>
+         *
+         * @return The root.
+         */
+        @Override
+        public EntryDataProto.EntryData getRoot() {
+
+            return root_ == null ? EntryDataProto.EntryData.getDefaultInstance() : root_;
+        }
+
+
+        /**
+         * <pre>
+         * The root EntryData represents the Document.
+         * </pre>
+         *
+         * <code>.EntryData root = 1;</code>
+         */
+        @Override
+        public EntryDataProto.EntryDataOrBuilder getRootOrBuilder() {
+
+            return root_ == null ? EntryDataProto.EntryData.getDefaultInstance() : root_;
+        }
+
+
+        private byte memoizedIsInitialized = -1;
+
+
+        @Override
+        public final boolean isInitialized() {
+
+            byte isInitialized = memoizedIsInitialized;
+          if (isInitialized == 1) {
+            return true;
+          }
+          if (isInitialized == 0) {
+            return false;
+          }
+
+            memoizedIsInitialized = 1;
+            return true;
+        }
+
+
+        @Override
+        public void writeTo(com.google.protobuf.CodedOutputStream output) throws java.io.IOException {
+
+            if (((bitField0_ & 0x00000001) != 0)) {
+                output.writeMessage(1, getRoot());
+            }
+            getUnknownFields().writeTo(output);
+        }
+
+
+        @Override
+        public int getSerializedSize() {
+
+            int size = memoizedSize;
+          if (size != -1) {
+            return size;
+          }
+
+            size = 0;
+            if (((bitField0_ & 0x00000001) != 0)) {
+                size += com.google.protobuf.CodedOutputStream.computeMessageSize(1, getRoot());
+            }
+            size += getUnknownFields().getSerializedSize();
+            memoizedSize = size;
+            return size;
+        }
+
+
+        @Override
+        public boolean equals(final Object obj) {
+
+            if (obj == this) {
+                return true;
+            }
+            if (!(obj instanceof DocumentStructure)) {
+                return super.equals(obj);
+            }
+            DocumentStructure other = (DocumentStructure) obj;
+
+          if (hasRoot() != other.hasRoot()) {
+            return false;
+          }
+            if (hasRoot()) {
+              if (!getRoot().equals(other.getRoot())) {
+                return false;
+              }
+            }
+          if (!getUnknownFields().equals(other.getUnknownFields())) {
+            return false;
+          }
+            return true;
+        }
+
+
+        @Override
+        public int hashCode() {
+
+            if (memoizedHashCode != 0) {
+                return memoizedHashCode;
+            }
+            int hash = 41;
+            hash = (19 * hash) + getDescriptor().hashCode();
+            if (hasRoot()) {
+                hash = (37 * hash) + ROOT_FIELD_NUMBER;
+                hash = (53 * hash) + getRoot().hashCode();
+            }
+            hash = (29 * hash) + getUnknownFields().hashCode();
+            memoizedHashCode = hash;
+            return hash;
+        }
+
+
+        public static DocumentStructure parseFrom(java.nio.ByteBuffer data) throws com.google.protobuf.InvalidProtocolBufferException {
+
+            return PARSER.parseFrom(data);
+        }
+
+
+        public static DocumentStructure parseFrom(java.nio.ByteBuffer data,
+                                                  com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws com.google.protobuf.InvalidProtocolBufferException {
+
+            return PARSER.parseFrom(data, extensionRegistry);
+        }
+
+
+        public static DocumentStructure parseFrom(com.google.protobuf.ByteString data) throws com.google.protobuf.InvalidProtocolBufferException {
+
+            return PARSER.parseFrom(data);
+        }
+
+
+        public static DocumentStructure parseFrom(com.google.protobuf.ByteString data,
+                                                  com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws com.google.protobuf.InvalidProtocolBufferException {
+
+            return PARSER.parseFrom(data, extensionRegistry);
+        }
+
+
+        public static DocumentStructure parseFrom(byte[] data) throws com.google.protobuf.InvalidProtocolBufferException {
+
+            return PARSER.parseFrom(data);
+        }
+
+
+        public static DocumentStructure parseFrom(byte[] data,
+                                                  com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws com.google.protobuf.InvalidProtocolBufferException {
+
+            return PARSER.parseFrom(data, extensionRegistry);
+        }
+
+
+        public static DocumentStructure parseFrom(java.io.InputStream input) throws java.io.IOException {
+
+            return com.google.protobuf.GeneratedMessage.parseWithIOException(PARSER, input);
+        }
+
+
+        public static DocumentStructure parseFrom(java.io.InputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws java.io.IOException {
+
+            return com.google.protobuf.GeneratedMessage.parseWithIOException(PARSER, input, extensionRegistry);
+        }
+
+
+        public static DocumentStructure parseDelimitedFrom(java.io.InputStream input) throws java.io.IOException {
+
+            return com.google.protobuf.GeneratedMessage.parseDelimitedWithIOException(PARSER, input);
+        }
+
+
+        public static DocumentStructure parseDelimitedFrom(java.io.InputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws java.io.IOException {
+
+            return com.google.protobuf.GeneratedMessage.parseDelimitedWithIOException(PARSER, input, extensionRegistry);
+        }
+
+
+        public static DocumentStructure parseFrom(com.google.protobuf.CodedInputStream input) throws java.io.IOException {
+
+            return com.google.protobuf.GeneratedMessage.parseWithIOException(PARSER, input);
+        }
+
+
+        public static DocumentStructure parseFrom(com.google.protobuf.CodedInputStream input,
+                                                  com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws java.io.IOException {
+
+            return com.google.protobuf.GeneratedMessage.parseWithIOException(PARSER, input, extensionRegistry);
+        }
+
+
+        @Override
+        public Builder newBuilderForType() {return newBuilder();}
+
+
+        public static Builder newBuilder() {
+
+            return DEFAULT_INSTANCE.toBuilder();
+        }
+
+
+        public static Builder newBuilder(DocumentStructure prototype) {
+
+            return DEFAULT_INSTANCE.toBuilder().mergeFrom(prototype);
+        }
+
+
+        @Override
+        public Builder toBuilder() {
+
+            return this == DEFAULT_INSTANCE ? new Builder() : new Builder().mergeFrom(this);
+        }
+
+
+        @Override
+        protected Builder newBuilderForType(BuilderParent parent) {
+
+            Builder builder = new Builder(parent);
+            return builder;
+        }
+
+
+        /**
+         * Protobuf type {@code DocumentStructure}
+         */
+        public static final class Builder extends com.google.protobuf.GeneratedMessage.Builder<Builder> implements
+                // @@protoc_insertion_point(builder_implements:DocumentStructure)
+                DocumentStructureOrBuilder {
+
+            public static final com.google.protobuf.Descriptors.Descriptor getDescriptor() {
+
+                return DocumentStructureProto.internal_static_DocumentStructure_descriptor;
+            }
+
+
+            @Override
+            protected FieldAccessorTable internalGetFieldAccessorTable() {
+
+                return DocumentStructureProto.internal_static_DocumentStructure_fieldAccessorTable.ensureFieldAccessorsInitialized(DocumentStructure.class, Builder.class);
+            }
+
+
+            // Construct using DocumentStructureOuterClass.DocumentStructure.newBuilder()
+            private Builder() {
+
+                maybeForceBuilderInitialization();
+            }
+
+
+            private Builder(BuilderParent parent) {
+
+                super(parent);
+                maybeForceBuilderInitialization();
+            }
+
+
+            private void maybeForceBuilderInitialization() {
+
+                if (com.google.protobuf.GeneratedMessage.alwaysUseFieldBuilders) {
+                    getRootFieldBuilder();
+                }
+            }
+
+
+            @Override
+            public Builder clear() {
+
+                super.clear();
+                bitField0_ = 0;
+                root_ = null;
+                if (rootBuilder_ != null) {
+                    rootBuilder_.dispose();
+                    rootBuilder_ = null;
+                }
+                return this;
+            }
+
+
+            @Override
+            public com.google.protobuf.Descriptors.Descriptor getDescriptorForType() {
+
+                return DocumentStructureProto.internal_static_DocumentStructure_descriptor;
+            }
+
+
+            @Override
+            public DocumentStructure getDefaultInstanceForType() {
+
+                return DocumentStructure.getDefaultInstance();
+            }
+
+
+            @Override
+            public DocumentStructure build() {
+
+                DocumentStructure result = buildPartial();
+                if (!result.isInitialized()) {
+                    throw newUninitializedMessageException(result);
+                }
+                return result;
+            }
+
+
+            @Override
+            public DocumentStructure buildPartial() {
+
+                DocumentStructure result = new DocumentStructure(this);
+                if (bitField0_ != 0) {
+                    buildPartial0(result);
+                }
+                onBuilt();
+                return result;
+            }
+
+
+            private void buildPartial0(DocumentStructure result) {
+
+                int from_bitField0_ = bitField0_;
+                int to_bitField0_ = 0;
+                if (((from_bitField0_ & 0x00000001) != 0)) {
+                    result.root_ = rootBuilder_ == null ? root_ : rootBuilder_.build();
+                    to_bitField0_ |= 0x00000001;
+                }
+                result.bitField0_ |= to_bitField0_;
+            }
+
+
+            @Override
+            public Builder mergeFrom(com.google.protobuf.Message other) {
+
+                if (other instanceof DocumentStructure) {
+                    return mergeFrom((DocumentStructure) other);
+                } else {
+                    super.mergeFrom(other);
+                    return this;
+                }
+            }
+
+
+            public Builder mergeFrom(DocumentStructure other) {
+
+              if (other == DocumentStructure.getDefaultInstance()) {
+                return this;
+              }
+                if (other.hasRoot()) {
+                    mergeRoot(other.getRoot());
+                }
+                this.mergeUnknownFields(other.getUnknownFields());
+                onChanged();
+                return this;
+            }
+
+
+            @Override
+            public final boolean isInitialized() {
+
+                return true;
+            }
+
+
+            @Override
+            public Builder mergeFrom(com.google.protobuf.CodedInputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws java.io.IOException {
+
+                if (extensionRegistry == null) {
+                    throw new NullPointerException();
+                }
+                try {
+                    boolean done = false;
+                    while (!done) {
+                        int tag = input.readTag();
+                        switch (tag) {
+                            case 0:
+                                done = true;
+                                break;
+                            case 10: {
+                                input.readMessage(getRootFieldBuilder().getBuilder(), extensionRegistry);
+                                bitField0_ |= 0x00000001;
+                                break;
+                            } // case 10
+                            default: {
+                                if (!super.parseUnknownField(input, extensionRegistry, tag)) {
+                                    done = true; // was an endgroup tag
+                                }
+                                break;
+                            } // default:
+                        } // switch (tag)
+                    } // while (!done)
+                } catch (com.google.protobuf.InvalidProtocolBufferException e) {
+                    throw e.unwrapIOException();
+                } finally {
+                    onChanged();
+                } // finally
+                return this;
+            }
+
+
+            private int bitField0_;
+
+            private EntryDataProto.EntryData root_;
+            private com.google.protobuf.SingleFieldBuilder<EntryDataProto.EntryData, EntryDataProto.EntryData.Builder, EntryDataProto.EntryDataOrBuilder> rootBuilder_;
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             *
+             * @return Whether the root field is set.
+             */
+            public boolean hasRoot() {
+
+                return ((bitField0_ & 0x00000001) != 0);
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             *
+             * @return The root.
+             */
+            public EntryDataProto.EntryData getRoot() {
+
+                if (rootBuilder_ == null) {
+                    return root_ == null ? EntryDataProto.EntryData.getDefaultInstance() : root_;
+                } else {
+                    return rootBuilder_.getMessage();
+                }
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             */
+            public Builder setRoot(EntryDataProto.EntryData value) {
+
+                if (rootBuilder_ == null) {
+                    if (value == null) {
+                        throw new NullPointerException();
+                    }
+                    root_ = value;
+                } else {
+                    rootBuilder_.setMessage(value);
+                }
+                bitField0_ |= 0x00000001;
+                onChanged();
+                return this;
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             */
+            public Builder setRoot(EntryDataProto.EntryData.Builder builderForValue) {
+
+                if (rootBuilder_ == null) {
+                    root_ = builderForValue.build();
+                } else {
+                    rootBuilder_.setMessage(builderForValue.build());
+                }
+                bitField0_ |= 0x00000001;
+                onChanged();
+                return this;
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             */
+            public Builder mergeRoot(EntryDataProto.EntryData value) {
+
+                if (rootBuilder_ == null) {
+                    if (((bitField0_ & 0x00000001) != 0) && root_ != null && root_ != EntryDataProto.EntryData.getDefaultInstance()) {
+                        getRootBuilder().mergeFrom(value);
+                    } else {
+                        root_ = value;
+                    }
+                } else {
+                    rootBuilder_.mergeFrom(value);
+                }
+                if (root_ != null) {
+                    bitField0_ |= 0x00000001;
+                    onChanged();
+                }
+                return this;
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             */
+            public Builder clearRoot() {
+
+                bitField0_ = (bitField0_ & ~0x00000001);
+                root_ = null;
+                if (rootBuilder_ != null) {
+                    rootBuilder_.dispose();
+                    rootBuilder_ = null;
+                }
+                onChanged();
+                return this;
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             */
+            public EntryDataProto.EntryData.Builder getRootBuilder() {
+
+                bitField0_ |= 0x00000001;
+                onChanged();
+                return getRootFieldBuilder().getBuilder();
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             */
+            public EntryDataProto.EntryDataOrBuilder getRootOrBuilder() {
+
+                if (rootBuilder_ != null) {
+                    return rootBuilder_.getMessageOrBuilder();
+                } else {
+                    return root_ == null ? EntryDataProto.EntryData.getDefaultInstance() : root_;
+                }
+            }
+
+
+            /**
+             * <pre>
+             * The root EntryData represents the Document.
+             * </pre>
+             *
+             * <code>.EntryData root = 1;</code>
+             */
+            private com.google.protobuf.SingleFieldBuilder<EntryDataProto.EntryData, EntryDataProto.EntryData.Builder, EntryDataProto.EntryDataOrBuilder> getRootFieldBuilder() {
+
+                if (rootBuilder_ == null) {
+                    rootBuilder_ = new com.google.protobuf.SingleFieldBuilder<EntryDataProto.EntryData, EntryDataProto.EntryData.Builder, EntryDataProto.EntryDataOrBuilder>(getRoot(),
+                                                                                                                                                                             getParentForChildren(),
+                                                                                                                                                                             isClean());
+                    root_ = null;
+                }
+                return rootBuilder_;
+            }
+
+            // @@protoc_insertion_point(builder_scope:DocumentStructure)
+        }
+
+        // @@protoc_insertion_point(class_scope:DocumentStructure)
+        private static final DocumentStructure DEFAULT_INSTANCE;
+
+        static {
+            DEFAULT_INSTANCE = new DocumentStructure();
+        }
+
+        public static DocumentStructure getDefaultInstance() {
+
+            return DEFAULT_INSTANCE;
+        }
+
+
+        private static final com.google.protobuf.Parser<DocumentStructure> PARSER = new com.google.protobuf.AbstractParser<DocumentStructure>() {
+            @Override
+            public DocumentStructure parsePartialFrom(com.google.protobuf.CodedInputStream input,
+                                                      com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws com.google.protobuf.InvalidProtocolBufferException {
+
+                Builder builder = newBuilder();
+                try {
+                    builder.mergeFrom(input, extensionRegistry);
+                } catch (com.google.protobuf.InvalidProtocolBufferException e) {
+                    throw e.setUnfinishedMessage(builder.buildPartial());
+                } catch (com.google.protobuf.UninitializedMessageException e) {
+                    throw e.asInvalidProtocolBufferException().setUnfinishedMessage(builder.buildPartial());
+                } catch (java.io.IOException e) {
+                    throw new com.google.protobuf.InvalidProtocolBufferException(e).setUnfinishedMessage(builder.buildPartial());
+                }
+                return builder.buildPartial();
+            }
+        };
+
+
+        public static com.google.protobuf.Parser<DocumentStructure> parser() {
+
+            return PARSER;
+        }
+
+
+        @Override
+        public com.google.protobuf.Parser<DocumentStructure> getParserForType() {
+
+            return PARSER;
+        }
+
+
+        @Override
+        public DocumentStructure getDefaultInstanceForType() {
+
+            return DEFAULT_INSTANCE;
+        }
+
+    }
+
+    private static final com.google.protobuf.Descriptors.Descriptor internal_static_DocumentStructure_descriptor;
+    private static final com.google.protobuf.GeneratedMessage.FieldAccessorTable internal_static_DocumentStructure_fieldAccessorTable;
+
+
+    public static com.google.protobuf.Descriptors.FileDescriptor getDescriptor() {
+
+        return descriptor;
+    }
+
+
+    private static com.google.protobuf.Descriptors.FileDescriptor descriptor;
+
+    static {
+        String[] descriptorData = {"\n\027DocumentStructure.proto\032\017EntryData.pro"
+                                   + "to\"-\n\021DocumentStructure\022\030\n\004root\030\001 \001(\0132\n."
+                                   + "EntryDatab\006proto3"};
+        descriptor = com.google.protobuf.Descriptors.FileDescriptor.internalBuildGeneratedFileFrom(descriptorData,
+                                                                                                   new com.google.protobuf.Descriptors.FileDescriptor[]{EntryDataProto.getDescriptor(),});
+        internal_static_DocumentStructure_descriptor = getDescriptor().getMessageTypes()
+                .get(0);
+        internal_static_DocumentStructure_fieldAccessorTable = new com.google.protobuf.GeneratedMessage.FieldAccessorTable(internal_static_DocumentStructure_descriptor,
+                                                                                                                           new String[]{"Root",});
+        descriptor.resolveAllFeaturesImmutable();
+        EntryDataProto.getDescriptor();
+    }
+
+    // @@protoc_insertion_point(outer_class_scope)
+}

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentStructureWrapper.java

@@ -0,0 +1,126 @@
+package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;
+
+import static com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureProto.DocumentStructure;
+
+import java.awt.geom.Rectangle2D;
+import java.io.ObjectStreamException;
+import java.io.Serializable;
+import java.util.Arrays;
+import java.util.List;
+import java.util.stream.Stream;
+
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.EntryDataProto.EntryData;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.AllArgsConstructor;
+import lombok.Getter;
+
+@Getter
+@AllArgsConstructor
+public class DocumentStructureWrapper implements Serializable {
+
+    private final DocumentStructure documentStructure;
+
+
+    @Schema(description = "Object containing the extra field names, a table has in its properties field.")
+    public static class TableProperties implements Serializable {
+
+        public static final String NUMBER_OF_ROWS = "numberOfRows";
+        public static final String NUMBER_OF_COLS = "numberOfCols";
+
+    }
+
+    @Schema(description = "Object containing the extra field names, an Image has in its properties field.")
+    public static class ImageProperties implements Serializable {
+
+        public static final String TRANSPARENT = "transparent";
+        public static final String IMAGE_TYPE = "imageType";
+        public static final String POSITION = "position";
+        public static final String ID = "id";
+
+        public static final String REPRESENTATION_HASH = "representationHash";
+
+    }
+
+    @Schema(description = "Object containing the extra field names, a table cell has in its properties field.")
+    public static class TableCellProperties implements Serializable {
+
+        public static final String B_BOX = "bBox";
+        public static final String ROW = "row";
+        public static final String COL = "col";
+        public static final String HEADER = "header";
+
+    }
+
+    @Schema(description = "Object containing the extra field names, a duplicate paragraph has in its properties field.")
+    public static class DuplicateParagraphProperties implements Serializable {
+
+        public static final String UNSORTED_TEXTBLOCK_ID = "utbid";
+
+    }
+
+    public static final String RECTANGLE_DELIMITER = ";";
+
+
+    public static Rectangle2D parseRectangle2D(String bBox) {
+
+        List<Float> floats = Arrays.stream(bBox.split(RECTANGLE_DELIMITER))
+                .map(Float::parseFloat)
+                .toList();
+        return new Rectangle2D.Float(floats.get(0), floats.get(1), floats.get(2), floats.get(3));
+    }
+
+
+    public static double[] parseRepresentationVector(String representationHash) {
+
+        String[] stringArray = representationHash.split("[,\\s]+");
+        double[] doubleArray = new double[stringArray.length];
+        for (int i = 0; i < stringArray.length; i++) {
+            doubleArray[i] = Double.parseDouble(stringArray[i]);
+        }
+
+        return doubleArray;
+    }
+
+
+    public EntryData get(List<Integer> tocId) {
+
+        if (tocId.isEmpty()) {
+            return documentStructure.getRoot();
+        }
+        EntryData entry = documentStructure.getRoot().getChildrenList()
+                .get(tocId.get(0));
+        for (int id : tocId.subList(1, tocId.size())) {
+            entry = entry.getChildrenList()
+                    .get(id);
+        }
+        return entry;
+    }
+
+
+    public Stream<EntryData> streamAllEntries() {
+
+        return Stream.concat(Stream.of(documentStructure.getRoot()),
+                             documentStructure.getRoot().getChildrenList()
+                                     .stream())
+                .flatMap(DocumentStructureWrapper::flatten);
+    }
+
+
+    public String toString() {
+
+        return String.join("\n",
+                           streamAllEntries().map(EntryData::toString)
+                                   .toList());
+    }
+
+
+    private static Stream<EntryData> flatten(EntryData entry) {
+
+        return Stream.concat(Stream.of(entry),
+                             entry.getChildrenList()
+                                     .stream()
+                                     .flatMap(DocumentStructureWrapper::flatten));
+    }
+
+}

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentTextData.java

@@ -10,6 +10,7 @@ import lombok.Data;
 import lombok.NoArgsConstructor;
 import lombok.experimental.FieldDefaults;
 
+@Deprecated
 @Data
 @Builder
 @NoArgsConstructor

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/LayoutEngine.java

@@ -1,5 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;
 
+@Deprecated
 public enum LayoutEngine {
     ALGORITHM,
     AI,

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/LayoutEngineProto.java

@@ -0,0 +1,193 @@
+package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;
+
+// Generated by the protocol buffer compiler.  DO NOT EDIT!
+// NO CHECKED-IN PROTOBUF GENCODE
+// source: LayoutEngine.proto
+// Protobuf Java Version: 4.27.1
+@SuppressWarnings("all")
+public final class LayoutEngineProto {
+
+    private LayoutEngineProto() {}
+
+
+    static {
+        com.google.protobuf.RuntimeVersion.validateProtobufGencodeVersion(com.google.protobuf.RuntimeVersion.RuntimeDomain.PUBLIC,
+                /* major= */ 4,
+                /* minor= */ 27,
+                /* patch= */ 1,
+                /* suffix= */ "", LayoutEngineProto.class.getName());
+    }
+
+    public static void registerAllExtensions(com.google.protobuf.ExtensionRegistryLite registry) {
+
+    }
+
+
+    public static void registerAllExtensions(com.google.protobuf.ExtensionRegistry registry) {
+
+        registerAllExtensions((com.google.protobuf.ExtensionRegistryLite) registry);
+    }
+
+
+    /**
+     * Protobuf enum {@code LayoutEngine}
+     */
+    public enum LayoutEngine implements com.google.protobuf.ProtocolMessageEnum {
+        /**
+         * <code>ALGORITHM = 0;</code>
+         */
+        ALGORITHM(0),
+        /**
+         * <code>AI = 1;</code>
+         */
+        AI(1),
+        /**
+         * <code>OUTLINE = 2;</code>
+         */
+        OUTLINE(2),
+        UNRECOGNIZED(-1),
+        ;
+
+
+        static {
+            com.google.protobuf.RuntimeVersion.validateProtobufGencodeVersion(com.google.protobuf.RuntimeVersion.RuntimeDomain.PUBLIC,
+                    /* major= */ 4,
+                    /* minor= */ 27,
+                    /* patch= */ 1,
+                    /* suffix= */ "", LayoutEngine.class.getName());
+        }
+
+        /**
+         * <code>ALGORITHM = 0;</code>
+         */
+        public static final int ALGORITHM_VALUE = 0;
+        /**
+         * <code>AI = 1;</code>
+         */
+        public static final int AI_VALUE = 1;
+        /**
+         * <code>OUTLINE = 2;</code>
+         */
+        public static final int OUTLINE_VALUE = 2;
+
+
+        public final int getNumber() {
+
+            if (this == UNRECOGNIZED) {
+                throw new IllegalArgumentException("Can't get the number of an unknown enum value.");
+            }
+            return value;
+        }
+
+
+        /**
+         * @param value The numeric wire value of the corresponding enum entry.
+         * @return The enum associated with the given numeric wire value.
+         * @deprecated Use {@link #forNumber(int)} instead.
+         */
+        @Deprecated
+        public static LayoutEngine valueOf(int value) {
+
+            return forNumber(value);
+        }
+
+
+        /**
+         * @param value The numeric wire value of the corresponding enum entry.
+         * @return The enum associated with the given numeric wire value.
+         */
+        public static LayoutEngine forNumber(int value) {
+
+            switch (value) {
+                case 0:
+                    return ALGORITHM;
+                case 1:
+                    return AI;
+                case 2:
+                    return OUTLINE;
+                default:
+                    return null;
+            }
+        }
+
+
+        public static com.google.protobuf.Internal.EnumLiteMap<LayoutEngine> internalGetValueMap() {
+
+            return internalValueMap;
+        }
+
+
+        private static final com.google.protobuf.Internal.EnumLiteMap<LayoutEngine> internalValueMap = new com.google.protobuf.Internal.EnumLiteMap<LayoutEngine>() {
+            public LayoutEngine findValueByNumber(int number) {
+
+                return LayoutEngine.forNumber(number);
+            }
+        };
+
+
+        public final com.google.protobuf.Descriptors.EnumValueDescriptor getValueDescriptor() {
+
+            if (this == UNRECOGNIZED) {
+                throw new IllegalStateException("Can't get the descriptor of an unrecognized enum value.");
+            }
+            return getDescriptor().getValues()
+                    .get(ordinal());
+        }
+
+
+        public final com.google.protobuf.Descriptors.EnumDescriptor getDescriptorForType() {
+
+            return getDescriptor();
+        }
+
+
+        public static final com.google.protobuf.Descriptors.EnumDescriptor getDescriptor() {
+
+            return LayoutEngineProto.getDescriptor().getEnumTypes()
+                    .get(0);
+        }
+
+
+        private static final LayoutEngine[] VALUES = values();
+
+
+        public static LayoutEngine valueOf(com.google.protobuf.Descriptors.EnumValueDescriptor desc) {
+
+            if (desc.getType() != getDescriptor()) {
+                throw new IllegalArgumentException("EnumValueDescriptor is not for this type.");
+            }
+            if (desc.getIndex() == -1) {
+                return UNRECOGNIZED;
+            }
+            return VALUES[desc.getIndex()];
+        }
+
+
+        private final int value;
+
+
+        private LayoutEngine(int value) {
+
+            this.value = value;
+        }
+
+        // @@protoc_insertion_point(enum_scope:LayoutEngine)
+    }
+
+
+    public static com.google.protobuf.Descriptors.FileDescriptor getDescriptor() {
+
+        return descriptor;
+    }
+
+
+    private static com.google.protobuf.Descriptors.FileDescriptor descriptor;
+
+    static {
+        String[] descriptorData = {"\n\022LayoutEngine.proto*2\n\014LayoutEngine\022\r\n\t" + "ALGORITHM\020\000\022\006\n\002AI\020\001\022\013\n\007OUTLINE\020\002b\006proto3"};
+        descriptor = com.google.protobuf.Descriptors.FileDescriptor.internalBuildGeneratedFileFrom(descriptorData, new com.google.protobuf.Descriptors.FileDescriptor[]{});
+        descriptor.resolveAllFeaturesImmutable();
+    }
+
+    // @@protoc_insertion_point(outer_class_scope)
+}

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/NodeType.java

@@ -3,6 +3,7 @@ package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;
 import java.io.Serializable;
 import java.util.Locale;
 
+@Deprecated
 public enum NodeType implements Serializable {
     DOCUMENT,
     SECTION,

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/NodeTypeProto.java

@@ -0,0 +1,274 @@
+package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;
+
+import java.util.Locale;
+// Generated by the protocol buffer compiler.  DO NOT EDIT!
+// NO CHECKED-IN PROTOBUF GENCODE
+// source: NodeType.proto
+// Protobuf Java Version: 4.27.1
+@SuppressWarnings("all")
+public final class NodeTypeProto {
+
+    private NodeTypeProto() {}
+
+
+    static {
+        com.google.protobuf.RuntimeVersion.validateProtobufGencodeVersion(com.google.protobuf.RuntimeVersion.RuntimeDomain.PUBLIC,
+                /* major= */ 4,
+                /* minor= */ 27,
+                /* patch= */ 1,
+                /* suffix= */ "", NodeTypeProto.class.getName());
+    }
+
+    public static void registerAllExtensions(com.google.protobuf.ExtensionRegistryLite registry) {
+
+    }
+
+
+    public static void registerAllExtensions(com.google.protobuf.ExtensionRegistry registry) {
+
+        registerAllExtensions((com.google.protobuf.ExtensionRegistryLite) registry);
+    }
+
+
+    /**
+     * Protobuf enum {@code NodeType}
+     */
+    public enum NodeType implements com.google.protobuf.ProtocolMessageEnum {
+        /**
+         * <code>DOCUMENT = 0;</code>
+         */
+        DOCUMENT(0),
+        /**
+         * <code>SECTION = 1;</code>
+         */
+        SECTION(1),
+        /**
+         * <code>SUPER_SECTION = 2;</code>
+         */
+        SUPER_SECTION(2),
+        /**
+         * <code>HEADLINE = 3;</code>
+         */
+        HEADLINE(3),
+        /**
+         * <code>PARAGRAPH = 4;</code>
+         */
+        PARAGRAPH(4),
+        /**
+         * <code>TABLE = 5;</code>
+         */
+        TABLE(5),
+        /**
+         * <code>TABLE_CELL = 6;</code>
+         */
+        TABLE_CELL(6),
+        /**
+         * <code>IMAGE = 7;</code>
+         */
+        IMAGE(7),
+        /**
+         * <code>HEADER = 8;</code>
+         */
+        HEADER(8),
+        /**
+         * <code>FOOTER = 9;</code>
+         */
+        FOOTER(9),
+        UNRECOGNIZED(-1),
+        ;
+
+
+        public String toString() {
+
+            return this.name().charAt(0) + this.name().substring(1).toLowerCase(Locale.ROOT);
+        }
+
+
+        static {
+            com.google.protobuf.RuntimeVersion.validateProtobufGencodeVersion(com.google.protobuf.RuntimeVersion.RuntimeDomain.PUBLIC,
+                    /* major= */ 4,
+                    /* minor= */ 27,
+                    /* patch= */ 1,
+                    /* suffix= */ "", NodeType.class.getName());
+        }
+
+        /**
+         * <code>DOCUMENT = 0;</code>
+         */
+        public static final int DOCUMENT_VALUE = 0;
+        /**
+         * <code>SECTION = 1;</code>
+         */
+        public static final int SECTION_VALUE = 1;
+        /**
+         * <code>SUPER_SECTION = 2;</code>
+         */
+        public static final int SUPER_SECTION_VALUE = 2;
+        /**
+         * <code>HEADLINE = 3;</code>
+         */
+        public static final int HEADLINE_VALUE = 3;
+        /**
+         * <code>PARAGRAPH = 4;</code>
+         */
+        public static final int PARAGRAPH_VALUE = 4;
+        /**
+         * <code>TABLE = 5;</code>
+         */
+        public static final int TABLE_VALUE = 5;
+        /**
+         * <code>TABLE_CELL = 6;</code>
+         */
+        public static final int TABLE_CELL_VALUE = 6;
+        /**
+         * <code>IMAGE = 7;</code>
+         */
+        public static final int IMAGE_VALUE = 7;
+        /**
+         * <code>HEADER = 8;</code>
+         */
+        public static final int HEADER_VALUE = 8;
+        /**
+         * <code>FOOTER = 9;</code>
+         */
+        public static final int FOOTER_VALUE = 9;
+
+
+        public final int getNumber() {
+
+            if (this == UNRECOGNIZED) {
+                throw new IllegalArgumentException("Can't get the number of an unknown enum value.");
+            }
+            return value;
+        }
+
+
+        /**
+         * @param value The numeric wire value of the corresponding enum entry.
+         * @return The enum associated with the given numeric wire value.
+         * @deprecated Use {@link #forNumber(int)} instead.
+         */
+        @Deprecated
+        public static NodeType valueOf(int value) {
+
+            return forNumber(value);
+        }
+
+
+        /**
+         * @param value The numeric wire value of the corresponding enum entry.
+         * @return The enum associated with the given numeric wire value.
+         */
+        public static NodeType forNumber(int value) {
+
+            switch (value) {
+                case 0:
+                    return DOCUMENT;
+                case 1:
+                    return SECTION;
+                case 2:
+                    return SUPER_SECTION;
+                case 3:
+                    return HEADLINE;
+                case 4:
+                    return PARAGRAPH;
+                case 5:
+                    return TABLE;
+                case 6:
+                    return TABLE_CELL;
+                case 7:
+                    return IMAGE;
+                case 8:
+                    return HEADER;
+                case 9:
+                    return FOOTER;
+                default:
+                    return null;
+            }
+        }
+
+
+        public static com.google.protobuf.Internal.EnumLiteMap<NodeType> internalGetValueMap() {
+
+            return internalValueMap;
+        }
+
+
+        private static final com.google.protobuf.Internal.EnumLiteMap<NodeType> internalValueMap = new com.google.protobuf.Internal.EnumLiteMap<NodeType>() {
+            public NodeType findValueByNumber(int number) {
+
+                return NodeType.forNumber(number);
+            }
+        };
+
+
+        public final com.google.protobuf.Descriptors.EnumValueDescriptor getValueDescriptor() {
+
+            if (this == UNRECOGNIZED) {
+                throw new IllegalStateException("Can't get the descriptor of an unrecognized enum value.");
+            }
+            return getDescriptor().getValues()
+                    .get(ordinal());
+        }
+
+
+        public final com.google.protobuf.Descriptors.EnumDescriptor getDescriptorForType() {
+
+            return getDescriptor();
+        }
+
+
+        public static final com.google.protobuf.Descriptors.EnumDescriptor getDescriptor() {
+
+            return NodeTypeProto.getDescriptor().getEnumTypes()
+                    .get(0);
+        }
+
+
+        private static final NodeType[] VALUES = values();
+
+
+        public static NodeType valueOf(com.google.protobuf.Descriptors.EnumValueDescriptor desc) {
+
+            if (desc.getType() != getDescriptor()) {
+                throw new IllegalArgumentException("EnumValueDescriptor is not for this type.");
+            }
+            if (desc.getIndex() == -1) {
+                return UNRECOGNIZED;
+            }
+            return VALUES[desc.getIndex()];
+        }
+
+
+        private final int value;
+
+
+        private NodeType(int value) {
+
+            this.value = value;
+        }
+
+        // @@protoc_insertion_point(enum_scope:NodeType)
+    }
+
+
+    public static com.google.protobuf.Descriptors.FileDescriptor getDescriptor() {
+
+        return descriptor;
+    }
+
+
+    private static com.google.protobuf.Descriptors.FileDescriptor descriptor;
+
+    static {
+        String[] descriptorData = {"\n\016NodeType.proto*\223\001\n\010NodeType\022\014\n\010DOCUMEN"
+                                   + "T\020\000\022\013\n\007SECTION\020\001\022\021\n\rSUPER_SECTION\020\002\022\014\n\010H"
+                                   + "EADLINE\020\003\022\r\n\tPARAGRAPH\020\004\022\t\n\005TABLE\020\005\022\016\n\nT"
+                                   + "ABLE_CELL\020\006\022\t\n\005IMAGE\020\007\022\n\n\006HEADER\020\010\022\n\n\006FO"
+                                   + "OTER\020\tb\006proto3"};
+        descriptor = com.google.protobuf.Descriptors.FileDescriptor.internalBuildGeneratedFileFrom(descriptorData, new com.google.protobuf.Descriptors.FileDescriptor[]{});
+        descriptor.resolveAllFeaturesImmutable();
+    }
+
+    // @@protoc_insertion_point(outer_class_scope)
+}

layoutparser-service/layoutparser-service-internal-api/src/main/resources/DocumentPage.proto

@@ -0,0 +1,21 @@
+syntax = "proto3";
+
+message AllDocumentPages {
+
+  repeated DocumentPage documentPages = 1;
+}
+
+message DocumentPage {
+  // The page number, starting with 1.
+  int32 number = 1;
+
+  // The page height in PDF user units.
+  int32 height = 2;
+
+  // The page width in PDF user units.
+  int32 width = 3;
+
+  // The page rotation as specified by the PDF.
+  int32 rotation = 4;
+}
+

layoutparser-service/layoutparser-service-internal-api/src/main/resources/DocumentPositionData.proto

@@ -0,0 +1,25 @@
+syntax = "proto3";
+
+message AllDocumentPositionData {
+
+  repeated DocumentPositionData documentPositionData = 1;
+}
+
+message DocumentPositionData {
+  // Identifier of the text block.
+  int64 id = 1;
+
+  // For each string coordinate in the search text of the text block, the array contains an entry relating the string coordinate to the position coordinate.
+  // This is required due to the text and position coordinates not being equal.
+  repeated int32 stringIdxToPositionIdx = 2;
+
+  // The bounding box for each glyph as a rectangle. This matrix is of size (n,4), where n is the number of glyphs in the text block.
+  // The second dimension specifies the rectangle with the value x, y, width, height, with x, y specifying the lower left corner.
+  // In order to access this information, the stringIdxToPositionIdx array must be used to transform the coordinates.
+  repeated Position positions = 3;
+
+  // Definition of a BoundingBox that contains x, y, width, and height.
+  message Position {
+    repeated float value = 1;
+  }
+}

layoutparser-service/layoutparser-service-internal-api/src/main/resources/DocumentStructure.proto

@@ -0,0 +1,8 @@
+syntax = "proto3";
+
+import "EntryData.proto";
+
+message DocumentStructure {
+  // The root EntryData represents the Document.
+  EntryData root = 1;
+}

layoutparser-service/layoutparser-service-internal-api/src/main/resources/DocumentTextData.proto

@@ -0,0 +1,29 @@
+syntax = "proto3";
+
+message AllDocumentTextData {
+
+  repeated DocumentTextData documentTextData = 1;
+}
+
+message DocumentTextData {
+  // Identifier of the text block.
+  int64 id = 1;
+
+  // The page the text block occurs on.
+  int64 page = 2;
+
+  // The text of the text block.
+  string searchText = 3;
+
+  // Each text block is assigned a number on a page, starting from 0.
+  int32 numberOnPage = 4;
+
+  // The text blocks are ordered, this number represents the start of the text block as a string offset.
+  int32 start = 5;
+
+  // The text blocks are ordered, this number represents the end of the text block as a string offset.
+  int32 end = 6;
+
+  // The line breaks in the text of this semantic node in string offsets. They are exclusive end. At the end of each semantic node there is an implicit linebreak.
+  repeated int32 lineBreaks = 7;
+}

layoutparser-service/layoutparser-service-internal-api/src/main/resources/EntryData.proto

@@ -0,0 +1,27 @@
+syntax = "proto3";
+
+import "LayoutEngine.proto";
+import "NodeType.proto";
+
+message EntryData {
+  // Type of the semantic node.
+  NodeType type = 1;
+
+  // Specifies the position in the parsed tree structure.
+  repeated int32 treeId = 2;
+
+  // Specifies the text block IDs associated with this semantic node.
+  repeated int64 atomicBlockIds = 3;
+
+  // Specifies the pages this semantic node appears on.
+  repeated int64 pageNumbers = 4;
+
+  // Some semantic nodes have additional information, this information is stored in this Map.
+  map<string, string> properties = 5;
+
+  // All child Entries of this Entry.
+  repeated EntryData children = 6;
+
+  // Describes the origin of the semantic node.
+  repeated LayoutEngine engines = 7;
+}

layoutparser-service/layoutparser-service-internal-api/src/main/resources/LayoutEngine.proto

@@ -0,0 +1,7 @@
+syntax = "proto3";
+
+enum LayoutEngine {
+  ALGORITHM = 0;
+  AI = 1;
+  OUTLINE = 2;
+}

layoutparser-service/layoutparser-service-internal-api/src/main/resources/NodeType.proto

@@ -0,0 +1,14 @@
+syntax = "proto3";
+
+enum NodeType {
+  DOCUMENT = 0;
+  SECTION = 1;
+  SUPER_SECTION = 2;
+  HEADLINE = 3;
+  PARAGRAPH = 4;
+  TABLE = 5;
+  TABLE_CELL = 6;
+  IMAGE = 7;
+  HEADER = 8;
+  FOOTER = 9;
+}

layoutparser-service/layoutparser-service-processor/build.gradle.kts

@@ -16,8 +16,10 @@ dependencies {
         exclude("org.springframework.boot", "spring-boot-starter-security")
         exclude("org.springframework.boot", "spring-boot-starter-validation")
     }
-    implementation("com.knecon.fforesight:tenant-commons:0.30.0")
-    implementation("com.iqser.red.commons:storage-commons:2.45.0")
+    implementation("com.knecon.fforesight:tenant-commons:0.30.0") {
+        exclude("com.iqser.red.commons", "storage-commons")
+    }
+    implementation("com.iqser.red.commons:storage-commons:2.50.0")
 
     implementation("org.apache.pdfbox:pdfbox:${pdfBoxVersion}")
     implementation("org.apache.pdfbox:pdfbox-tools:${pdfBoxVersion}")
@@ -34,4 +36,5 @@ dependencies {
     implementation("com.pdftron:PDFNet:10.11.0")
     implementation("org.apache.commons:commons-text:1.12.0")
 
+    implementation("com.google.protobuf:protobuf-java-util:4.27.1")
 }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/LayoutParsingPipeline.java

@@ -20,7 +20,7 @@ import org.apache.pdfbox.pdmodel.common.PDRectangle;
 import org.apache.pdfbox.pdmodel.documentinterchange.markedcontent.PDMarkedContent;
 import org.springframework.stereotype.Service;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingFinishedEvent;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingRequest;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/LayoutParsingStorageService.java

@@ -11,7 +11,9 @@ import java.nio.file.Path;
 import java.nio.file.Paths;
 import java.nio.file.StandardOpenOption;
 import java.util.Optional;
+import java.util.concurrent.CompletableFuture;
 
+import org.springframework.core.task.TaskExecutor;
 import org.springframework.stereotype.Service;
 
 import com.fasterxml.jackson.databind.ObjectMapper;
@@ -39,6 +41,8 @@ public class LayoutParsingStorageService {
     private final StorageService storageService;
     private final ObjectMapper objectMapper;
 
+    private final TaskExecutor taskExecutor;
+
 
     @Observed(name = "LayoutParsingStorageService", contextualName = "get-origin-file")
     public File getOriginFile(String storageId) throws IOException {
@@ -100,13 +104,35 @@ public class LayoutParsingStorageService {
     }
 
 
+    @SneakyThrows
     @Observed(name = "LayoutParsingStorageService", contextualName = "store-document-data")
     public void storeDocumentData(LayoutParsingRequest layoutParsingRequest, DocumentData documentData) {
 
-        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.structureFileStorageId(), documentData.getDocumentStructure());
-        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.textBlockFileStorageId(), documentData.getDocumentTextData());
-        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.positionBlockFileStorageId(), documentData.getDocumentPositions());
-        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.pageFileStorageId(), documentData.getDocumentPages());
+        Runnable storeDocumentStructureRunnable = () -> storageService.storeProtoObject(TenantContext.getTenantId(),
+                                                                                       layoutParsingRequest.structureFileStorageId(),
+                                                                                       documentData.getDocumentStructure());
+
+        CompletableFuture<Void> storeDocumentStructureFuture = CompletableFuture.runAsync(storeDocumentStructureRunnable, taskExecutor);
+
+        Runnable storeDocumentTextDataRunnable = () -> storageService.storeProtoObject(TenantContext.getTenantId(),
+                                                                                      layoutParsingRequest.textBlockFileStorageId(),
+                                                                                      documentData.getDocumentTextData());
+
+        CompletableFuture<Void> storeDocumentTextDataFuture = CompletableFuture.runAsync(storeDocumentTextDataRunnable, taskExecutor);
+
+        Runnable storeDocumentPositionsRunnable = () -> storageService.storeProtoObject(TenantContext.getTenantId(),
+                                                                                       layoutParsingRequest.positionBlockFileStorageId(),
+                                                                                       documentData.getDocumentPositions());
+
+        CompletableFuture<Void> storeDocumentPositionsFuture = CompletableFuture.runAsync(storeDocumentPositionsRunnable, taskExecutor);
+
+        Runnable storeDocumentPagesRunnable = () -> storageService.storeProtoObject(TenantContext.getTenantId(),
+                                                                                   layoutParsingRequest.pageFileStorageId(),
+                                                                                   documentData.getDocumentPages());
+
+        CompletableFuture<Void> storeDocumentPagesFuture = CompletableFuture.runAsync(storeDocumentPagesRunnable, taskExecutor);
+
+        CompletableFuture.allOf(storeDocumentStructureFuture, storeDocumentTextDataFuture, storeDocumentPositionsFuture, storeDocumentPagesFuture).join();
     }
 
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/AbstractPageBlock.java

@@ -4,7 +4,7 @@ import java.util.HashSet;
 import java.util.Set;
 
 import com.fasterxml.jackson.annotation.JsonIgnore;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngineProto.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.processor.docstrum.model.BoundingBox;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageBlock;
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/DocumentTree.java

@@ -8,7 +8,7 @@ import java.util.LinkedList;
 import java.util.List;
 import java.util.stream.Stream;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.GenericSemanticNode;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/AbstractSemanticNode.java

@@ -6,7 +6,7 @@ import java.util.List;
 import java.util.Map;
 import java.util.Set;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngineProto.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.TextEntity;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Document.java

@@ -9,7 +9,7 @@ import java.util.Set;
 import java.util.stream.Collectors;
 import java.util.stream.Stream;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Footer.java

@@ -1,6 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Header.java

@@ -1,6 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Headline.java

@@ -1,6 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Image.java

@@ -6,7 +6,7 @@ import java.util.HashMap;
 import java.util.Map;
 import java.util.Set;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Paragraph.java

@@ -1,6 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Section.java

@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 
 import lombok.AccessLevel;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/SemanticNode.java

@@ -12,8 +12,8 @@ import java.util.Set;
 import java.util.stream.Collectors;
 import java.util.stream.Stream;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngineProto.LayoutEngine;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.EntityType;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/SuperSection.java

@@ -1,6 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 
 import lombok.AccessLevel;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/Table.java

@@ -12,8 +12,8 @@ import java.util.Set;
 import java.util.stream.IntStream;
 import java.util.stream.Stream;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngineProto.LayoutEngine;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.TextEntity;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/nodes/TableCell.java

@@ -4,7 +4,7 @@ import java.awt.geom.Rectangle2D;
 import java.util.HashMap;
 import java.util.Map;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlockCollector;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/textblock/AtomicTextBlock.java

@@ -1,5 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock;
 
+import static com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextDataProto.DocumentTextData;
 import static java.lang.String.format;
 
 import java.awt.geom.Rectangle2D;
@@ -13,8 +14,8 @@ import java.util.Map;
 import java.util.Set;
 import java.util.stream.Collectors;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.DocumentPositionData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.DocumentPositionData.Position;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
@@ -153,11 +154,9 @@ public class AtomicTextBlock implements TextBlock {
                 .page(page)
                 .textRange(new TextRange(documentTextData.getStart(), documentTextData.getEnd()))
                 .searchText(documentTextData.getSearchText())
-                .lineBreaks(Arrays.stream(documentTextData.getLineBreaks()).boxed()
-                                    .toList())
-                .stringIdxToPositionIdx(Arrays.stream(documentPositionData.getStringIdxToPositionIdx()).boxed()
-                                                .toList())
-                .positions(toRectangle2DList(documentPositionData.getPositions()))
+                .lineBreaks(documentTextData.getLineBreaksList())
+                .stringIdxToPositionIdx(documentPositionData.getStringIdxToPositionIdxList())
+                .positions(toRectangle2DList(documentPositionData.getPositionsList()))
                 .parent(parent)
                 .build();
     }
@@ -171,6 +170,14 @@ public class AtomicTextBlock implements TextBlock {
     }
 
 
+    private static List<Rectangle2D> toRectangle2DList(List<Position> positions) {
+
+        return positions.stream()
+                .map(pos -> (Rectangle2D) new Rectangle2D.Float(pos.getValue(0), pos.getValue(1), pos.getValue(2), pos.getValue(3)))
+                .toList();
+    }
+
+
     public CharSequence getLine(int lineNumber) {
 
         if (lineNumber >= numberOfLines() || lineNumber < 0) {

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/blockification/BlockificationPostprocessingService.java

@@ -11,7 +11,7 @@ import org.apache.commons.lang3.StringUtils;
 import org.apache.commons.text.similarity.LevenshteinDistance;
 import org.springframework.stereotype.Service;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngineProto.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationPage;
 import com.knecon.fforesight.service.layoutparser.processor.model.PageBlockType;
 import com.knecon.fforesight.service.layoutparser.processor.model.SectionIdentifier;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/mapper/DocumentDataMapper.java

@@ -1,6 +1,10 @@
 package com.knecon.fforesight.service.layoutparser.processor.services.mapper;
 
+import static com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextDataProto.AllDocumentTextData;
+import static com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextDataProto.DocumentTextData;
+
 import java.awt.geom.Rectangle2D;
+import java.util.ArrayList;
 import java.util.HashMap;
 import java.util.HashSet;
 import java.util.List;
@@ -9,11 +13,15 @@ import java.util.Set;
 import java.util.stream.Collectors;
 
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPage;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPageProto.AllDocumentPages;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPageProto.DocumentPage;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.AllDocumentPositionData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.DocumentPositionData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.DocumentPositionData.Position;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureProto.DocumentStructure;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureWrapper;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.EntryDataProto.EntryData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngineProto.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.DuplicatedParagraph;
@@ -32,44 +40,57 @@ public class DocumentDataMapper {
     public DocumentData toDocumentData(Document document) {
 
         List<DocumentTextData> documentTextData = document.streamTerminalTextBlocksInOrder()
-                .flatMap(textBlock -> textBlock.getAtomicTextBlocks().stream())
+                .flatMap(textBlock -> textBlock.getAtomicTextBlocks()
+                        .stream())
                 .distinct()
                 .map(DocumentDataMapper::toAtomicTextBlockData)
                 .toList();
 
+        AllDocumentTextData allDocumentTextData = AllDocumentTextData.newBuilder().addAllDocumentTextData(documentTextData).build();
+
         List<DocumentPositionData> atomicPositionBlockData = document.streamTerminalTextBlocksInOrder()
-                .flatMap(textBlock -> textBlock.getAtomicTextBlocks().stream())
+                .flatMap(textBlock -> textBlock.getAtomicTextBlocks()
+                        .stream())
                 .distinct()
                 .map(DocumentDataMapper::toAtomicPositionBlockData)
                 .toList();
 
-        Set<Long> nonEmptyTextBlocks = documentTextData.stream().mapToLong(DocumentTextData::getId).boxed().collect(Collectors.toSet());
+        AllDocumentPositionData allDocumentPositionData = AllDocumentPositionData.newBuilder().addAllDocumentPositionData(atomicPositionBlockData).build();
 
-        List<DocumentPage> documentPageData = document.getPages().stream().map(DocumentDataMapper::toPageData).toList();
-        DocumentStructure tableOfContentsData = toDocumentTreeData(document.getDocumentTree());
+        Set<Long> nonEmptyTextBlocks = documentTextData.stream()
+                .mapToLong(DocumentTextData::getId).boxed()
+                .collect(Collectors.toSet());
+
+        List<DocumentPage> documentPageData = document.getPages()
+                .stream()
+                .map(DocumentDataMapper::toPageData)
+                .toList();
+
+        AllDocumentPages allDocumentPages = AllDocumentPages.newBuilder().addAllDocumentPages(documentPageData).build();
+        DocumentStructureWrapper tableOfContentsData = toDocumentTreeData(document.getDocumentTree());
         return DocumentData.builder()
-                .documentTextData(documentTextData.toArray(new DocumentTextData[0]))
-                .documentPositions(atomicPositionBlockData.toArray(new DocumentPositionData[0]))
-                .documentPages(documentPageData.toArray(new DocumentPage[0]))
-                .documentStructure(tableOfContentsData)
+                .documentTextData(allDocumentTextData)
+                .documentPositions(allDocumentPositionData)
+                .documentPages(allDocumentPages)
+                .documentStructureWrapper(tableOfContentsData)
                 .build();
     }
 
 
-    private DocumentStructure toDocumentTreeData(DocumentTree documentTree) {
+    private DocumentStructureWrapper toDocumentTreeData(DocumentTree documentTree) {
 
-        return new DocumentStructure(toEntryData(documentTree.getRoot()));
+        return new DocumentStructureWrapper(DocumentStructure.newBuilder().setRoot(toEntryData(documentTree.getRoot())).build());
     }
 
 
-    private DocumentStructure.EntryData toEntryData(DocumentTree.Entry entry) {
+    private EntryData toEntryData(DocumentTree.Entry entry) {
 
-        Long[] atomicTextBlocks;
+        List<Long> atomicTextBlocks;
 
         if (entry.getNode().isLeaf()) {
             atomicTextBlocks = toAtomicTextBlockIds(entry.getNode().getLeafTextBlock());
         } else {
-            atomicTextBlocks = new Long[]{};
+            atomicTextBlocks = new ArrayList<>();
         }
 
         Map<String, String> properties = switch (entry.getType()) {
@@ -81,77 +102,81 @@ public class DocumentDataMapper {
             default -> new HashMap<>();
         };
 
-        DocumentStructure.EntryData.EntryDataBuilder documentBuilder = DocumentStructure.EntryData.builder()
-                .treeId(toPrimitiveIntArray(entry.getTreeId()))
-                .children(entry.getChildren().stream().map(DocumentDataMapper::toEntryData).toList())
-                .type(entry.getType())
-                .atomicBlockIds(atomicTextBlocks)
-                .pageNumbers(entry.getNode().getPages().stream().map(Page::getNumber).map(Integer::longValue).toArray(Long[]::new))
-                .properties(properties);
+        var documentBuilder = EntryData.newBuilder()
+                .addAllTreeId(entry.getTreeId())
+                .addAllChildren(entry.getChildren()
+                                        .stream()
+                                        .map(DocumentDataMapper::toEntryData)
+                                        .toList())
+                .setType(entry.getType())
+                .addAllAtomicBlockIds(atomicTextBlocks)
+                .addAllPageNumbers(entry.getNode().getPages()
+                                           .stream()
+                                           .map(Page::getNumber)
+                                           .map(Integer::longValue)
+                                           .toList())
+                .putAllProperties(properties);
         if (entry.getNode() != null) {
-            documentBuilder.engines(entry.getNode().getEngines());
+            documentBuilder.addAllEngines(entry.getNode().getEngines());
         } else {
-            documentBuilder.engines(new HashSet<>(Set.of(LayoutEngine.ALGORITHM)));
+            documentBuilder.addAllEngines(new HashSet<>(Set.of(LayoutEngine.ALGORITHM)));
         }
         return documentBuilder.build();
     }
 
 
-    private Long[] toAtomicTextBlockIds(TextBlock textBlock) {
+    private List<Long> toAtomicTextBlockIds(TextBlock textBlock) {
 
-        return textBlock.getAtomicTextBlocks().stream().map(AtomicTextBlock::getId).toArray(Long[]::new);
+        return textBlock.getAtomicTextBlocks()
+                .stream()
+                .map(AtomicTextBlock::getId)
+                .toList();
     }
 
 
     private DocumentPage toPageData(Page p) {
 
-        return DocumentPage.builder().rotation(p.getRotation()).height(p.getHeight()).width(p.getWidth()).number(p.getNumber()).build();
+        return DocumentPage.newBuilder().setRotation(p.getRotation()).setHeight(p.getHeight()).setWidth(p.getWidth()).setNumber(p.getNumber()).build();
     }
 
 
     private DocumentTextData toAtomicTextBlockData(AtomicTextBlock atomicTextBlock) {
 
-        return DocumentTextData.builder()
-                .id(atomicTextBlock.getId())
-                .page(atomicTextBlock.getPage().getNumber().longValue())
-                .searchText(atomicTextBlock.getSearchText())
-                .numberOnPage(atomicTextBlock.getNumberOnPage())
-                .start(atomicTextBlock.getTextRange().start())
-                .end(atomicTextBlock.getTextRange().end())
-                .lineBreaks(toPrimitiveIntArray(atomicTextBlock.getLineBreaks()))
+        return DocumentTextData.newBuilder()
+                .setId(atomicTextBlock.getId())
+                .setPage(atomicTextBlock.getPage().getNumber().longValue())
+                .setSearchText(atomicTextBlock.getSearchText())
+                .setNumberOnPage(atomicTextBlock.getNumberOnPage())
+                .setStart(atomicTextBlock.getTextRange().start())
+                .setEnd(atomicTextBlock.getTextRange().end())
+                .addAllLineBreaks(atomicTextBlock.getLineBreaks())
                 .build();
     }
 
 
     private DocumentPositionData toAtomicPositionBlockData(AtomicTextBlock atomicTextBlock) {
 
-        return DocumentPositionData.builder()
-                .id(atomicTextBlock.getId())
-                .positions(toPrimitiveFloatMatrix(atomicTextBlock.getPositions()))
-                .stringIdxToPositionIdx(toPrimitiveIntArray(atomicTextBlock.getStringIdxToPositionIdx()))
+        return DocumentPositionData.newBuilder()
+                .setId(atomicTextBlock.getId())
+                .addAllPositions(toPositions(atomicTextBlock.getPositions()))
+                .addAllStringIdxToPositionIdx(atomicTextBlock.getStringIdxToPositionIdx())
                 .build();
     }
 
 
-    private static float[][] toPrimitiveFloatMatrix(List<Rectangle2D> positions) {
+    private static List<Position> toPositions(List<Rectangle2D> rects) {
 
-        float[][] positionMatrix = new float[positions.size()][];
-        for (int i = 0; i < positions.size(); i++) {
-            positionMatrix[i] = toArray(positions.get(i));
+        List<Position> positions = new ArrayList<>();
+        for (Rectangle2D rect : rects) {
+            positions.add(toPosition(rect));
         }
-        return positionMatrix;
+        return positions;
     }
 
 
-    private static float[] toArray(Rectangle2D positions) {
+    private static Position toPosition(Rectangle2D rect) {
 
-        return new float[]{(float) positions.getMinX(), (float) positions.getMinY(), (float) positions.getWidth(), (float) positions.getHeight()};
-    }
-
-
-    private int[] toPrimitiveIntArray(List<Integer> list) {
-
-        return list.stream().mapToInt(Integer::intValue).toArray();
+        return Position.newBuilder().addValue((float) rect.getMinX()).addValue((float) rect.getMinY()).addValue((float) rect.getWidth()).addValue((float) rect.getHeight()).build();
     }
 
 }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/mapper/DocumentGraphMapper.java

@@ -8,10 +8,10 @@ import java.util.Map;
 import java.util.NoSuchElementException;
 
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPage;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPageProto.DocumentPage;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.AllDocumentPositionData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextDataProto.AllDocumentTextData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.EntryDataProto.EntryData;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.DuplicatedParagraph;
@@ -41,27 +41,29 @@ public class DocumentGraphMapper {
         DocumentTree documentTree = new DocumentTree(document);
         Context context = new Context(documentData, documentTree);
 
-        context.pages.addAll(Arrays.stream(documentData.getDocumentPages())
+        context.pages.addAll(documentData.getDocumentPages().getDocumentPagesList()
+                                     .stream()
                                      .map(DocumentGraphMapper::buildPage)
                                      .toList());
 
-        context.documentTree.getRoot().getChildren().addAll(buildEntries(documentData.getDocumentStructure().getRoot().getChildren(), context));
+        context.documentTree.getRoot().getChildren().addAll(buildEntries(documentData.getDocumentStructure().getRoot().getChildrenList(), context));
 
         document.setDocumentTree(context.documentTree);
         document.setPages(new HashSet<>(context.pages));
-        document.setNumberOfPages(documentData.getDocumentPages().length);
+        document.setNumberOfPages(documentData.getDocumentPages().getDocumentPagesCount());
 
         document.setTextBlock(document.getTextBlock());
         return document;
     }
 
 
-    private List<DocumentTree.Entry> buildEntries(List<DocumentStructure.EntryData> entries, Context context) {
+    private List<DocumentTree.Entry> buildEntries(List<EntryData> entries, Context context) {
 
         List<DocumentTree.Entry> newEntries = new LinkedList<>();
-        for (DocumentStructure.EntryData entryData : entries) {
+        for (EntryData entryData : entries) {
 
-            List<Page> pages = Arrays.stream(entryData.getPageNumbers())
+            List<Page> pages = entryData.getPageNumbersList()
+                    .stream()
                     .map(pageNumber -> getPage(pageNumber, context))
                     .toList();
 
@@ -74,12 +76,12 @@ public class DocumentGraphMapper {
                 case FOOTER -> buildFooter(context);
                 case TABLE -> buildTable(context, entryData.getProperties());
                 case TABLE_CELL -> buildTableCell(context, entryData.getProperties());
-                case IMAGE -> buildImage(context, entryData.getProperties(), entryData.getPageNumbers());
+                case IMAGE -> buildImage(context, entryData.getProperties(), entryData.getPageNumbersList());
                 default -> throw new UnsupportedOperationException("Not yet implemented for type " + entryData.getType());
             };
 
-            if (entryData.getAtomicBlockIds().length > 0) {
-                TextBlock textBlock = toTextBlock(entryData.getAtomicBlockIds(), context, node);
+            if (entryData.getAtomicBlockIdsCount() > 0) {
+                TextBlock textBlock = toTextBlock(entryData.getAtomicBlockIdsList(), context, node);
                 node.setLeafTextBlock(textBlock);
                 switch (entryData.getType()) {
                     case HEADER -> pages.forEach(page -> page.setHeader((Header) node));
@@ -89,11 +91,10 @@ public class DocumentGraphMapper {
                             .forEach(atb -> atb.getPage().getTextBlocksOnPage().add(atb));
                 }
             }
-            List<Integer> treeId = Arrays.stream(entryData.getTreeId()).boxed()
-                    .toList();
+            List<Integer> treeId = entryData.getTreeIdList();
             node.setTreeId(treeId);
 
-            newEntries.add(DocumentTree.Entry.builder().treeId(treeId).children(buildEntries(entryData.getChildren(), context)).node(node).build());
+            newEntries.add(DocumentTree.Entry.builder().treeId(treeId).children(buildEntries(entryData.getChildrenList(), context)).node(node).build());
         }
         return newEntries;
     }
@@ -105,10 +106,10 @@ public class DocumentGraphMapper {
     }
 
 
-    private Image buildImage(Context context, Map<String, String> properties, Long[] pageNumbers) {
+    private Image buildImage(Context context, Map<String, String> properties, List<Long> pageNumbers) {
 
-        assert pageNumbers.length == 1;
-        Page page = getPage(pageNumbers[0], context);
+        assert pageNumbers.size() == 1;
+        Page page = getPage(pageNumbers.get(0), context);
         var builder = Image.builder();
         PropertiesMapper.parseImageProperties(properties, builder);
         return builder.documentTree(context.documentTree).page(page).build();
@@ -161,7 +162,7 @@ public class DocumentGraphMapper {
 
             DuplicatedParagraph duplicatedParagraph = DuplicatedParagraph.builder().documentTree(context.documentTree).build();
 
-            Long[] unsortedTextblockIds = PropertiesMapper.getUnsortedTextblockIds(properties);
+            var unsortedTextblockIds = PropertiesMapper.getUnsortedTextblockIds(properties);
             duplicatedParagraph.setUnsortedLeafTextBlock(toTextBlock(unsortedTextblockIds, context, duplicatedParagraph));
             return duplicatedParagraph;
 
@@ -171,9 +172,9 @@ public class DocumentGraphMapper {
     }
 
 
-    private TextBlock toTextBlock(Long[] atomicTextBlockIds, Context context, SemanticNode parent) {
+    private TextBlock toTextBlock(List<Long> atomicTextBlockIds, Context context, SemanticNode parent) {
 
-        return Arrays.stream(atomicTextBlockIds)
+        return atomicTextBlockIds.stream()
                 .map(atomicTextBlockId -> getAtomicTextBlock(context, parent, atomicTextBlockId))
                 .collect(new TextBlockCollector());
     }
@@ -181,10 +182,13 @@ public class DocumentGraphMapper {
 
     private AtomicTextBlock getAtomicTextBlock(Context context, SemanticNode parent, Long atomicTextBlockId) {
 
-        return AtomicTextBlock.fromAtomicTextBlockData(context.documentTextDataBlockData.get(Math.toIntExact(atomicTextBlockId)),
-                                                       context.atomicPositionBlockData.get(Math.toIntExact(atomicTextBlockId)),
+        return AtomicTextBlock.fromAtomicTextBlockData(context.documentTextDataBlockData.getDocumentTextDataList()
+                                                               .get(Math.toIntExact(atomicTextBlockId)),
+                                                       context.atomicPositionBlockData.getDocumentPositionDataList()
+                                                               .get(Math.toIntExact(atomicTextBlockId)),
                                                        parent,
-                                                       getPage(context.documentTextDataBlockData.get(Math.toIntExact(atomicTextBlockId)).getPage(), context));
+                                                       getPage(context.documentTextDataBlockData.getDocumentTextDataList()
+                                                                       .get(Math.toIntExact(atomicTextBlockId)).getPage(), context));
     }
 
 
@@ -207,18 +211,16 @@ public class DocumentGraphMapper {
 
         private final DocumentTree documentTree;
         private final List<Page> pages;
-        private final List<DocumentTextData> documentTextDataBlockData;
-        private final List<DocumentPositionData> atomicPositionBlockData;
+        private final AllDocumentTextData documentTextDataBlockData;
+        private final AllDocumentPositionData atomicPositionBlockData;
 
 
         Context(DocumentData documentData, DocumentTree documentTree) {
 
             this.documentTree = documentTree;
             this.pages = new LinkedList<>();
-            this.documentTextDataBlockData = Arrays.stream(documentData.getDocumentTextData())
-                    .toList();
-            this.atomicPositionBlockData = Arrays.stream(documentData.getDocumentPositions())
-                    .toList();
+            this.documentTextDataBlockData = documentData.getDocumentTextData();
+            this.atomicPositionBlockData = documentData.getDocumentPositions();
 
         }
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/mapper/PropertiesMapper.java

@@ -3,10 +3,11 @@ package com.knecon.fforesight.service.layoutparser.processor.services.mapper;
 import java.awt.geom.Rectangle2D;
 import java.util.Arrays;
 import java.util.HashMap;
+import java.util.List;
 import java.util.Locale;
 import java.util.Map;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureWrapper;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.DuplicatedParagraph;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.ImageType;
@@ -20,11 +21,11 @@ public class PropertiesMapper {
     public static Map<String, String> buildImageProperties(Image image) {
 
         Map<String, String> properties = new HashMap<>();
-        properties.put(DocumentStructure.ImageProperties.IMAGE_TYPE, image.getImageType().toString());
-        properties.put(DocumentStructure.ImageProperties.TRANSPARENT, String.valueOf(image.isTransparent()));
-        properties.put(DocumentStructure.ImageProperties.POSITION, toString(image.getPosition()));
-        properties.put(DocumentStructure.ImageProperties.ID, image.getId());
-        properties.put(DocumentStructure.ImageProperties.REPRESENTATION_HASH, image.getRepresentationHash());
+        properties.put(DocumentStructureWrapper.ImageProperties.IMAGE_TYPE, image.getImageType().toString());
+        properties.put(DocumentStructureWrapper.ImageProperties.TRANSPARENT, String.valueOf(image.isTransparent()));
+        properties.put(DocumentStructureWrapper.ImageProperties.POSITION, toString(image.getPosition()));
+        properties.put(DocumentStructureWrapper.ImageProperties.ID, image.getId());
+        properties.put(DocumentStructureWrapper.ImageProperties.REPRESENTATION_HASH, image.getRepresentationHash());
         return properties;
     }
 
@@ -32,15 +33,19 @@ public class PropertiesMapper {
     public static Map<String, String> buildTableCellProperties(TableCell tableCell) {
 
         Map<String, String> properties = new HashMap<>();
-        properties.put(DocumentStructure.TableCellProperties.ROW, String.valueOf(tableCell.getRow()));
-        properties.put(DocumentStructure.TableCellProperties.COL, String.valueOf(tableCell.getCol()));
-        properties.put(DocumentStructure.TableCellProperties.HEADER, String.valueOf(tableCell.isHeader()));
+        properties.put(DocumentStructureWrapper.TableCellProperties.ROW, String.valueOf(tableCell.getRow()));
+        properties.put(DocumentStructureWrapper.TableCellProperties.COL, String.valueOf(tableCell.getCol()));
+        properties.put(DocumentStructureWrapper.TableCellProperties.HEADER, String.valueOf(tableCell.isHeader()));
 
         if (tableCell.getPages().size() > 1 || tableCell.getBBox().keySet().size() > 1) {
             throw new IllegalArgumentException("TableCell can only occur on a single page!");
         }
-        String bBoxString = toString(tableCell.getBBox().get(tableCell.getPages().stream().findFirst().get()));
-        properties.put(DocumentStructure.TableCellProperties.B_BOX, bBoxString);
+        String bBoxString = toString(tableCell.getBBox()
+                                             .get(tableCell.getPages()
+                                                          .stream()
+                                                          .findFirst()
+                                                          .get()));
+        properties.put(DocumentStructureWrapper.TableCellProperties.B_BOX, bBoxString);
 
         return properties;
     }
@@ -49,59 +54,62 @@ public class PropertiesMapper {
     public static Map<String, String> buildTableProperties(Table table) {
 
         Map<String, String> properties = new HashMap<>();
-        properties.put(DocumentStructure.TableProperties.NUMBER_OF_ROWS, String.valueOf(table.getNumberOfRows()));
-        properties.put(DocumentStructure.TableProperties.NUMBER_OF_COLS, String.valueOf(table.getNumberOfCols()));
+        properties.put(DocumentStructureWrapper.TableProperties.NUMBER_OF_ROWS, String.valueOf(table.getNumberOfRows()));
+        properties.put(DocumentStructureWrapper.TableProperties.NUMBER_OF_COLS, String.valueOf(table.getNumberOfCols()));
         return properties;
     }
 
 
     public static void parseImageProperties(Map<String, String> properties, Image.ImageBuilder builder) {
 
-        builder.imageType(parseImageType(properties.get(DocumentStructure.ImageProperties.IMAGE_TYPE)));
-        builder.transparent(Boolean.parseBoolean(properties.get(DocumentStructure.ImageProperties.TRANSPARENT)));
-        builder.position(DocumentStructure.parseRectangle2D(properties.get(DocumentStructure.ImageProperties.POSITION)));
+        builder.imageType(parseImageType(properties.get(DocumentStructureWrapper.ImageProperties.IMAGE_TYPE)));
+        builder.transparent(Boolean.parseBoolean(properties.get(DocumentStructureWrapper.ImageProperties.TRANSPARENT)));
+        builder.position(DocumentStructureWrapper.parseRectangle2D(properties.get(DocumentStructureWrapper.ImageProperties.POSITION)));
     }
 
 
     public static void parseTableCellProperties(Map<String, String> properties, TableCell.TableCellBuilder builder) {
 
-        builder.row(Integer.parseInt(properties.get(DocumentStructure.TableCellProperties.ROW)));
-        builder.col(Integer.parseInt(properties.get(DocumentStructure.TableCellProperties.COL)));
-        builder.header(Boolean.parseBoolean(properties.get(DocumentStructure.TableCellProperties.HEADER)));
-        builder.bBox(DocumentStructure.parseRectangle2D(properties.get(DocumentStructure.TableCellProperties.B_BOX)));
+        builder.row(Integer.parseInt(properties.get(DocumentStructureWrapper.TableCellProperties.ROW)));
+        builder.col(Integer.parseInt(properties.get(DocumentStructureWrapper.TableCellProperties.COL)));
+        builder.header(Boolean.parseBoolean(properties.get(DocumentStructureWrapper.TableCellProperties.HEADER)));
+        builder.bBox(DocumentStructureWrapper.parseRectangle2D(properties.get(DocumentStructureWrapper.TableCellProperties.B_BOX)));
     }
 
 
     public static void parseTableProperties(Map<String, String> properties, Table.TableBuilder builder) {
 
-        builder.numberOfRows(Integer.parseInt(properties.get(DocumentStructure.TableProperties.NUMBER_OF_ROWS)));
-        builder.numberOfCols(Integer.parseInt(properties.get(DocumentStructure.TableProperties.NUMBER_OF_COLS)));
+        builder.numberOfRows(Integer.parseInt(properties.get(DocumentStructureWrapper.TableProperties.NUMBER_OF_ROWS)));
+        builder.numberOfCols(Integer.parseInt(properties.get(DocumentStructureWrapper.TableProperties.NUMBER_OF_COLS)));
     }
 
 
     public static Map<String, String> buildDuplicateParagraphProperties(DuplicatedParagraph duplicatedParagraph) {
 
         Map<String, String> properties = new HashMap<>();
-        properties.put(DocumentStructure.DuplicateParagraphProperties.UNSORTED_TEXTBLOCK_ID, Arrays.toString(toAtomicTextBlockIds(duplicatedParagraph.getUnsortedLeafTextBlock())));
+        properties.put(DocumentStructureWrapper.DuplicateParagraphProperties.UNSORTED_TEXTBLOCK_ID,
+                       Arrays.toString(toAtomicTextBlockIds(duplicatedParagraph.getUnsortedLeafTextBlock())));
         return properties;
     }
 
 
     public static boolean isDuplicateParagraph(Map<String, String> properties) {
 
-        return properties.containsKey(DocumentStructure.DuplicateParagraphProperties.UNSORTED_TEXTBLOCK_ID);
+        return properties.containsKey(DocumentStructureWrapper.DuplicateParagraphProperties.UNSORTED_TEXTBLOCK_ID);
     }
 
 
-    public static Long[] getUnsortedTextblockIds(Map<String, String> properties) {
+    public static List<Long> getUnsortedTextblockIds(Map<String, String> properties) {
 
-        return toLongArray(properties.get(DocumentStructure.DuplicateParagraphProperties.UNSORTED_TEXTBLOCK_ID));
+        return toLongList(properties.get(DocumentStructureWrapper.DuplicateParagraphProperties.UNSORTED_TEXTBLOCK_ID));
     }
 
 
-    public static Long[] toLongArray(String ids) {
+    public static List<Long> toLongList(String ids) {
 
-        return Arrays.stream(ids.substring(1, ids.length() - 1).trim().split(",")).map(Long::valueOf).toArray(Long[]::new);
+        return Arrays.stream(ids.substring(1, ids.length() - 1).trim().split(","))
+                .map(Long::valueOf)
+                .toList();
     }
 
 
@@ -120,20 +128,23 @@ public class PropertiesMapper {
     public static String toString(Rectangle2D rectangle2D) {
 
         return String.format(Locale.US,
-                "%f%s%f%s%f%s%f",
-                rectangle2D.getX(),
-                DocumentStructure.RECTANGLE_DELIMITER,
-                rectangle2D.getY(),
-                DocumentStructure.RECTANGLE_DELIMITER,
-                rectangle2D.getWidth(),
-                DocumentStructure.RECTANGLE_DELIMITER,
-                rectangle2D.getHeight());
+                             "%f%s%f%s%f%s%f",
+                             rectangle2D.getX(),
+                             DocumentStructureWrapper.RECTANGLE_DELIMITER,
+                             rectangle2D.getY(),
+                             DocumentStructureWrapper.RECTANGLE_DELIMITER,
+                             rectangle2D.getWidth(),
+                             DocumentStructureWrapper.RECTANGLE_DELIMITER,
+                             rectangle2D.getHeight());
     }
 
 
     private static Long[] toAtomicTextBlockIds(TextBlock textBlock) {
 
-        return textBlock.getAtomicTextBlocks().stream().map(AtomicTextBlock::getId).toArray(Long[]::new);
+        return textBlock.getAtomicTextBlocks()
+                .stream()
+                .map(AtomicTextBlock::getId)
+                .toArray(Long[]::new);
     }
 
 }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/mapper/TaasDocumentDataMapper.java

@@ -6,7 +6,7 @@ import java.util.Locale;
 import java.util.concurrent.atomic.AtomicInteger;
 import java.util.stream.IntStream;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.taas.ParagraphData;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.taas.Range;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.taas.ResearchDocumentData;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/utils/PdfVisualisationUtility.java

@@ -14,7 +14,7 @@ import org.apache.pdfbox.pdmodel.PDPageContentStream;
 import org.apache.pdfbox.pdmodel.font.PDType1Font;
 import org.apache.pdfbox.pdmodel.font.Standard14Fonts;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
@@ -117,6 +117,7 @@ public class PdfVisualisationUtility {
             case TABLE -> Color.ORANGE;
             case TABLE_CELL -> Color.GRAY;
             case IMAGE -> Color.MAGENTA;
+            case UNRECOGNIZED -> Color.PINK;
         }).build();
     }
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/utils/ProtobufUtil.java

@@ -0,0 +1,42 @@
+package com.knecon.fforesight.service.layoutparser.processor.utils;
+
+import java.io.BufferedOutputStream;
+import java.io.File;
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.util.zip.GZIPOutputStream;
+
+import com.google.protobuf.Message;
+import com.google.protobuf.MessageOrBuilder;
+import com.google.protobuf.Struct;
+import com.google.protobuf.util.JsonFormat;
+
+import lombok.SneakyThrows;
+import lombok.experimental.UtilityClass;
+
+@UtilityClass
+public class ProtobufUtil {
+
+    public static String toJson(MessageOrBuilder messageOrBuilder) throws IOException {
+        return JsonFormat.printer().print(messageOrBuilder);
+    }
+
+    @SuppressWarnings("unchecked")
+    public static Message fromJson(String json) throws IOException {
+        Message.Builder structBuilder = Struct.newBuilder();
+        JsonFormat.parser().ignoringUnknownFields().merge(json, structBuilder);
+        return structBuilder.build();
+    }
+
+
+    @SneakyThrows
+    public <T extends Message> File serializeToTempFile(T any) {
+        var tempFile = File.createTempFile("storage-protobuf", ".data");
+
+        try (var fos = new GZIPOutputStream(new BufferedOutputStream(new FileOutputStream(tempFile)))) {
+            any.writeTo(fos);
+            return tempFile;
+        }
+    }
+
+}

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/visualization/LayoutGrid.java

@@ -16,7 +16,7 @@ import java.util.stream.Collectors;
 import java.util.stream.Stream;
 
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.DuplicatedParagraph;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Headline;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
@@ -125,7 +125,7 @@ public class LayoutGrid extends LayoutGridLayerConfig {
 
         Map<Page, Rectangle2D> bBoxMap = section.getBBox();
 
-        List<SemanticNode> subSections = section.streamAllSubNodesOfType(NodeType.SECTION)
+        List<SemanticNode> subSections = section.streamAllSubNodesOfType(NodeTypeProto.NodeType.SECTION)
                 .toList();
         Integer maxChildDepth = subSections.stream()
                 .map(node -> node.getTreeId().size())

layoutparser-service/layoutparser-service-server/build.gradle.kts

@@ -29,7 +29,7 @@ dependencies {
     implementation(project(":layoutparser-service-processor"))
     implementation(project(":layoutparser-service-internal-api"))
 
-    implementation("com.iqser.red.commons:storage-commons:2.45.0")
+    implementation("com.iqser.red.commons:storage-commons:2.50.0")
     implementation("com.knecon.fforesight:tenant-commons:0.30.0")
     implementation("com.knecon.fforesight:tracing-commons:0.5.0")
     implementation("com.knecon.fforesight:lifecycle-commons:0.6.0")
@@ -38,11 +38,14 @@ dependencies {
     implementation("com.amazonaws:aws-java-sdk-s3:1.12.536")
     implementation("org.springframework.cloud:spring-cloud-starter-openfeign:4.0.4")
     implementation("net.logstash.logback:logstash-logback-encoder:7.4")
+    implementation("ch.qos.logback:logback-classic")
+
     implementation("com.pdftron:PDFNet:10.11.0")
 
     // for integration testing only
     testImplementation(project(":viewer-doc-processor"))
     testImplementation(project(":layoutparser-service-internal-api"))
+    testImplementation("com.google.protobuf:protobuf-java-util:4.27.1")
 
     testImplementation("org.springframework.boot:spring-boot-starter-amqp:${springBootStarterVersion}")
     testImplementation("org.springframework.boot:spring-boot-starter-test:${springBootStarterVersion}")

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/graph/DocumentGraphJsonWritingTest.java

@@ -70,7 +70,7 @@ public class DocumentGraphJsonWritingTest extends BuildDocumentTest {
         var stem = Path.of("/tmp/DocumentGraphJsonWritingTest");
         stem.toFile().mkdirs();
         var tmpFilePath = stem.resolve(filename.getFileName());
-        mapper.writeValue(new FileOutputStream(new File(tmpFilePath + "_structure" + ".json")), documentData.getDocumentStructure());
+        mapper.writeValue(new FileOutputStream(new File(tmpFilePath + "_structure" + ".json")), documentData.getDocumentStructureWrapper());
         mapper.writeValue(new FileOutputStream(new File(tmpFilePath + "_text" + ".json")), documentData.getDocumentTextData());
         mapper.writeValue(new FileOutputStream(new File(tmpFilePath + "_positions" + ".json")), documentData.getDocumentPositions());
         mapper.writeValue(new FileOutputStream(new File(tmpFilePath + "_pages" + ".json")), documentData.getDocumentPages());

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/graph/DocumentGraphMappingTest.java

@@ -6,11 +6,13 @@ import static org.junit.jupiter.api.Assertions.assertTrue;
 import org.junit.jupiter.api.Test;
 
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPage;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPageProto.AllDocumentPages;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionDataProto.AllDocumentPositionData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureProto.DocumentStructure;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureWrapper;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextDataProto.AllDocumentTextData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.EntryDataProto.EntryData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
 import com.knecon.fforesight.service.layoutparser.processor.services.mapper.DocumentDataMapper;
@@ -32,21 +34,21 @@ public class DocumentGraphMappingTest extends BuildDocumentTest {
         Document document = buildGraph(filename);
         DocumentData documentData = DocumentDataMapper.toDocumentData(document);
 
-        storageService.storeJSONObject(TenantContext.getTenantId(), filename + "_PAGES" + ".json", documentData.getDocumentPages());
-        storageService.storeJSONObject(TenantContext.getTenantId(), filename + "_TEXT" + ".json", documentData.getDocumentTextData());
-        storageService.storeJSONObject(TenantContext.getTenantId(), filename + "_POSITIONS" + ".json", documentData.getDocumentPositions());
-        storageService.storeJSONObject(TenantContext.getTenantId(), filename + "_STRUCTURE" + ".json", documentData.getDocumentStructure());
+        storageService.storeProtoObject(TenantContext.getTenantId(), filename + "_PAGES" + ".proto", documentData.getDocumentPages());
+        storageService.storeProtoObject(TenantContext.getTenantId(), filename + "_TEXT" + ".proto", documentData.getDocumentTextData());
+        storageService.storeProtoObject(TenantContext.getTenantId(), filename + "_POSITIONS" + ".proto", documentData.getDocumentPositions());
+        storageService.storeProtoObject(TenantContext.getTenantId(), filename + "_STRUCTURE" + ".proto", documentData.getDocumentStructure());
 
-        DocumentPage[] pageData = storageService.readJSONObject(TenantContext.getTenantId(), filename + "_PAGES" + ".json", DocumentPage[].class);
-        DocumentTextData[] atomicTextBlockData = storageService.readJSONObject(TenantContext.getTenantId(), filename + "_TEXT" + ".json", DocumentTextData[].class);
-        DocumentPositionData[] atomicPositionBlockData = storageService.readJSONObject(TenantContext.getTenantId(),
-                filename + "_POSITIONS" + ".json",
-                DocumentPositionData[].class);
-        DocumentStructure documentTreeData = storageService.readJSONObject(TenantContext.getTenantId(), filename + "_STRUCTURE" + ".json", DocumentStructure.class);
+        AllDocumentPages pageData = storageService.readProtoObject(TenantContext.getTenantId(), filename + "_PAGES" + ".proto", AllDocumentPages.parser());
+        AllDocumentTextData atomicTextBlockData = storageService.readProtoObject(TenantContext.getTenantId(), filename + "_TEXT" + ".proto", AllDocumentTextData.parser());
+        AllDocumentPositionData atomicPositionBlockData = storageService.readProtoObject(TenantContext.getTenantId(),
+                                                                                         filename + "_POSITIONS" + ".proto",
+                                                                                         AllDocumentPositionData.parser());
+        DocumentStructure documentTreeData = storageService.readProtoObject(TenantContext.getTenantId(), filename + "_STRUCTURE" + ".proto", DocumentStructure.parser());
 
         DocumentData documentData2 = DocumentData.builder()
                 .documentPages(pageData)
-                .documentStructure(documentTreeData)
+                .documentStructureWrapper(new DocumentStructureWrapper(documentTreeData))
                 .documentTextData(atomicTextBlockData)
                 .documentPositions(atomicPositionBlockData)
                 .build();
@@ -73,10 +75,9 @@ public class DocumentGraphMappingTest extends BuildDocumentTest {
 
     private static boolean allTablesHavePositiveNumberOfRowsAndColumns(DocumentData documentData) {
 
-        return documentData.getDocumentStructure()
-                .streamAllEntries()
+        return documentData.getDocumentStructureWrapper().streamAllEntries()
                 .filter(entryData -> entryData.getType().equals(NodeType.TABLE))
-                .map(DocumentStructure.EntryData::getProperties)
+                .map(EntryData::getPropertiesMap)
                 .map(properties -> {
                     var builder = Table.builder();
                     PropertiesMapper.parseTableProperties(properties, builder);

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/services/RulingCleaningServiceTest.java

@@ -17,8 +17,9 @@ import org.junit.jupiter.api.Test;
 import org.springframework.core.io.ClassPathResource;
 
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentData;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructureWrapper;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.EntryDataProto.EntryData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.model.PageContents;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
@@ -119,7 +120,7 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
                         Map.of("file",filename.toFile().toString())));
         DocumentData documentDataBefore = DocumentDataMapper.toDocumentData(documentGraphBefore);
         DocumentData documentDataAfter = DocumentDataMapper.toDocumentData(documentGraphAfter);
-        if (!compareStructures(documentDataBefore.getDocumentStructure(), documentDataAfter.getDocumentStructure())) {
+        if (!compareStructures(documentDataBefore.getDocumentStructureWrapper(), documentDataAfter.getDocumentStructureWrapper())) {
             String tmpFileNameBefore = "/tmp/before." + filename.getFileName().toString();
             try (PDDocument pdDocument = Loader.loadPDF(filename.toFile())) {
                 PdfDraw.drawDocumentGraph(pdDocument, documentGraphBefore);
@@ -136,11 +137,11 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
 
 
     @SneakyThrows
-    private boolean compareStructures(DocumentStructure structure1, DocumentStructure structure2) {
+    private boolean compareStructures(DocumentStructureWrapper structure1, DocumentStructureWrapper structure2) {
 
         List<Table> listStructure1 = structure1.streamAllEntries()
                 .filter(entryData -> entryData.getType().equals(NodeType.TABLE))
-                .map(DocumentStructure.EntryData::getProperties)
+                .map(EntryData::getPropertiesMap)
                 .map(properties -> {
                     var builder = Table.builder();
                     PropertiesMapper.parseTableProperties(properties, builder);
@@ -150,7 +151,7 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
 
         List<Table> listStructure2 = structure2.streamAllEntries()
                 .filter(entryData -> entryData.getType().equals(NodeType.TABLE))
-                .map(DocumentStructure.EntryData::getProperties)
+                .map(EntryData::getPropertiesMap)
                 .map(properties -> {
                     var builder = Table.builder();
                     PropertiesMapper.parseTableProperties(properties, builder);

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/utils/visualizations/PdfDraw.java

@@ -17,7 +17,7 @@ import org.apache.pdfbox.pdmodel.font.Standard14Fonts;
 import org.apache.pdfbox.util.Matrix;
 import org.springframework.core.io.ClassPathResource;
 
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeTypeProto.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
@@ -233,6 +233,7 @@ public class PdfDraw {
             case TABLE -> Color.ORANGE;
             case TABLE_CELL -> Color.GRAY;
             case IMAGE -> Color.MAGENTA;
+            case UNRECOGNIZED -> Color.PINK;
         }).build();
     }