docxmlater 8.1.0 → 8.3.0

package/dist/ms-doc/subdocuments/SubdocumentParser.js

@@ -0,0 +1,208 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.SubdocumentParser = exports.SubdocumentParseError = void 0;
+class SubdocumentParseError extends Error {
+    constructor(message) {
+        super(message);
+        this.name = 'SubdocumentParseError';
+    }
+}
+exports.SubdocumentParseError = SubdocumentParseError;
+class SubdocumentParser {
+    fib;
+    fullText;
+    constructor(fib, fullText) {
+        this.fib = fib;
+        this.fullText = fullText;
+    }
+    getSubdocumentRanges() {
+        const rgLw = this.fib.rgLw97;
+        const mainEnd = rgLw.ccpText;
+        const footnotesStart = mainEnd;
+        const footnotesEnd = footnotesStart + rgLw.ccpFtn;
+        const headersStart = footnotesEnd;
+        const headersEnd = headersStart + rgLw.ccpHdd;
+        const commentsStart = headersEnd;
+        const commentsEnd = commentsStart + rgLw.ccpAtn;
+        const endnotesStart = commentsEnd;
+        const endnotesEnd = endnotesStart + rgLw.ccpEdn;
+        const textboxesStart = endnotesEnd;
+        const textboxesEnd = textboxesStart + rgLw.ccpTxbx;
+        const headerTextboxesStart = textboxesEnd;
+        const headerTextboxesEnd = headerTextboxesStart + rgLw.ccpHdrTxbx;
+        const result = {
+            main: { cpStart: 0, cpEnd: mainEnd },
+        };
+        if (rgLw.ccpFtn > 0) {
+            result.footnotes = { cpStart: footnotesStart, cpEnd: footnotesEnd };
+        }
+        if (rgLw.ccpHdd > 0) {
+            result.headers = { cpStart: headersStart, cpEnd: headersEnd };
+        }
+        if (rgLw.ccpAtn > 0) {
+            result.comments = { cpStart: commentsStart, cpEnd: commentsEnd };
+        }
+        if (rgLw.ccpEdn > 0) {
+            result.endnotes = { cpStart: endnotesStart, cpEnd: endnotesEnd };
+        }
+        if (rgLw.ccpTxbx > 0) {
+            result.textboxes = { cpStart: textboxesStart, cpEnd: textboxesEnd };
+        }
+        if (rgLw.ccpHdrTxbx > 0) {
+            result.headerTextboxes = { cpStart: headerTextboxesStart, cpEnd: headerTextboxesEnd };
+        }
+        return result;
+    }
+    getMainDocumentText() {
+        const ranges = this.getSubdocumentRanges();
+        return this.fullText.substring(0, ranges.main.cpEnd);
+    }
+    getFootnotesText() {
+        const ranges = this.getSubdocumentRanges();
+        if (!ranges.footnotes)
+            return null;
+        return this.fullText.substring(ranges.footnotes.cpStart, ranges.footnotes.cpEnd);
+    }
+    getHeadersFootersText() {
+        const ranges = this.getSubdocumentRanges();
+        if (!ranges.headers)
+            return null;
+        return this.fullText.substring(ranges.headers.cpStart, ranges.headers.cpEnd);
+    }
+    getCommentsText() {
+        const ranges = this.getSubdocumentRanges();
+        if (!ranges.comments)
+            return null;
+        return this.fullText.substring(ranges.comments.cpStart, ranges.comments.cpEnd);
+    }
+    getEndnotesText() {
+        const ranges = this.getSubdocumentRanges();
+        if (!ranges.endnotes)
+            return null;
+        return this.fullText.substring(ranges.endnotes.cpStart, ranges.endnotes.cpEnd);
+    }
+    parseHeadersFooters() {
+        const results = [];
+        const hfText = this.getHeadersFootersText();
+        if (!hfText)
+            return results;
+        const entries = hfText.split('\r');
+        let entryIndex = 0;
+        let sectionIndex = 0;
+        const entriesPerSection = 6;
+        const types = [
+            { type: 'header', subtype: 'even' },
+            { type: 'header', subtype: 'odd' },
+            { type: 'footer', subtype: 'even' },
+            { type: 'footer', subtype: 'odd' },
+            { type: 'header', subtype: 'first' },
+            { type: 'footer', subtype: 'first' },
+        ];
+        const ranges = this.getSubdocumentRanges();
+        let cpOffset = ranges.headers?.cpStart ?? 0;
+        for (const entry of entries) {
+            const slotIndex = entryIndex % entriesPerSection;
+            const typeInfo = types[slotIndex];
+            if (entry.length > 0 && typeInfo) {
+                results.push({
+                    type: typeInfo.type === 'header' ? 'header' : 'footer',
+                    cpStart: cpOffset,
+                    cpEnd: cpOffset + entry.length,
+                    text: entry,
+                    sectionIndex,
+                    headerFooterType: typeInfo.subtype,
+                });
+            }
+            cpOffset += entry.length + 1;
+            entryIndex++;
+            if (entryIndex % entriesPerSection === 0) {
+                sectionIndex++;
+            }
+        }
+        return results;
+    }
+    parseFootnotes() {
+        const results = [];
+        const fnText = this.getFootnotesText();
+        if (!fnText)
+            return results;
+        const entries = fnText.split('\r');
+        const ranges = this.getSubdocumentRanges();
+        let cpOffset = ranges.footnotes?.cpStart ?? 0;
+        for (const entry of entries) {
+            if (entry.length > 0) {
+                results.push({
+                    type: 'footnote',
+                    cpStart: cpOffset,
+                    cpEnd: cpOffset + entry.length,
+                    text: entry,
+                });
+            }
+            cpOffset += entry.length + 1;
+        }
+        return results;
+    }
+    parseEndnotes() {
+        const results = [];
+        const enText = this.getEndnotesText();
+        if (!enText)
+            return results;
+        const entries = enText.split('\r');
+        const ranges = this.getSubdocumentRanges();
+        let cpOffset = ranges.endnotes?.cpStart ?? 0;
+        for (const entry of entries) {
+            if (entry.length > 0) {
+                results.push({
+                    type: 'endnote',
+                    cpStart: cpOffset,
+                    cpEnd: cpOffset + entry.length,
+                    text: entry,
+                });
+            }
+            cpOffset += entry.length + 1;
+        }
+        return results;
+    }
+    parseComments() {
+        const results = [];
+        const cmtText = this.getCommentsText();
+        if (!cmtText)
+            return results;
+        const entries = cmtText.split('\r');
+        const ranges = this.getSubdocumentRanges();
+        let cpOffset = ranges.comments?.cpStart ?? 0;
+        for (const entry of entries) {
+            if (entry.length > 0) {
+                results.push({
+                    type: 'comment',
+                    cpStart: cpOffset,
+                    cpEnd: cpOffset + entry.length,
+                    text: entry,
+                });
+            }
+            cpOffset += entry.length + 1;
+        }
+        return results;
+    }
+    parseAll() {
+        const results = [];
+        const mainText = this.getMainDocumentText();
+        results.push({
+            type: 'main',
+            cpStart: 0,
+            cpEnd: mainText.length,
+            text: mainText,
+        });
+        results.push(...this.parseHeadersFooters());
+        results.push(...this.parseFootnotes());
+        results.push(...this.parseEndnotes());
+        results.push(...this.parseComments());
+        return results;
+    }
+    static parse(fib, fullText) {
+        const parser = new SubdocumentParser(fib, fullText);
+        return parser.parseAll();
+    }
+}
+exports.SubdocumentParser = SubdocumentParser;
+//# sourceMappingURL=SubdocumentParser.js.map

package/dist/ms-doc/subdocuments/SubdocumentParser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"SubdocumentParser.js","sourceRoot":"","sources":["../../../src/ms-doc/subdocuments/SubdocumentParser.ts"],"names":[],"mappings":";;;AAsBA,MAAa,qBAAsB,SAAQ,KAAK;IAC9C,YAAY,OAAe;QACzB,KAAK,CAAC,OAAO,CAAC,CAAC;QACf,IAAI,CAAC,IAAI,GAAG,uBAAuB,CAAC;IACtC,CAAC;CACF;AALD,sDAKC;AA+CD,MAAa,iBAAiB;IACpB,GAAG,CAAM;IACT,QAAQ,CAAS;IAEzB,YAAY,GAAQ,EAAE,QAAgB;QACpC,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC;QACf,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;IAC3B,CAAC;IAKD,oBAAoB;QASlB,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC;QAI7B,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;QAG7B,MAAM,cAAc,GAAG,OAAO,CAAC;QAC/B,MAAM,YAAY,GAAG,cAAc,GAAG,IAAI,CAAC,MAAM,CAAC;QAGlD,MAAM,YAAY,GAAG,YAAY,CAAC;QAClC,MAAM,UAAU,GAAG,YAAY,GAAG,IAAI,CAAC,MAAM,CAAC;QAG9C,MAAM,aAAa,GAAG,UAAU,CAAC;QACjC,MAAM,WAAW,GAAG,aAAa,GAAG,IAAI,CAAC,MAAM,CAAC;QAGhD,MAAM,aAAa,GAAG,WAAW,CAAC;QAClC,MAAM,WAAW,GAAG,aAAa,GAAG,IAAI,CAAC,MAAM,CAAC;QAGhD,MAAM,cAAc,GAAG,WAAW,CAAC;QACnC,MAAM,YAAY,GAAG,cAAc,GAAG,IAAI,CAAC,OAAO,CAAC;QAGnD,MAAM,oBAAoB,GAAG,YAAY,CAAC;QAC1C,MAAM,kBAAkB,GAAG,oBAAoB,GAAG,IAAI,CAAC,UAAU,CAAC;QAElE,MAAM,MAAM,GAAiD;YAC3D,IAAI,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;SACrC,CAAC;QAEF,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACpB,MAAM,CAAC,SAAS,GAAG,EAAE,OAAO,EAAE,cAAc,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC;QACtE,CAAC;QACD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACpB,MAAM,CAAC,OAAO,GAAG,EAAE,OAAO,EAAE,YAAY,EAAE,KAAK,EAAE,UAAU,EAAE,CAAC;QAChE,CAAC;QACD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACpB,MAAM,CAAC,QAAQ,GAAG,EAAE,OAAO,EAAE,aAAa,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC;QACnE,CAAC;QACD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACpB,MAAM,CAAC,QAAQ,GAAG,EAAE,OAAO,EAAE,aAAa,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC;QACnE,CAAC;QACD,IAAI,IAAI,CAAC,OAAO,GAAG,CAAC,EAAE,CAAC;YACrB,MAAM,CAAC,SAAS,GAAG,EAAE,OAAO,EAAE,cAAc,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC;QACtE,CAAC;QACD,IAAI,IAAI,CAAC,UAAU,GAAG,CAAC,EAAE,CAAC;YACxB,MAAM,CAAC,eAAe,GAAG,EAAE,OAAO,EAAE,oBAAoB,EAAE,KAAK,EAAE,kBAAkB,EAAE,CAAC;QACxF,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAKD,mBAAmB;QACjB,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACvD,CAAC;IAKD,gBAAgB;QACd,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,CAAC,MAAM,CAAC,SAAS;YAAE,OAAO,IAAI,CAAC;QACnC,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,SAAS,CAAC,OAAO,EAAE,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;IACnF,CAAC;IAKD,qBAAqB;QACnB,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,CAAC,MAAM,CAAC,OAAO;YAAE,OAAO,IAAI,CAAC;QACjC,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,OAAO,CAAC,OAAO,EAAE,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;IAC/E,CAAC;IAKD,eAAe;QACb,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,CAAC,MAAM,CAAC,QAAQ;YAAE,OAAO,IAAI,CAAC;QAClC,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IACjF,CAAC;IAKD,eAAe;QACb,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,CAAC,MAAM,CAAC,QAAQ;YAAE,OAAO,IAAI,CAAC;QAClC,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IACjF,CAAC;IAeD,mBAAmB;QACjB,MAAM,OAAO,GAAwB,EAAE,CAAC;QACxC,MAAM,MAAM,GAAG,IAAI,CAAC,qBAAqB,EAAE,CAAC;QAE5C,IAAI,CAAC,MAAM;YAAE,OAAO,OAAO,CAAC;QAG5B,MAAM,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACnC,IAAI,UAAU,GAAG,CAAC,CAAC;QACnB,IAAI,YAAY,GAAG,CAAC,CAAC;QAGrB,MAAM,iBAAiB,GAAG,CAAC,CAAC;QAC5B,MAAM,KAAK,GAAoE;YAC7E,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE;YACnC,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE;YAClC,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE;YACnC,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE;YAClC,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE;YACpC,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE;SACrC,CAAC;QAEF,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,QAAQ,GAAG,MAAM,CAAC,OAAO,EAAE,OAAO,IAAI,CAAC,CAAC;QAE5C,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,MAAM,SAAS,GAAG,UAAU,GAAG,iBAAiB,CAAC;YACjD,MAAM,QAAQ,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC;YAElC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,QAAQ,EAAE,CAAC;gBACjC,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,QAAQ,CAAC,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ;oBACtD,OAAO,EAAE,QAAQ;oBACjB,KAAK,EAAE,QAAQ,GAAG,KAAK,CAAC,MAAM;oBAC9B,IAAI,EAAE,KAAK;oBACX,YAAY;oBACZ,gBAAgB,EAAE,QAAQ,CAAC,OAAO;iBACnC,CAAC,CAAC;YACL,CAAC;YAED,QAAQ,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;YAC7B,UAAU,EAAE,CAAC;YAGb,IAAI,UAAU,GAAG,iBAAiB,KAAK,CAAC,EAAE,CAAC;gBACzC,YAAY,EAAE,CAAC;YACjB,CAAC;QACH,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAKD,cAAc;QACZ,MAAM,OAAO,GAAwB,EAAE,CAAC;QACxC,MAAM,MAAM,GAAG,IAAI,CAAC,gBAAgB,EAAE,CAAC;QAEvC,IAAI,CAAC,MAAM;YAAE,OAAO,OAAO,CAAC;QAG5B,MAAM,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACnC,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,QAAQ,GAAG,MAAM,CAAC,SAAS,EAAE,OAAO,IAAI,CAAC,CAAC;QAE9C,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,UAAU;oBAChB,OAAO,EAAE,QAAQ;oBACjB,KAAK,EAAE,QAAQ,GAAG,KAAK,CAAC,MAAM;oBAC9B,IAAI,EAAE,KAAK;iBACZ,CAAC,CAAC;YACL,CAAC;YACD,QAAQ,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;QAC/B,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAKD,aAAa;QACX,MAAM,OAAO,GAAwB,EAAE,CAAC;QACxC,MAAM,MAAM,GAAG,IAAI,CAAC,eAAe,EAAE,CAAC;QAEtC,IAAI,CAAC,MAAM;YAAE,OAAO,OAAO,CAAC;QAE5B,MAAM,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACnC,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,QAAQ,GAAG,MAAM,CAAC,QAAQ,EAAE,OAAO,IAAI,CAAC,CAAC;QAE7C,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,SAAS;oBACf,OAAO,EAAE,QAAQ;oBACjB,KAAK,EAAE,QAAQ,GAAG,KAAK,CAAC,MAAM;oBAC9B,IAAI,EAAE,KAAK;iBACZ,CAAC,CAAC;YACL,CAAC;YACD,QAAQ,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;QAC/B,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAKD,aAAa;QACX,MAAM,OAAO,GAAwB,EAAE,CAAC;QACxC,MAAM,OAAO,GAAG,IAAI,CAAC,eAAe,EAAE,CAAC;QAEvC,IAAI,CAAC,OAAO;YAAE,OAAO,OAAO,CAAC;QAE7B,MAAM,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACpC,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;QAC3C,IAAI,QAAQ,GAAG,MAAM,CAAC,QAAQ,EAAE,OAAO,IAAI,CAAC,CAAC;QAE7C,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,SAAS;oBACf,OAAO,EAAE,QAAQ;oBACjB,KAAK,EAAE,QAAQ,GAAG,KAAK,CAAC,MAAM;oBAC9B,IAAI,EAAE,KAAK;iBACZ,CAAC,CAAC;YACL,CAAC;YACD,QAAQ,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;QAC/B,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAKD,QAAQ;QACN,MAAM,OAAO,GAAwB,EAAE,CAAC;QAGxC,MAAM,QAAQ,GAAG,IAAI,CAAC,mBAAmB,EAAE,CAAC;QAC5C,OAAO,CAAC,IAAI,CAAC;YACX,IAAI,EAAE,MAAM;YACZ,OAAO,EAAE,CAAC;YACV,KAAK,EAAE,QAAQ,CAAC,MAAM;YACtB,IAAI,EAAE,QAAQ;SACf,CAAC,CAAC;QAGH,OAAO,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,mBAAmB,EAAE,CAAC,CAAC;QAG5C,OAAO,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,cAAc,EAAE,CAAC,CAAC;QAGvC,OAAO,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC,CAAC;QAGtC,OAAO,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC,CAAC;QAEtC,OAAO,OAAO,CAAC;IACjB,CAAC;IAKD,MAAM,CAAC,KAAK,CAAC,GAAQ,EAAE,QAAgB;QACrC,MAAM,MAAM,GAAG,IAAI,iBAAiB,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;QACpD,OAAO,MAAM,CAAC,QAAQ,EAAE,CAAC;IAC3B,CAAC;CACF;AApTD,8CAoTC"}

package/dist/ms-doc/tables/TableParser.d.ts

@@ -0,0 +1,29 @@
+import { TableDefinition, TableProperties } from '../types/DocTypes';
+export declare class TableParseError extends Error {
+    constructor(message: string);
+}
+export interface TableBoundary {
+    cpStart: number;
+    cpEnd: number;
+    rowCount: number;
+    maxColumns: number;
+}
+export declare class TableParser {
+    private text;
+    private paragraphs;
+    constructor(documentText: string);
+    findTables(): TableBoundary[];
+    parseTable(boundary: TableBoundary): TableDefinition;
+    parseAllTables(): TableDefinition[];
+    static hasTableContent(text: string): boolean;
+    static extractRows(text: string): string[][];
+}
+export declare class TAPParser {
+    private data;
+    private view;
+    constructor(tapData: Uint8Array);
+    parse(): TableProperties;
+    parseCellWidths(cellCount: number): number[];
+    static isTableRow(papxData: Uint8Array): boolean;
+}
+//# sourceMappingURL=TableParser.d.ts.map

package/dist/ms-doc/tables/TableParser.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"TableParser.d.ts","sourceRoot":"","sources":["../../../src/ms-doc/tables/TableParser.ts"],"names":[],"mappings":"AAcA,OAAO,EAAE,eAAe,EAA2C,eAAe,EAAE,MAAM,mBAAmB,CAAC;AAK9G,qBAAa,eAAgB,SAAQ,KAAK;gBAC5B,OAAO,EAAE,MAAM;CAI5B;AAmBD,MAAM,WAAW,aAAa;IAE5B,OAAO,EAAE,MAAM,CAAC;IAEhB,KAAK,EAAE,MAAM,CAAC;IAEd,QAAQ,EAAE,MAAM,CAAC;IAEjB,UAAU,EAAE,MAAM,CAAC;CACpB;AAKD,qBAAa,WAAW;IACtB,OAAO,CAAC,IAAI,CAAS;IACrB,OAAO,CAAC,UAAU,CAAW;gBAEjB,YAAY,EAAE,MAAM;IAShC,UAAU,IAAI,aAAa,EAAE;IAgD7B,UAAU,CAAC,QAAQ,EAAE,aAAa,GAAG,eAAe;IAiDpD,cAAc,IAAI,eAAe,EAAE;IAQnC,MAAM,CAAC,eAAe,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO;IAQ7C,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,EAAE;CAc7C;AAQD,qBAAa,SAAS;IACpB,OAAO,CAAC,IAAI,CAAa;IACzB,OAAO,CAAC,IAAI,CAAW;gBAEX,OAAO,EAAE,UAAU;IAQ/B,KAAK,IAAI,eAAe;IAiDxB,eAAe,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,EAAE;IAsB5C,MAAM,CAAC,UAAU,CAAC,QAAQ,EAAE,UAAU,GAAG,OAAO;CAejD"}

package/dist/ms-doc/tables/TableParser.js

@@ -0,0 +1,176 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.TAPParser = exports.TableParser = exports.TableParseError = void 0;
+const Constants_1 = require("../types/Constants");
+class TableParseError extends Error {
+    constructor(message) {
+        super(message);
+        this.name = 'TableParseError';
+    }
+}
+exports.TableParseError = TableParseError;
+class TableParser {
+    text;
+    paragraphs;
+    constructor(documentText) {
+        this.text = documentText;
+        this.paragraphs = documentText.split(/[\r\n\x0D]/);
+    }
+    findTables() {
+        const tables = [];
+        const cellMark = String.fromCharCode(Constants_1.SPECIAL_CHARS.CELL);
+        let currentTable = null;
+        let cpPosition = 0;
+        for (const para of this.paragraphs) {
+            const hasCell = para.includes(cellMark);
+            if (hasCell) {
+                const cellCount = (para.match(new RegExp(cellMark, 'g')) ?? []).length;
+                if (!currentTable) {
+                    currentTable = {
+                        cpStart: cpPosition,
+                        cpEnd: cpPosition + para.length,
+                        rowCount: 1,
+                        maxColumns: cellCount,
+                    };
+                }
+                else {
+                    currentTable.cpEnd = cpPosition + para.length;
+                    currentTable.rowCount++;
+                    currentTable.maxColumns = Math.max(currentTable.maxColumns, cellCount);
+                }
+            }
+            else if (currentTable) {
+                tables.push(currentTable);
+                currentTable = null;
+            }
+            cpPosition += para.length + 1;
+        }
+        if (currentTable) {
+            tables.push(currentTable);
+        }
+        return tables;
+    }
+    parseTable(boundary) {
+        const cellMark = String.fromCharCode(Constants_1.SPECIAL_CHARS.CELL);
+        const rows = [];
+        let cpPosition = 0;
+        let rowIndex = 0;
+        for (const para of this.paragraphs) {
+            if (cpPosition >= boundary.cpStart && cpPosition <= boundary.cpEnd) {
+                if (para.includes(cellMark)) {
+                    const cellTexts = para.split(cellMark).filter((_, idx, arr) => idx < arr.length - 1);
+                    const cells = cellTexts.map((text, colIndex) => ({
+                        text: text.trim(),
+                        rowIndex,
+                        colIndex,
+                        rowSpan: 1,
+                        colSpan: 1,
+                    }));
+                    rows.push({
+                        index: rowIndex,
+                        cells,
+                        isHeader: rowIndex === 0,
+                    });
+                    rowIndex++;
+                }
+            }
+            if (cpPosition > boundary.cpEnd) {
+                break;
+            }
+            cpPosition += para.length + 1;
+        }
+        return {
+            rows,
+            rowCount: rows.length,
+            columnCount: boundary.maxColumns,
+            cpStart: boundary.cpStart,
+            cpEnd: boundary.cpEnd,
+        };
+    }
+    parseAllTables() {
+        const boundaries = this.findTables();
+        return boundaries.map((b) => this.parseTable(b));
+    }
+    static hasTableContent(text) {
+        const cellMark = String.fromCharCode(Constants_1.SPECIAL_CHARS.CELL);
+        return text.includes(cellMark);
+    }
+    static extractRows(text) {
+        const cellMark = String.fromCharCode(Constants_1.SPECIAL_CHARS.CELL);
+        const paragraphs = text.split(/[\r\n\x0D]/);
+        const rows = [];
+        for (const para of paragraphs) {
+            if (para.includes(cellMark)) {
+                const cells = para.split(cellMark).filter((_, idx, arr) => idx < arr.length - 1);
+                rows.push(cells.map((c) => c.trim()));
+            }
+        }
+        return rows;
+    }
+}
+exports.TableParser = TableParser;
+class TAPParser {
+    data;
+    view;
+    constructor(tapData) {
+        this.data = tapData;
+        this.view = new DataView(this.data.buffer, this.data.byteOffset, this.data.byteLength);
+    }
+    parse() {
+        const props = {};
+        if (this.data.length < 4) {
+            return props;
+        }
+        let offset = 0;
+        if (offset + 2 <= this.data.length) {
+            const jc = this.view.getInt16(offset, true);
+            switch (jc) {
+                case 0:
+                    props.alignment = 'left';
+                    break;
+                case 1:
+                    props.alignment = 'center';
+                    break;
+                case 2:
+                    props.alignment = 'right';
+                    break;
+            }
+            offset += 2;
+        }
+        if (offset + 2 <= this.data.length) {
+            props.cellSpacing = this.view.getInt16(offset, true);
+            offset += 2;
+        }
+        if (offset + 2 <= this.data.length) {
+            const height = this.view.getInt16(offset, true);
+            props.rowHeight = Math.abs(height);
+            props.rowHeightRule = height < 0 ? 'atLeast' : 'exact';
+            offset += 2;
+        }
+        return props;
+    }
+    parseCellWidths(cellCount) {
+        const widths = [];
+        const widthsOffset = 24;
+        if (widthsOffset + cellCount * 2 > this.data.length) {
+            return widths;
+        }
+        for (let i = 0; i < cellCount; i++) {
+            const width = this.view.getInt16(widthsOffset + i * 2, true);
+            widths.push(width);
+        }
+        return widths;
+    }
+    static isTableRow(papxData) {
+        const view = new DataView(papxData.buffer, papxData.byteOffset, papxData.byteLength);
+        for (let i = 0; i < papxData.length - 2; i++) {
+            const sprm = view.getUint16(i, true);
+            if (sprm === 0x2416) {
+                return true;
+            }
+        }
+        return false;
+    }
+}
+exports.TAPParser = TAPParser;
+//# sourceMappingURL=TableParser.js.map

package/dist/ms-doc/tables/TableParser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"TableParser.js","sourceRoot":"","sources":["../../../src/ms-doc/tables/TableParser.ts"],"names":[],"mappings":";;;AAaA,kDAAmD;AAMnD,MAAa,eAAgB,SAAQ,KAAK;IACxC,YAAY,OAAe;QACzB,KAAK,CAAC,OAAO,CAAC,CAAC;QACf,IAAI,CAAC,IAAI,GAAG,iBAAiB,CAAC;IAChC,CAAC;CACF;AALD,0CAKC;AAiCD,MAAa,WAAW;IACd,IAAI,CAAS;IACb,UAAU,CAAW;IAE7B,YAAY,YAAoB;QAC9B,IAAI,CAAC,IAAI,GAAG,YAAY,CAAC;QAEzB,IAAI,CAAC,UAAU,GAAG,YAAY,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC;IACrD,CAAC;IAKD,UAAU;QACR,MAAM,MAAM,GAAoB,EAAE,CAAC;QACnC,MAAM,QAAQ,GAAG,MAAM,CAAC,YAAY,CAAC,yBAAa,CAAC,IAAI,CAAC,CAAC;QAEzD,IAAI,YAAY,GAAyB,IAAI,CAAC;QAC9C,IAAI,UAAU,GAAG,CAAC,CAAC;QAEnB,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YACnC,MAAM,OAAO,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;YAExC,IAAI,OAAO,EAAE,CAAC;gBAEZ,MAAM,SAAS,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,MAAM,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;gBAEvE,IAAI,CAAC,YAAY,EAAE,CAAC;oBAElB,YAAY,GAAG;wBACb,OAAO,EAAE,UAAU;wBACnB,KAAK,EAAE,UAAU,GAAG,IAAI,CAAC,MAAM;wBAC/B,QAAQ,EAAE,CAAC;wBACX,UAAU,EAAE,SAAS;qBACtB,CAAC;gBACJ,CAAC;qBAAM,CAAC;oBAEN,YAAY,CAAC,KAAK,GAAG,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC;oBAC9C,YAAY,CAAC,QAAQ,EAAE,CAAC;oBACxB,YAAY,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,YAAY,CAAC,UAAU,EAAE,SAAS,CAAC,CAAC;gBACzE,CAAC;YACH,CAAC;iBAAM,IAAI,YAAY,EAAE,CAAC;gBAExB,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;gBAC1B,YAAY,GAAG,IAAI,CAAC;YACtB,CAAC;YAED,UAAU,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;QAChC,CAAC;QAGD,IAAI,YAAY,EAAE,CAAC;YACjB,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QAC5B,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAKD,UAAU,CAAC,QAAuB;QAChC,MAAM,QAAQ,GAAG,MAAM,CAAC,YAAY,CAAC,yBAAa,CAAC,IAAI,CAAC,CAAC;QACzD,MAAM,IAAI,GAAyB,EAAE,CAAC;QAEtC,IAAI,UAAU,GAAG,CAAC,CAAC;QACnB,IAAI,QAAQ,GAAG,CAAC,CAAC;QAEjB,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YAEnC,IAAI,UAAU,IAAI,QAAQ,CAAC,OAAO,IAAI,UAAU,IAAI,QAAQ,CAAC,KAAK,EAAE,CAAC;gBACnE,IAAI,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC;oBAE5B,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;oBACrF,MAAM,KAAK,GAA0B,SAAS,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAC;wBACtE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;wBACjB,QAAQ;wBACR,QAAQ;wBACR,OAAO,EAAE,CAAC;wBACV,OAAO,EAAE,CAAC;qBACX,CAAC,CAAC,CAAC;oBAEJ,IAAI,CAAC,IAAI,CAAC;wBACR,KAAK,EAAE,QAAQ;wBACf,KAAK;wBACL,QAAQ,EAAE,QAAQ,KAAK,CAAC;qBACzB,CAAC,CAAC;oBACH,QAAQ,EAAE,CAAC;gBACb,CAAC;YACH,CAAC;YAED,IAAI,UAAU,GAAG,QAAQ,CAAC,KAAK,EAAE,CAAC;gBAChC,MAAM;YACR,CAAC;YAED,UAAU,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;QAChC,CAAC;QAED,OAAO;YACL,IAAI;YACJ,QAAQ,EAAE,IAAI,CAAC,MAAM;YACrB,WAAW,EAAE,QAAQ,CAAC,UAAU;YAChC,OAAO,EAAE,QAAQ,CAAC,OAAO;YACzB,KAAK,EAAE,QAAQ,CAAC,KAAK;SACtB,CAAC;IACJ,CAAC;IAKD,cAAc;QACZ,MAAM,UAAU,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC;QACrC,OAAO,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;IACnD,CAAC;IAKD,MAAM,CAAC,eAAe,CAAC,IAAY;QACjC,MAAM,QAAQ,GAAG,MAAM,CAAC,YAAY,CAAC,yBAAa,CAAC,IAAI,CAAC,CAAC;QACzD,OAAO,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACjC,CAAC;IAKD,MAAM,CAAC,WAAW,CAAC,IAAY;QAC7B,MAAM,QAAQ,GAAG,MAAM,CAAC,YAAY,CAAC,yBAAa,CAAC,IAAI,CAAC,CAAC;QACzD,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC;QAC5C,MAAM,IAAI,GAAe,EAAE,CAAC;QAE5B,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,IAAI,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC;gBAC5B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;gBACjF,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;YACxC,CAAC;QACH,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;CACF;AA5ID,kCA4IC;AAQD,MAAa,SAAS;IACZ,IAAI,CAAa;IACjB,IAAI,CAAW;IAEvB,YAAY,OAAmB;QAC7B,IAAI,CAAC,IAAI,GAAG,OAAO,CAAC;QACpB,IAAI,CAAC,IAAI,GAAG,IAAI,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IACzF,CAAC;IAKD,KAAK;QACH,MAAM,KAAK,GAAoB,EAAE,CAAC;QAElC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACzB,OAAO,KAAK,CAAC;QACf,CAAC;QAKD,IAAI,MAAM,GAAG,CAAC,CAAC;QAGf,IAAI,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACnC,MAAM,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;YAC5C,QAAQ,EAAE,EAAE,CAAC;gBACX,KAAK,CAAC;oBACJ,KAAK,CAAC,SAAS,GAAG,MAAM,CAAC;oBACzB,MAAM;gBACR,KAAK,CAAC;oBACJ,KAAK,CAAC,SAAS,GAAG,QAAQ,CAAC;oBAC3B,MAAM;gBACR,KAAK,CAAC;oBACJ,KAAK,CAAC,SAAS,GAAG,OAAO,CAAC;oBAC1B,MAAM;YACV,CAAC;YACD,MAAM,IAAI,CAAC,CAAC;QACd,CAAC;QAGD,IAAI,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACnC,KAAK,CAAC,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;YACrD,MAAM,IAAI,CAAC,CAAC;QACd,CAAC;QAGD,IAAI,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACnC,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;YAChD,KAAK,CAAC,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;YACnC,KAAK,CAAC,aAAa,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC;YACvD,MAAM,IAAI,CAAC,CAAC;QACd,CAAC;QAED,OAAO,KAAK,CAAC;IACf,CAAC;IAKD,eAAe,CAAC,SAAiB;QAC/B,MAAM,MAAM,GAAa,EAAE,CAAC;QAI5B,MAAM,YAAY,GAAG,EAAE,CAAC;QAExB,IAAI,YAAY,GAAG,SAAS,GAAG,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACpD,OAAO,MAAM,CAAC;QAChB,CAAC;QAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,YAAY,GAAG,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC;YAC7D,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACrB,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAKD,MAAM,CAAC,UAAU,CAAC,QAAoB;QAGpC,MAAM,IAAI,GAAG,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,EAAE,QAAQ,CAAC,UAAU,EAAE,QAAQ,CAAC,UAAU,CAAC,CAAC;QAErF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;YACrC,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;gBAEpB,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QAED,OAAO,KAAK,CAAC;IACf,CAAC;CACF;AAlGD,8BAkGC"}

package/dist/ms-doc/text/PieceTable.d.ts

@@ -0,0 +1,21 @@
+import { PieceTable, TextRange } from '../types/DocTypes';
+export declare class PieceTableError extends Error {
+    constructor(message: string);
+}
+export declare class PieceTableParser {
+    private data;
+    private view;
+    private wordDocumentStream;
+    constructor(clxData: Uint8Array, wordDocumentStream: Uint8Array);
+    parse(): PieceTable;
+    private parsePcdt;
+    private parsePlcPcd;
+    private parsePcd;
+    extractText(): TextRange[];
+    private extractPieceText;
+    private decodeANSI;
+    private decodeUTF16LE;
+    getFullText(): string;
+    static extractText(clxData: Uint8Array, wordDocumentStream: Uint8Array): TextRange[];
+}
+//# sourceMappingURL=PieceTable.d.ts.map

package/dist/ms-doc/text/PieceTable.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"PieceTable.d.ts","sourceRoot":"","sources":["../../../src/ms-doc/text/PieceTable.ts"],"names":[],"mappings":"AAaA,OAAO,EAAE,UAAU,EAAmB,SAAS,EAAE,MAAM,mBAAmB,CAAC;AAK3E,qBAAa,eAAgB,SAAQ,KAAK;gBAC5B,OAAO,EAAE,MAAM;CAI5B;AAeD,qBAAa,gBAAgB;IAC3B,OAAO,CAAC,IAAI,CAAa;IACzB,OAAO,CAAC,IAAI,CAAW;IACvB,OAAO,CAAC,kBAAkB,CAAa;gBAM3B,OAAO,EAAE,UAAU,EAAE,kBAAkB,EAAE,UAAU;IAS/D,KAAK,IAAI,UAAU;IA2BnB,OAAO,CAAC,SAAS;IAcjB,OAAO,CAAC,WAAW;IAuCnB,OAAO,CAAC,QAAQ;IA2BhB,WAAW,IAAI,SAAS,EAAE;IAmB1B,OAAO,CAAC,gBAAgB;IAiBxB,OAAO,CAAC,UAAU;IAiDlB,OAAO,CAAC,aAAa;IAcrB,WAAW,IAAI,MAAM;IAQrB,MAAM,CAAC,WAAW,CAAC,OAAO,EAAE,UAAU,EAAE,kBAAkB,EAAE,UAAU,GAAG,SAAS,EAAE;CAIrF"}

package/dist/ms-doc/text/PieceTable.js

@@ -0,0 +1,171 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.PieceTableParser = exports.PieceTableError = void 0;
+class PieceTableError extends Error {
+    constructor(message) {
+        super(message);
+        this.name = 'PieceTableError';
+    }
+}
+exports.PieceTableError = PieceTableError;
+const CLX_TYPES = {
+    PRC: 0x01,
+    PCDT: 0x02,
+};
+class PieceTableParser {
+    data;
+    view;
+    wordDocumentStream;
+    constructor(clxData, wordDocumentStream) {
+        this.data = clxData;
+        this.view = new DataView(this.data.buffer, this.data.byteOffset, this.data.byteLength);
+        this.wordDocumentStream = wordDocumentStream;
+    }
+    parse() {
+        let offset = 0;
+        while (offset < this.data.length) {
+            const clxType = this.data[offset];
+            if (clxType === CLX_TYPES.PRC) {
+                offset++;
+                const prcLen = this.view.getInt16(offset, true);
+                offset += 2 + prcLen;
+            }
+            else if (clxType === CLX_TYPES.PCDT) {
+                offset++;
+                return this.parsePcdt(offset);
+            }
+            else {
+                throw new PieceTableError(`Unknown Clx type: 0x${clxType?.toString(16)}`);
+            }
+        }
+        throw new PieceTableError('No Pcdt found in Clx structure');
+    }
+    parsePcdt(offset) {
+        const pcdtLen = this.view.getUint32(offset, true);
+        offset += 4;
+        return this.parsePlcPcd(offset, pcdtLen);
+    }
+    parsePlcPcd(offset, length) {
+        const PCD_SIZE = 8;
+        const numPieces = (length - 4) / 12;
+        if (!Number.isInteger(numPieces) || numPieces < 0) {
+            throw new PieceTableError(`Invalid PlcPcd length: ${length}`);
+        }
+        const cps = [];
+        for (let i = 0; i <= numPieces; i++) {
+            cps.push(this.view.getUint32(offset + i * 4, true));
+        }
+        const pcdOffset = offset + (numPieces + 1) * 4;
+        const pieces = [];
+        for (let i = 0; i < numPieces; i++) {
+            const pcd = this.parsePcd(pcdOffset + i * PCD_SIZE, cps[i] ?? 0, cps[i + 1] ?? 0);
+            pieces.push(pcd);
+        }
+        return { cps, pieces };
+    }
+    parsePcd(offset, cpStart, cpEnd) {
+        const fcRaw = this.view.getUint32(offset + 2, true);
+        const prm = this.view.getUint16(offset + 6, true);
+        const fCompressed = (fcRaw & 0x40000000) !== 0;
+        let fc = fcRaw & 0x3fffffff;
+        if (fCompressed) {
+            fc = fc >> 1;
+        }
+        return {
+            fc,
+            fCompressed,
+            prm,
+            cpStart,
+            cpEnd,
+        };
+    }
+    extractText() {
+        const pieceTable = this.parse();
+        const ranges = [];
+        for (const piece of pieceTable.pieces) {
+            const text = this.extractPieceText(piece);
+            ranges.push({
+                cpStart: piece.cpStart,
+                cpEnd: piece.cpEnd,
+                text,
+            });
+        }
+        return ranges;
+    }
+    extractPieceText(piece) {
+        const charCount = piece.cpEnd - piece.cpStart;
+        if (piece.fCompressed) {
+            const bytes = this.wordDocumentStream.slice(piece.fc, piece.fc + charCount);
+            return this.decodeANSI(bytes);
+        }
+        else {
+            const bytes = this.wordDocumentStream.slice(piece.fc, piece.fc + charCount * 2);
+            return this.decodeUTF16LE(bytes);
+        }
+    }
+    decodeANSI(bytes) {
+        const cp1252Map = {
+            0x80: 0x20ac,
+            0x82: 0x201a,
+            0x83: 0x0192,
+            0x84: 0x201e,
+            0x85: 0x2026,
+            0x86: 0x2020,
+            0x87: 0x2021,
+            0x88: 0x02c6,
+            0x89: 0x2030,
+            0x8a: 0x0160,
+            0x8b: 0x2039,
+            0x8c: 0x0152,
+            0x8e: 0x017d,
+            0x91: 0x2018,
+            0x92: 0x2019,
+            0x93: 0x201c,
+            0x94: 0x201d,
+            0x95: 0x2022,
+            0x96: 0x2013,
+            0x97: 0x2014,
+            0x98: 0x02dc,
+            0x99: 0x2122,
+            0x9a: 0x0161,
+            0x9b: 0x203a,
+            0x9c: 0x0153,
+            0x9e: 0x017e,
+            0x9f: 0x0178,
+        };
+        const chars = [];
+        for (let i = 0; i < bytes.length; i++) {
+            const byte = bytes[i];
+            if (byte === undefined)
+                continue;
+            if (byte >= 0x80 && byte <= 0x9f && cp1252Map[byte] !== undefined) {
+                chars.push(String.fromCharCode(cp1252Map[byte]));
+            }
+            else {
+                chars.push(String.fromCharCode(byte));
+            }
+        }
+        return chars.join('');
+    }
+    decodeUTF16LE(bytes) {
+        const chars = [];
+        for (let i = 0; i < bytes.length - 1; i += 2) {
+            const low = bytes[i];
+            const high = bytes[i + 1];
+            if (low === undefined || high === undefined)
+                break;
+            chars.push(String.fromCharCode(low | (high << 8)));
+        }
+        return chars.join('');
+    }
+    getFullText() {
+        const ranges = this.extractText();
+        return ranges.map((r) => r.text).join('');
+    }
+    static extractText(clxData, wordDocumentStream) {
+        const parser = new PieceTableParser(clxData, wordDocumentStream);
+        return parser.extractText();
+    }
+}
+exports.PieceTableParser = PieceTableParser;
+//# sourceMappingURL=PieceTable.js.map

package/dist/ms-doc/text/PieceTable.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"PieceTable.js","sourceRoot":"","sources":["../../../src/ms-doc/text/PieceTable.ts"],"names":[],"mappings":";;;AAkBA,MAAa,eAAgB,SAAQ,KAAK;IACxC,YAAY,OAAe;QACzB,KAAK,CAAC,OAAO,CAAC,CAAC;QACf,IAAI,CAAC,IAAI,GAAG,iBAAiB,CAAC;IAChC,CAAC;CACF;AALD,0CAKC;AAKD,MAAM,SAAS,GAAG;IAEhB,GAAG,EAAE,IAAI;IAET,IAAI,EAAE,IAAI;CACF,CAAC;AAKX,MAAa,gBAAgB;IACnB,IAAI,CAAa;IACjB,IAAI,CAAW;IACf,kBAAkB,CAAa;IAMvC,YAAY,OAAmB,EAAE,kBAA8B;QAC7D,IAAI,CAAC,IAAI,GAAG,OAAO,CAAC;QACpB,IAAI,CAAC,IAAI,GAAG,IAAI,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;QACvF,IAAI,CAAC,kBAAkB,GAAG,kBAAkB,CAAC;IAC/C,CAAC;IAKD,KAAK;QACH,IAAI,MAAM,GAAG,CAAC,CAAC;QAGf,OAAO,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACjC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAElC,IAAI,OAAO,KAAK,SAAS,CAAC,GAAG,EAAE,CAAC;gBAE9B,MAAM,EAAE,CAAC;gBACT,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;gBAChD,MAAM,IAAI,CAAC,GAAG,MAAM,CAAC;YACvB,CAAC;iBAAM,IAAI,OAAO,KAAK,SAAS,CAAC,IAAI,EAAE,CAAC;gBAEtC,MAAM,EAAE,CAAC;gBACT,OAAO,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;YAChC,CAAC;iBAAM,CAAC;gBACN,MAAM,IAAI,eAAe,CAAC,uBAAuB,OAAO,EAAE,QAAQ,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;YAC5E,CAAC;QACH,CAAC;QAED,MAAM,IAAI,eAAe,CAAC,gCAAgC,CAAC,CAAC;IAC9D,CAAC;IAKO,SAAS,CAAC,MAAc;QAE9B,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;QAClD,MAAM,IAAI,CAAC,CAAC;QAEZ,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC3C,CAAC;IAQO,WAAW,CAAC,MAAc,EAAE,MAAc;QAChD,MAAM,QAAQ,GAAG,CAAC,CAAC;QAMnB,MAAM,SAAS,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC;QAEpC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,SAAS,CAAC,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;YAClD,MAAM,IAAI,eAAe,CAAC,0BAA0B,MAAM,EAAE,CAAC,CAAC;QAChE,CAAC;QAGD,MAAM,GAAG,GAAa,EAAE,CAAC;QACzB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACpC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC;QACtD,CAAC;QAGD,MAAM,SAAS,GAAG,MAAM,GAAG,CAAC,SAAS,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;QAC/C,MAAM,MAAM,GAAsB,EAAE,CAAC;QAErC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,MAAM,GAAG,GAAG,IAAI,CAAC,QAAQ,CAAC,SAAS,GAAG,CAAC,GAAG,QAAQ,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;YAClF,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QACnB,CAAC;QAED,OAAO,EAAE,GAAG,EAAE,MAAM,EAAE,CAAC;IACzB,CAAC;IAUO,QAAQ,CAAC,MAAc,EAAE,OAAe,EAAE,KAAa;QAE7D,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC;QACpD,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC;QAGlD,MAAM,WAAW,GAAG,CAAC,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC;QAI/C,IAAI,EAAE,GAAG,KAAK,GAAG,UAAU,CAAC;QAC5B,IAAI,WAAW,EAAE,CAAC;YAChB,EAAE,GAAG,EAAE,IAAI,CAAC,CAAC;QACf,CAAC;QAED,OAAO;YACL,EAAE;YACF,WAAW;YACX,GAAG;YACH,OAAO;YACP,KAAK;SACN,CAAC;IACJ,CAAC;IAKD,WAAW;QACT,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,EAAE,CAAC;QAChC,MAAM,MAAM,GAAgB,EAAE,CAAC;QAE/B,KAAK,MAAM,KAAK,IAAI,UAAU,CAAC,MAAM,EAAE,CAAC;YACtC,MAAM,IAAI,GAAG,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC,CAAC;YAC1C,MAAM,CAAC,IAAI,CAAC;gBACV,OAAO,EAAE,KAAK,CAAC,OAAO;gBACtB,KAAK,EAAE,KAAK,CAAC,KAAK;gBAClB,IAAI;aACL,CAAC,CAAC;QACL,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAKO,gBAAgB,CAAC,KAAsB;QAC7C,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC;QAE9C,IAAI,KAAK,CAAC,WAAW,EAAE,CAAC;YAEtB,MAAM,KAAK,GAAG,IAAI,CAAC,kBAAkB,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,EAAE,KAAK,CAAC,EAAE,GAAG,SAAS,CAAC,CAAC;YAC5E,OAAO,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;aAAM,CAAC;YAEN,MAAM,KAAK,GAAG,IAAI,CAAC,kBAAkB,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,EAAE,KAAK,CAAC,EAAE,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC;YAChF,OAAO,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC;QACnC,CAAC;IACH,CAAC;IAKO,UAAU,CAAC,KAAiB;QAEl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b,CAAC;QAEF,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACtB,IAAI,IAAI,KAAK,SAAS;gBAAE,SAAS;YAEjC,IAAI,IAAI,IAAI,IAAI,IAAI,IAAI,IAAI,IAAI,IAAI,SAAS,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;gBAClE,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,CAAE,CAAC,CAAC,CAAC;YACpD,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC;YACxC,CAAC;QACH,CAAC;QACD,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACxB,CAAC;IAKO,aAAa,CAAC,KAAiB;QACrC,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YAC7C,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACrB,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAC1B,IAAI,GAAG,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS;gBAAE,MAAM;YACnD,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,GAAG,GAAG,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;QACrD,CAAC;QACD,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACxB,CAAC;IAKD,WAAW;QACT,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QAClC,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAC5C,CAAC;IAKD,MAAM,CAAC,WAAW,CAAC,OAAmB,EAAE,kBAA8B;QACpE,MAAM,MAAM,GAAG,IAAI,gBAAgB,CAAC,OAAO,EAAE,kBAAkB,CAAC,CAAC;QACjE,OAAO,MAAM,CAAC,WAAW,EAAE,CAAC;IAC9B,CAAC;CACF;AA5OD,4CA4OC"}