docxmlater 0.2.0 → 0.3.0

package/dist/core/Document.js

@@ -41,7 +41,6 @@ const Table_1 = require("../elements/Table");
 const Section_1 = require("../elements/Section");
 const ImageManager_1 = require("../elements/ImageManager");
 const HeaderFooterManager_1 = require("../elements/HeaderFooterManager");
-const Hyperlink_1 = require("../elements/Hyperlink");
 const TableOfContents_1 = require("../elements/TableOfContents");
 const TableOfContentsElement_1 = require("../elements/TableOfContentsElement");
 const BookmarkManager_1 = require("../elements/BookmarkManager");
@@ -49,17 +48,19 @@ const Revision_1 = require("../elements/Revision");
 const RevisionManager_1 = require("../elements/RevisionManager");
 const CommentManager_1 = require("../elements/CommentManager");
 const Run_1 = require("../elements/Run");
-const XMLBuilder_1 = require("../xml/XMLBuilder");
-const XMLParser_1 = require("../xml/XMLParser");
 const StylesManager_1 = require("../formatting/StylesManager");
 const NumberingManager_1 = require("../formatting/NumberingManager");
 const RelationshipManager_1 = require("./RelationshipManager");
-class ImageRun {
+const DocumentParser_1 = require("./DocumentParser");
+const DocumentGenerator_1 = require("./DocumentGenerator");
+const DocumentValidator_1 = require("./DocumentValidator");
+class ImageRun extends Run_1.Run {
     constructor(image) {
-        this.image = image;
+        super('');
+        this.imageElement = image;
     }
     toXML() {
-        const drawing = this.image.toXML();
+        const drawing = this.imageElement.toXML();
         return {
             name: 'w:r',
             children: [drawing]
@@ -69,15 +70,25 @@ class ImageRun {
 class Document {
     constructor(zipHandler, options = {}, initDefaults = true) {
         this.bodyElements = [];
-        this.parseErrors = [];
         this.zipHandler = zipHandler || new ZipHandler_1.ZipHandler();
-        this.properties = options.properties || {};
-        this.maxMemoryUsagePercent = options.maxMemoryUsagePercent ?? 80;
-        this.strictParsing = options.strictParsing ?? false;
+        const strictParsing = options.strictParsing ?? false;
+        const memoryPercent = options.maxMemoryUsagePercent ?? 80;
+        this.parser = new DocumentParser_1.DocumentParser(strictParsing);
+        this.generator = new DocumentGenerator_1.DocumentGenerator();
+        this.validator = new DocumentValidator_1.DocumentValidator(memoryPercent, {
+            maxMemoryUsagePercent: options.maxMemoryUsagePercent,
+            maxRssMB: options.maxRssMB,
+            useAbsoluteLimit: options.useAbsoluteMemoryLimit,
+        });
+        this.properties = options.properties ? DocumentValidator_1.DocumentValidator.validateProperties(options.properties) : {};
         this.stylesManager = StylesManager_1.StylesManager.create();
         this.numberingManager = NumberingManager_1.NumberingManager.create();
         this.section = Section_1.Section.createLetter();
-        this.imageManager = ImageManager_1.ImageManager.create();
+        this.imageManager = ImageManager_1.ImageManager.create({
+            maxImageCount: options.maxImageCount,
+            maxTotalImageSizeMB: options.maxTotalImageSizeMB,
+            maxSingleImageSizeMB: options.maxSingleImageSizeMB,
+        });
         this.relationshipManager = RelationshipManager_1.RelationshipManager.create();
         this.headerFooterManager = HeaderFooterManager_1.HeaderFooterManager.create();
         this.bookmarkManager = BookmarkManager_1.BookmarkManager.create();
@@ -108,346 +119,20 @@ class Document {
         return doc;
     }
     initializeRequiredFiles() {
-        this.zipHandler.addFile(types_1.DOCX_PATHS.CONTENT_TYPES, this.generateContentTypes());
-        this.zipHandler.addFile(types_1.DOCX_PATHS.RELS, this.generateRels());
-        this.zipHandler.addFile(types_1.DOCX_PATHS.DOCUMENT, this.generateDocumentXml());
+        this.zipHandler.addFile(types_1.DOCX_PATHS.CONTENT_TYPES, this.generator.generateContentTypes());
+        this.zipHandler.addFile(types_1.DOCX_PATHS.RELS, this.generator.generateRels());
+        this.zipHandler.addFile(types_1.DOCX_PATHS.DOCUMENT, this.generator.generateDocumentXml(this.bodyElements, this.section));
         this.zipHandler.addFile('word/_rels/document.xml.rels', this.relationshipManager.generateXml());
         this.zipHandler.addFile(types_1.DOCX_PATHS.STYLES, this.stylesManager.generateStylesXml());
         this.zipHandler.addFile(types_1.DOCX_PATHS.NUMBERING, this.numberingManager.generateNumberingXml());
-        this.zipHandler.addFile(types_1.DOCX_PATHS.CORE_PROPS, this.generateCoreProps());
-        this.zipHandler.addFile(types_1.DOCX_PATHS.APP_PROPS, this.generateAppProps());
+        this.zipHandler.addFile(types_1.DOCX_PATHS.CORE_PROPS, this.generator.generateCoreProps(this.properties));
+        this.zipHandler.addFile(types_1.DOCX_PATHS.APP_PROPS, this.generator.generateAppProps());
     }
     async parseDocument() {
-        const docXml = this.zipHandler.getFileAsString(types_1.DOCX_PATHS.DOCUMENT);
-        if (!docXml) {
-            throw new Error('Invalid document: word/document.xml not found');
-        }
-        this.parseRelationships();
-        this.parseProperties();
-        this.parseBodyElements(docXml);
-    }
-    parseBodyElements(docXml) {
-        this.bodyElements = [];
-        try {
-            XMLParser_1.XMLParser.validateSize(docXml);
-        }
-        catch (error) {
-            const err = error instanceof Error ? error : new Error(String(error));
-            this.parseErrors.push({ element: 'document', error: err });
-            if (this.strictParsing) {
-                throw err;
-            }
-            return;
-        }
-        const bodyContent = XMLParser_1.XMLParser.extractBody(docXml);
-        if (!bodyContent) {
-            return;
-        }
-        const paragraphXmls = XMLParser_1.XMLParser.extractElements(bodyContent, 'w:p');
-        for (const paraXml of paragraphXmls) {
-            const paragraph = this.parseParagraph(paraXml);
-            if (paragraph) {
-                this.bodyElements.push(paragraph);
-            }
-        }
-        const hasTable = bodyContent.includes('<w:tbl');
-        if (hasTable) {
-            const err = new Error('Document contains tables which are not yet fully supported in Phase 2. Tables will be ignored.');
-            this.parseErrors.push({ element: 'table', error: err });
-            if (this.strictParsing) {
-                throw err;
-            }
-        }
-        this.validateLoadedContent();
-    }
-    validateLoadedContent() {
-        const paragraphs = this.bodyElements.filter((el) => el instanceof Paragraph_1.Paragraph);
-        if (paragraphs.length === 0) {
-            return;
-        }
-        let totalRuns = 0;
-        let emptyRuns = 0;
-        let runsWithText = 0;
-        for (const para of paragraphs) {
-            const runs = para.getRuns();
-            totalRuns += runs.length;
-            for (const run of runs) {
-                const text = run.getText();
-                if (text.length === 0) {
-                    emptyRuns++;
-                }
-                else {
-                    runsWithText++;
-                }
-            }
-        }
-        if (totalRuns > 0) {
-            const emptyPercentage = (emptyRuns / totalRuns) * 100;
-            if (emptyPercentage > 90 && emptyRuns > 10) {
-                const warning = new Error(`WARNING: Document appears to be corrupted or empty. ` +
-                    `${emptyRuns} out of ${totalRuns} runs (${emptyPercentage.toFixed(1)}%) have no text content. ` +
-                    `This may indicate:\n` +
-                    `  - The document was already corrupted before loading\n` +
-                    `  - Text content was stripped by another application\n` +
-                    `  - Encoding issues during document creation\n` +
-                    `Original document structure is preserved, but text may be lost.`);
-                this.parseErrors.push({ element: 'document-validation', error: warning });
-                console.warn(`\nDocXML Load Warning:\n${warning.message}\n`);
-            }
-            else if (emptyPercentage > 50 && emptyRuns > 5) {
-                const warning = new Error(`Document has ${emptyRuns} out of ${totalRuns} runs (${emptyPercentage.toFixed(1)}%) with no text. ` +
-                    `This is higher than normal and may indicate partial data loss.`);
-                this.parseErrors.push({ element: 'document-validation', error: warning });
-                console.warn(`\nDocXML Load Warning:\n${warning.message}\n`);
-            }
-        }
-    }
-    parseParagraph(paraXml) {
-        try {
-            const paragraph = new Paragraph_1.Paragraph();
-            this.parseParagraphProperties(paraXml, paragraph);
-            const hyperlinkXmls = XMLParser_1.XMLParser.extractElements(paraXml, 'w:hyperlink');
-            for (const hyperlinkXml of hyperlinkXmls) {
-                const hyperlink = this.parseHyperlink(hyperlinkXml);
-                if (hyperlink) {
-                    paragraph.addHyperlink(hyperlink);
-                }
-            }
-            let paraXmlWithoutHyperlinks = paraXml;
-            for (const hyperlinkXml of hyperlinkXmls) {
-                paraXmlWithoutHyperlinks = paraXmlWithoutHyperlinks.replace(hyperlinkXml, '');
-            }
-            const runXmls = XMLParser_1.XMLParser.extractElements(paraXmlWithoutHyperlinks, 'w:r');
-            for (const runXml of runXmls) {
-                const run = this.parseRun(runXml);
-                if (run) {
-                    paragraph.addRun(run);
-                }
-            }
-            return paragraph;
-        }
-        catch (error) {
-            const err = error instanceof Error ? error : new Error(String(error));
-            this.parseErrors.push({ element: 'paragraph', error: err });
-            if (this.strictParsing) {
-                throw new Error(`Failed to parse paragraph: ${err.message}`);
-            }
-            return null;
-        }
-    }
-    parseParagraphProperties(paraXml, paragraph) {
-        const pPrMatch = paraXml.match(/<w:pPr[^>]*>([\s\S]*?)<\/w:pPr>/);
-        if (!pPrMatch || !pPrMatch[1]) {
-            return;
-        }
-        const pPr = pPrMatch[1];
-        const alignMatch = pPr.match(/<w:jc\s+w:val="([^"]+)"/);
-        if (alignMatch && alignMatch[1]) {
-            const alignment = alignMatch[1];
-            paragraph.setAlignment(alignment);
-        }
-        const styleMatch = pPr.match(/<w:pStyle\s+w:val="([^"]+)"/);
-        if (styleMatch && styleMatch[1]) {
-            paragraph.setStyle(styleMatch[1]);
-        }
-        const indMatch = pPr.match(/<w:ind([^>]+)\/>/);
-        if (indMatch && indMatch[1]) {
-            const indStr = indMatch[1];
-            const leftMatch = indStr.match(/w:left="(\d+)"/);
-            const rightMatch = indStr.match(/w:right="(\d+)"/);
-            const firstLineMatch = indStr.match(/w:firstLine="(\d+)"/);
-            if (leftMatch && leftMatch[1]) {
-                paragraph.setLeftIndent(parseInt(leftMatch[1], 10));
-            }
-            if (rightMatch && rightMatch[1]) {
-                paragraph.setRightIndent(parseInt(rightMatch[1], 10));
-            }
-            if (firstLineMatch && firstLineMatch[1]) {
-                paragraph.setFirstLineIndent(parseInt(firstLineMatch[1], 10));
-            }
-        }
-        const spacingMatch = pPr.match(/<w:spacing([^>]+)\/>/);
-        if (spacingMatch && spacingMatch[1]) {
-            const spacingStr = spacingMatch[1];
-            const beforeMatch = spacingStr.match(/w:before="(\d+)"/);
-            const afterMatch = spacingStr.match(/w:after="(\d+)"/);
-            const lineMatch = spacingStr.match(/w:line="(\d+)"/);
-            if (beforeMatch && beforeMatch[1]) {
-                paragraph.setSpaceBefore(parseInt(beforeMatch[1], 10));
-            }
-            if (afterMatch && afterMatch[1]) {
-                paragraph.setSpaceAfter(parseInt(afterMatch[1], 10));
-            }
-            if (lineMatch && lineMatch[1]) {
-                const lineRule = spacingStr.match(/w:lineRule="([^"]+)"/);
-                paragraph.setLineSpacing(parseInt(lineMatch[1], 10), lineRule && lineRule[1] ? lineRule[1] : undefined);
-            }
-        }
-        if (pPr.includes('<w:keepNext'))
-            paragraph.setKeepNext(true);
-        if (pPr.includes('<w:keepLines'))
-            paragraph.setKeepLines(true);
-        if (pPr.includes('<w:pageBreakBefore'))
-            paragraph.setPageBreakBefore(true);
-    }
-    parseRun(runXml) {
-        try {
-            const text = XMLBuilder_1.XMLBuilder.unescapeXml(XMLParser_1.XMLParser.extractText(runXml));
-            const run = new Run_1.Run(text);
-            this.parseRunProperties(runXml, run);
-            return run;
-        }
-        catch (error) {
-            const err = error instanceof Error ? error : new Error(String(error));
-            this.parseErrors.push({ element: 'run', error: err });
-            if (this.strictParsing) {
-                throw new Error(`Failed to parse run: ${err.message}`);
-            }
-            return null;
-        }
-    }
-    parseRunProperties(runXml, run) {
-        const rPrMatch = runXml.match(/<w:rPr[^>]*>([\s\S]*?)<\/w:rPr>/);
-        if (!rPrMatch || !rPrMatch[1]) {
-            return;
-        }
-        const rPr = rPrMatch[1];
-        if (rPr.includes('<w:b/>') || rPr.includes('<w:b ')) {
-            run.setBold(true);
-        }
-        if (rPr.includes('<w:i/>') || rPr.includes('<w:i ')) {
-            run.setItalic(true);
-        }
-        const underlineMatch = rPr.match(/<w:u\s+w:val="([^"]+)"/);
-        if (underlineMatch && underlineMatch[1]) {
-            const underlineStyle = underlineMatch[1];
-            run.setUnderline(underlineStyle);
-        }
-        else if (rPr.includes('<w:u/>')) {
-            run.setUnderline(true);
-        }
-        if (rPr.includes('<w:strike/>') || rPr.includes('<w:strike ')) {
-            run.setStrike(true);
-        }
-        const vertAlignMatch = rPr.match(/<w:vertAlign\s+w:val="([^"]+)"/);
-        if (vertAlignMatch && vertAlignMatch[1]) {
-            if (vertAlignMatch[1] === 'subscript') {
-                run.setSubscript(true);
-            }
-            else if (vertAlignMatch[1] === 'superscript') {
-                run.setSuperscript(true);
-            }
-        }
-        const fontMatch = rPr.match(/<w:rFonts[^>]+w:ascii="([^"]+)"/);
-        if (fontMatch && fontMatch[1]) {
-            run.setFont(fontMatch[1]);
-        }
-        const sizeMatch = rPr.match(/<w:sz\s+w:val="(\d+)"/);
-        if (sizeMatch && sizeMatch[1]) {
-            const halfPoints = parseInt(sizeMatch[1], 10);
-            run.setSize(halfPoints / 2);
-        }
-        const colorMatch = rPr.match(/<w:color\s+w:val="([^"]+)"/);
-        if (colorMatch && colorMatch[1]) {
-            run.setColor(colorMatch[1]);
-        }
-        const highlightMatch = rPr.match(/<w:highlight\s+w:val="([^"]+)"/);
-        if (highlightMatch && highlightMatch[1]) {
-            const highlightColor = highlightMatch[1];
-            run.setHighlight(highlightColor);
-        }
-        if (rPr.includes('<w:smallCaps/>') || rPr.includes('<w:smallCaps ')) {
-            run.setSmallCaps(true);
-        }
-        if (rPr.includes('<w:caps/>') || rPr.includes('<w:caps ')) {
-            run.setAllCaps(true);
-        }
-    }
-    parseHyperlink(hyperlinkXml) {
-        try {
-            const relationshipId = XMLParser_1.XMLParser.extractAttribute(hyperlinkXml, 'r:id');
-            const anchor = XMLParser_1.XMLParser.extractAttribute(hyperlinkXml, 'w:anchor');
-            const tooltip = XMLParser_1.XMLParser.extractAttribute(hyperlinkXml, 'w:tooltip');
-            if (!relationshipId && !anchor) {
-                return null;
-            }
-            const runXmls = XMLParser_1.XMLParser.extractElements(hyperlinkXml, 'w:r');
-            let text = '';
-            let formatting;
-            for (const runXml of runXmls) {
-                text += XMLBuilder_1.XMLBuilder.unescapeXml(XMLParser_1.XMLParser.extractText(runXml));
-                if (!formatting) {
-                    const run = this.parseRun(runXml);
-                    if (run) {
-                        formatting = run.getFormatting();
-                    }
-                }
-            }
-            let url;
-            if (relationshipId) {
-                const relationship = this.relationshipManager.getRelationship(relationshipId);
-                if (relationship && relationship.getType().includes('hyperlink')) {
-                    url = relationship.getTarget();
-                }
-            }
-            return new Hyperlink_1.Hyperlink({
-                url,
-                anchor,
-                text: text || 'Link',
-                formatting,
-                tooltip,
-                relationshipId,
-            });
-        }
-        catch (error) {
-            const err = error instanceof Error ? error : new Error(String(error));
-            this.parseErrors.push({ element: 'hyperlink', error: err });
-            if (this.strictParsing) {
-                throw new Error(`Failed to parse hyperlink: ${err.message}`);
-            }
-            return null;
-        }
-    }
-    parseRelationships() {
-        const relsPath = 'word/_rels/document.xml.rels';
-        const relsXml = this.zipHandler.getFileAsString(relsPath);
-        if (relsXml) {
-            this.relationshipManager = RelationshipManager_1.RelationshipManager.fromXml(relsXml);
-        }
-        else {
-            this.relationshipManager.addStyles();
-            this.relationshipManager.addNumbering();
-        }
-    }
-    parseProperties() {
-        const coreXml = this.zipHandler.getFileAsString(types_1.DOCX_PATHS.CORE_PROPS);
-        if (!coreXml) {
-            return;
-        }
-        const extractTag = (xml, tag) => {
-            const match = xml.match(new RegExp(`<${tag}[^>]*>([^<]*)</${tag}>`));
-            return match && match[1] ? XMLBuilder_1.XMLBuilder.unescapeXml(match[1]) : undefined;
-        };
-        this.properties = {
-            title: extractTag(coreXml, 'dc:title'),
-            subject: extractTag(coreXml, 'dc:subject'),
-            creator: extractTag(coreXml, 'dc:creator'),
-            keywords: extractTag(coreXml, 'cp:keywords'),
-            description: extractTag(coreXml, 'dc:description'),
-            lastModifiedBy: extractTag(coreXml, 'cp:lastModifiedBy'),
-        };
-        const revisionStr = extractTag(coreXml, 'cp:revision');
-        if (revisionStr) {
-            this.properties.revision = parseInt(revisionStr, 10);
-        }
-        const createdStr = extractTag(coreXml, 'dcterms:created');
-        if (createdStr) {
-            this.properties.created = new Date(createdStr);
-        }
-        const modifiedStr = extractTag(coreXml, 'dcterms:modified');
-        if (modifiedStr) {
-            this.properties.modified = new Date(modifiedStr);
-        }
+        const result = await this.parser.parseDocument(this.zipHandler, this.relationshipManager);
+        this.bodyElements = result.bodyElements;
+        this.properties = result.properties;
+        this.relationshipManager = result.relationshipManager;
     }
     addParagraph(paragraph) {
         this.bodyElements.push(paragraph);
@@ -504,48 +189,21 @@ class Document {
         return this;
     }
     setProperties(properties) {
-        this.properties = { ...this.properties, ...properties };
+        const validated = DocumentValidator_1.DocumentValidator.validateProperties(properties);
+        this.properties = { ...this.properties, ...validated };
         return this;
     }
     getProperties() {
         return { ...this.properties };
     }
-    validateBeforeSave() {
-        const paragraphs = this.getParagraphs();
-        if (paragraphs.length === 0) {
-            console.warn('\nDocXML Save Warning:\n' +
-                'Document has no paragraphs. You are saving an empty document.\n');
-            return;
-        }
-        let totalRuns = 0;
-        let emptyRuns = 0;
-        for (const para of paragraphs) {
-            const runs = para.getRuns();
-            totalRuns += runs.length;
-            for (const run of runs) {
-                if (run.getText().length === 0) {
-                    emptyRuns++;
-                }
-            }
-        }
-        if (totalRuns > 0) {
-            const emptyPercentage = (emptyRuns / totalRuns) * 100;
-            if (emptyPercentage > 90 && emptyRuns > 10) {
-                console.warn('\nDocXML Save Warning:\n' +
-                    `You are about to save a document where ${emptyRuns} out of ${totalRuns} runs (${emptyPercentage.toFixed(1)}%) are empty.\n` +
-                    'This may result in a document with no visible text content.\n' +
-                    'If this is unintentional, please review the document before saving.\n');
-            }
-        }
-    }
     async save(filePath) {
         const tempPath = `${filePath}.tmp.${Date.now()}`;
         try {
-            this.validateBeforeSave();
-            this.checkMemoryThreshold();
+            this.validator.validateBeforeSave(this.bodyElements);
+            this.validator.checkMemoryThreshold();
             await this.imageManager.loadAllImageData();
-            this.checkMemoryThreshold();
-            const sizeInfo = this.estimateSize();
+            this.validator.checkMemoryThreshold();
+            const sizeInfo = this.validator.estimateSize(this.bodyElements, this.imageManager);
             if (sizeInfo.warning) {
                 console.warn(`DocXML Warning: ${sizeInfo.warning}`);
             }
@@ -579,11 +237,11 @@ class Document {
     }
     async toBuffer() {
         try {
-            this.validateBeforeSave();
-            this.checkMemoryThreshold();
+            this.validator.validateBeforeSave(this.bodyElements);
+            this.validator.checkMemoryThreshold();
             await this.imageManager.loadAllImageData();
-            this.checkMemoryThreshold();
-            const sizeInfo = this.estimateSize();
+            this.validator.checkMemoryThreshold();
+            const sizeInfo = this.validator.estimateSize(this.bodyElements, this.imageManager);
             if (sizeInfo.warning) {
                 console.warn(`DocXML Warning: ${sizeInfo.warning}`);
             }
@@ -605,11 +263,11 @@ class Document {
         }
     }
     updateDocumentXml() {
-        const xml = this.generateDocumentXml();
+        const xml = this.generator.generateDocumentXml(this.bodyElements, this.section);
         this.zipHandler.updateFile(types_1.DOCX_PATHS.DOCUMENT, xml);
     }
     updateCoreProps() {
-        const xml = this.generateCoreProps();
+        const xml = this.generator.generateCoreProps(this.properties);
         this.zipHandler.updateFile(types_1.DOCX_PATHS.CORE_PROPS, xml);
     }
     updateStylesXml() {
@@ -620,78 +278,6 @@ class Document {
         const xml = this.numberingManager.generateNumberingXml();
         this.zipHandler.updateFile(types_1.DOCX_PATHS.NUMBERING, xml);
     }
-    generateContentTypes() {
-        return `<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
-<Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types">
-  <Default Extension="rels" ContentType="application/vnd.openxmlformats-package.relationships+xml"/>
-  <Default Extension="xml" ContentType="application/xml"/>
-  <Override PartName="/word/document.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.document.main+xml"/>
-  <Override PartName="/word/styles.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.styles+xml"/>
-  <Override PartName="/word/numbering.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.numbering+xml"/>
-  <Override PartName="/docProps/core.xml" ContentType="application/vnd.openxmlformats-package.core-properties+xml"/>
-  <Override PartName="/docProps/app.xml" ContentType="application/vnd.openxmlformats-officedocument.extended-properties+xml"/>
-</Types>`;
-    }
-    generateRels() {
-        return `<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
-<Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">
-  <Relationship Id="rId1" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument" Target="word/document.xml"/>
-  <Relationship Id="rId2" Type="http://schemas.openxmlformats.org/package/2006/relationships/metadata/core-properties" Target="docProps/core.xml"/>
-  <Relationship Id="rId3" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/extended-properties" Target="docProps/app.xml"/>
-</Relationships>`;
-    }
-    generateDocumentXml() {
-        const bodyXmls = [];
-        for (const element of this.bodyElements) {
-            const xml = element.toXML();
-            if (Array.isArray(xml)) {
-                bodyXmls.push(...xml);
-            }
-            else {
-                bodyXmls.push(xml);
-            }
-        }
-        bodyXmls.push(this.section.toXML());
-        return XMLBuilder_1.XMLBuilder.createDocument(bodyXmls);
-    }
-    generateCoreProps() {
-        const now = new Date();
-        const created = this.properties.created || now;
-        const modified = this.properties.modified || now;
-        const formatDate = (date) => {
-            return date.toISOString();
-        };
-        return `<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
-<cp:coreProperties xmlns:cp="http://schemas.openxmlformats.org/package/2006/metadata/core-properties"
-                   xmlns:dc="http://purl.org/dc/elements/1.1/"
-                   xmlns:dcterms="http://purl.org/dc/terms/"
-                   xmlns:dcmitype="http://purl.org/dc/dcmitype/"
-                   xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
-  <dc:title>${XMLBuilder_1.XMLBuilder.escapeXmlText(this.properties.title || '')}</dc:title>
-  <dc:subject>${XMLBuilder_1.XMLBuilder.escapeXmlText(this.properties.subject || '')}</dc:subject>
-  <dc:creator>${XMLBuilder_1.XMLBuilder.escapeXmlText(this.properties.creator || 'DocXML')}</dc:creator>
-  <cp:keywords>${XMLBuilder_1.XMLBuilder.escapeXmlText(this.properties.keywords || '')}</cp:keywords>
-  <dc:description>${XMLBuilder_1.XMLBuilder.escapeXmlText(this.properties.description || '')}</dc:description>
-  <cp:lastModifiedBy>${XMLBuilder_1.XMLBuilder.escapeXmlText(this.properties.lastModifiedBy || this.properties.creator || 'DocXML')}</cp:lastModifiedBy>
-  <cp:revision>${this.properties.revision || 1}</cp:revision>
-  <dcterms:created xsi:type="dcterms:W3CDTF">${formatDate(created)}</dcterms:created>
-  <dcterms:modified xsi:type="dcterms:W3CDTF">${formatDate(modified)}</dcterms:modified>
-</cp:coreProperties>`;
-    }
-    generateAppProps() {
-        return `<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
-<Properties xmlns="http://schemas.openxmlformats.org/officeDocument/2006/extended-properties"
-            xmlns:vt="http://schemas.openxmlformats.org/officeDocument/2006/docPropsVTypes">
-  <Application>DocXML</Application>
-  <DocSecurity>0</DocSecurity>
-  <ScaleCrop>false</ScaleCrop>
-  <Company></Company>
-  <LinksUpToDate>false</LinksUpToDate>
-  <SharedDoc>false</SharedDoc>
-  <HyperlinksChanged>false</HyperlinksChanged>
-  <AppVersion>0.1.0</AppVersion>
-</Properties>`;
-    }
     getStylesManager() {
         return this.stylesManager;
     }
@@ -800,38 +386,7 @@ class Document {
         return this.relationshipManager;
     }
     processHyperlinks() {
-        const paragraphs = this.getParagraphs();
-        const headers = this.headerFooterManager.getAllHeaders();
-        const footers = this.headerFooterManager.getAllFooters();
-        for (const header of headers) {
-            for (const element of header.header.getElements()) {
-                if (element instanceof Paragraph_1.Paragraph) {
-                    this.processHyperlinksInParagraph(element);
-                }
-            }
-        }
-        for (const footer of footers) {
-            for (const element of footer.footer.getElements()) {
-                if (element instanceof Paragraph_1.Paragraph) {
-                    this.processHyperlinksInParagraph(element);
-                }
-            }
-        }
-        for (const para of paragraphs) {
-            this.processHyperlinksInParagraph(para);
-        }
-    }
-    processHyperlinksInParagraph(paragraph) {
-        const content = paragraph.getContent();
-        for (const item of content) {
-            if (item instanceof Hyperlink_1.Hyperlink && item.isExternal() && !item.getRelationshipId()) {
-                const url = item.getUrl();
-                if (url) {
-                    const relationship = this.relationshipManager.addHyperlink(url);
-                    item.setRelationshipId(relationship.getId());
-                }
-            }
-        }
+        this.generator.processHyperlinks(this.bodyElements, this.headerFooterManager, this.relationshipManager);
     }
     saveImages() {
         const images = this.imageManager.getAllImages();
@@ -871,43 +426,9 @@ class Document {
         }
     }
     updateContentTypesWithImagesHeadersFootersAndComments() {
-        const contentTypes = this.generateContentTypesWithImagesHeadersFootersAndComments();
+        const contentTypes = this.generator.generateContentTypesWithImagesHeadersFootersAndComments(this.imageManager, this.headerFooterManager, this.commentManager);
         this.zipHandler.updateFile(types_1.DOCX_PATHS.CONTENT_TYPES, contentTypes);
     }
-    generateContentTypesWithImagesHeadersFootersAndComments() {
-        const images = this.imageManager.getAllImages();
-        const headers = this.headerFooterManager.getAllHeaders();
-        const footers = this.headerFooterManager.getAllFooters();
-        const hasComments = this.commentManager.getCount() > 0;
-        const extensions = new Set();
-        for (const entry of images) {
-            extensions.add(entry.image.getExtension());
-        }
-        let xml = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?>\n';
-        xml += '<Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types">\n';
-        xml += '  <Default Extension="rels" ContentType="application/vnd.openxmlformats-package.relationships+xml"/>\n';
-        xml += '  <Default Extension="xml" ContentType="application/xml"/>\n';
-        for (const ext of extensions) {
-            const mimeType = ImageManager_1.ImageManager.getMimeType(ext);
-            xml += `  <Default Extension="${ext}" ContentType="${mimeType}"/>\n`;
-        }
-        xml += '  <Override PartName="/word/document.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.document.main+xml"/>\n';
-        xml += '  <Override PartName="/word/styles.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.styles+xml"/>\n';
-        xml += '  <Override PartName="/word/numbering.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.numbering+xml"/>\n';
-        for (const entry of headers) {
-            xml += `  <Override PartName="/word/${entry.filename}" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.header+xml"/>\n`;
-        }
-        for (const entry of footers) {
-            xml += `  <Override PartName="/word/${entry.filename}" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.footer+xml"/>\n`;
-        }
-        if (hasComments) {
-            xml += '  <Override PartName="/word/comments.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.comments+xml"/>\n';
-        }
-        xml += '  <Override PartName="/docProps/core.xml" ContentType="application/vnd.openxmlformats-package.core-properties+xml"/>\n';
-        xml += '  <Override PartName="/docProps/app.xml" ContentType="application/vnd.openxmlformats-officedocument.extended-properties+xml"/>\n';
-        xml += '</Types>';
-        return xml;
-    }
     getBookmarkManager() {
         return this.bookmarkManager;
     }
@@ -1020,53 +541,14 @@ class Document {
         return this.revisionManager.getAllDeletions();
     }
     getParseWarnings() {
-        return [...this.parseErrors];
-    }
-    checkMemoryThreshold() {
-        const { heapUsed, heapTotal } = process.memoryUsage();
-        const usagePercent = (heapUsed / heapTotal) * 100;
-        if (usagePercent > this.maxMemoryUsagePercent) {
-            throw new Error(`Memory usage critical (${usagePercent.toFixed(1)}% of ${(heapTotal / 1024 / 1024).toFixed(0)}MB heap). ` +
-                `Cannot process document safely. Consider:\n` +
-                `- Reducing document size\n` +
-                `- Optimizing/compressing images\n` +
-                `- Splitting into multiple documents\n` +
-                `- Increasing Node.js heap size (--max-old-space-size)`);
-        }
+        return this.parser.getParseErrors();
     }
     estimateSize() {
-        const paragraphCount = this.getParagraphCount();
-        const tableCount = this.getTableCount();
-        const imageCount = this.imageManager.getImageCount();
-        const estimatedXml = (paragraphCount * 200) + (tableCount * 1000) + 50000;
-        const imageBytes = this.imageManager.getTotalSize();
-        const totalBytes = estimatedXml + imageBytes;
-        const totalMB = totalBytes / (1024 * 1024);
-        const WARNING_MB = 50;
-        const ERROR_MB = 100;
-        let warning;
-        if (totalMB > ERROR_MB) {
-            warning = `Document size (${totalMB.toFixed(1)}MB) exceeds recommended maximum of ${ERROR_MB}MB. ` +
-                `This may cause memory issues. Consider splitting into multiple documents or optimizing images.`;
-        }
-        else if (totalMB > WARNING_MB) {
-            warning = `Document size (${totalMB.toFixed(1)}MB) exceeds ${WARNING_MB}MB. ` +
-                `Large documents may take longer to process and use significant memory.`;
-        }
-        return {
-            paragraphs: paragraphCount,
-            tables: tableCount,
-            images: imageCount,
-            estimatedXmlBytes: estimatedXml,
-            imageBytes,
-            totalEstimatedBytes: totalBytes,
-            totalEstimatedMB: parseFloat(totalMB.toFixed(2)),
-            warning,
-        };
+        return this.validator.estimateSize(this.bodyElements, this.imageManager);
     }
     dispose() {
         this.bodyElements = [];
-        this.parseErrors = [];
+        this.parser.clearParseErrors();
         this.stylesManager = StylesManager_1.StylesManager.create();
         this.numberingManager = NumberingManager_1.NumberingManager.create();
         this.imageManager.clear();
@@ -1078,31 +560,7 @@ class Document {
         this.commentManager.clear();
     }
     getSizeStats() {
-        const estimate = this.estimateSize();
-        const warnings = [];
-        if (estimate.warning) {
-            warnings.push(estimate.warning);
-        }
-        const formatBytes = (bytes) => {
-            if (bytes < 1024)
-                return `${bytes} B`;
-            if (bytes < 1024 * 1024)
-                return `${(bytes / 1024).toFixed(1)} KB`;
-            return `${(bytes / (1024 * 1024)).toFixed(1)} MB`;
-        };
-        return {
-            elements: {
-                paragraphs: estimate.paragraphs,
-                tables: estimate.tables,
-                images: estimate.images,
-            },
-            size: {
-                xml: formatBytes(estimate.estimatedXmlBytes),
-                images: formatBytes(estimate.imageBytes),
-                total: formatBytes(estimate.totalEstimatedBytes),
-            },
-            warnings,
-        };
+        return this.validator.getSizeStats(this.bodyElements, this.imageManager);
     }
 }
 exports.Document = Document;