@usejunior/docx-core 0.0.1 → 0.1.1

package/dist/primitives/reject_changes.js

@@ -0,0 +1,371 @@
+/**
+ * reject_changes — reject all tracked changes in a OOXML document body.
+ *
+ * Produces the document as it was before tracked changes were made by:
+ * - Removing w:ins elements AND their content (insertions undone)
+ * - Unwrapping w:del elements and converting w:delText → w:t (deletions restored)
+ * - Unwrapping w:moveFrom (keep at original position), removing w:moveTo and content
+ * - Restoring original properties from *PrChange records
+ * - Preserving cross-paragraph bookmark boundaries when removing inserted paragraphs
+ * - Stripping paragraph-level revision markers and rsidDel attributes
+ *
+ * Operates on the W3C DOM (`@xmldom/xmldom`).
+ */
+import { OOXML } from './namespaces.js';
+const W_NS = OOXML.W_NS;
+// ── DOM helpers (internal) ──────────────────────────────────────────
+function isW(node, localName) {
+    return (node.nodeType === 1 &&
+        node.namespaceURI === W_NS &&
+        node.localName === localName);
+}
+function isWElement(node) {
+    return node.nodeType === 1 && node.namespaceURI === W_NS;
+}
+function getDepth(node) {
+    let depth = 0;
+    let cur = node.parentNode;
+    while (cur) {
+        depth++;
+        cur = cur.parentNode;
+    }
+    return depth;
+}
+function collectByLocalName(container, localName) {
+    return Array.from(container.getElementsByTagNameNS(W_NS, localName));
+}
+function removeAllByLocalName(container, localName) {
+    const elements = collectByLocalName(container, localName);
+    let count = 0;
+    for (const el of elements) {
+        if (el.parentNode) {
+            el.parentNode.removeChild(el);
+            count++;
+        }
+    }
+    return count;
+}
+function unwrapAllByLocalName(container, localName) {
+    const elements = collectByLocalName(container, localName);
+    // Sort deepest-first to handle nested wrappers correctly
+    elements.sort((a, b) => getDepth(b) - getDepth(a));
+    let count = 0;
+    for (const el of elements) {
+        const parent = el.parentNode;
+        if (!parent)
+            continue;
+        while (el.firstChild) {
+            parent.insertBefore(el.firstChild, el);
+        }
+        parent.removeChild(el);
+        count++;
+    }
+    return count;
+}
+/**
+ * Check if a paragraph has a paragraph-level revision marker.
+ * Pattern: w:p > w:pPr > w:rPr > w:ins (or w:del)
+ */
+function paragraphHasParaMarker(p, markerLocalName) {
+    for (let i = 0; i < p.childNodes.length; i++) {
+        const child = p.childNodes[i];
+        if (!isW(child, 'pPr'))
+            continue;
+        for (let j = 0; j < child.childNodes.length; j++) {
+            const pPrChild = child.childNodes[j];
+            if (!isW(pPrChild, 'rPr'))
+                continue;
+            for (let k = 0; k < pPrChild.childNodes.length; k++) {
+                const rPrChild = pPrChild.childNodes[k];
+                if (isW(rPrChild, markerLocalName))
+                    return true;
+            }
+        }
+    }
+    return false;
+}
+/**
+ * Check if a node (or its descendants) contains any w:r elements.
+ */
+function containsRun(node) {
+    if (isW(node, 'r'))
+        return true;
+    for (let i = 0; i < node.childNodes.length; i++) {
+        if (containsRun(node.childNodes[i]))
+            return true;
+    }
+    return false;
+}
+// Tags that are "inserted" content — removal candidates when rejecting
+const INSERTED_LOCALS = new Set(['ins', 'moveTo']);
+// Tags that are "kept" content when rejecting
+const KEPT_LOCALS = new Set(['del', 'moveFrom']);
+// Range marker tags to ignore
+const RANGE_MARKER_LOCALS = new Set([
+    'moveFromRangeStart', 'moveFromRangeEnd',
+    'moveToRangeStart', 'moveToRangeEnd',
+]);
+/**
+ * Determine if a paragraph's only content lives inside w:ins or w:moveTo
+ * (no w:r outside those wrappers, and no w:del/w:moveFrom siblings).
+ * These paragraphs should be removed when rejecting changes.
+ */
+function paragraphHasOnlyInsertedContent(p) {
+    for (let i = 0; i < p.childNodes.length; i++) {
+        const child = p.childNodes[i];
+        if (child.nodeType !== 1)
+            continue;
+        const el = child;
+        if (el.namespaceURI !== W_NS)
+            continue;
+        const local = el.localName;
+        // If the paragraph has kept content wrappers, it stays
+        if (KEPT_LOCALS.has(local))
+            return false;
+        // Skip pPr, range markers, and inserted wrappers
+        if (local === 'pPr' || RANGE_MARKER_LOCALS.has(local) || INSERTED_LOCALS.has(local))
+            continue;
+        // A bare w:r or any other element that contains runs means live content
+        if (local === 'r' || containsRun(el))
+            return false;
+    }
+    // We passed all children without finding live content — but there must be
+    // at least one inserted wrapper for this to be an "inserted content" paragraph
+    for (let i = 0; i < p.childNodes.length; i++) {
+        const child = p.childNodes[i];
+        if (child.nodeType === 1 && isWElement(child) && INSERTED_LOCALS.has(child.localName)) {
+            return true;
+        }
+    }
+    return false;
+}
+// Property change element local names (all 6 types)
+const PR_CHANGE_LOCALS = [
+    'rPrChange', 'pPrChange', 'sectPrChange',
+    'tblPrChange', 'trPrChange', 'tcPrChange',
+];
+/**
+ * Relocate orphaned bookmarks from a paragraph being removed to an
+ * adjacent kept paragraph. This preserves cross-paragraph bookmark boundaries.
+ */
+function relocateBookmarks(p, paragraphsToRemove) {
+    const parent = p.parentNode;
+    if (!parent)
+        return;
+    // Find an adjacent kept paragraph (prefer next, fallback to previous)
+    let target = null;
+    let sibling = p.nextSibling;
+    while (sibling) {
+        if (sibling.nodeType === 1 && isW(sibling, 'p') && !paragraphsToRemove.has(sibling)) {
+            target = sibling;
+            break;
+        }
+        sibling = sibling.nextSibling;
+    }
+    if (!target) {
+        sibling = p.previousSibling;
+        while (sibling) {
+            if (sibling.nodeType === 1 && isW(sibling, 'p') && !paragraphsToRemove.has(sibling)) {
+                target = sibling;
+                break;
+            }
+            sibling = sibling.previousSibling;
+        }
+    }
+    if (!target)
+        return;
+    // Move bookmarkStart/bookmarkEnd elements that are direct children of the paragraph
+    const toMove = [];
+    for (let i = 0; i < p.childNodes.length; i++) {
+        const child = p.childNodes[i];
+        if (child.nodeType !== 1)
+            continue;
+        const el = child;
+        if (isW(el, 'bookmarkStart') || isW(el, 'bookmarkEnd')) {
+            toMove.push(el);
+        }
+    }
+    // Also check sibling bookmarks (sibling-style: <bookmarkStart/><p/><bookmarkEnd/>)
+    let prev = p.previousSibling;
+    while (prev) {
+        if (prev.nodeType === 1) {
+            const el = prev;
+            if (isW(el, 'bookmarkStart') || isW(el, 'bookmarkEnd')) {
+                toMove.push(el);
+                prev = prev.previousSibling;
+                continue;
+            }
+            break;
+        }
+        prev = prev.previousSibling;
+    }
+    let next = p.nextSibling;
+    while (next) {
+        if (next.nodeType === 1) {
+            const el = next;
+            if (isW(el, 'bookmarkStart') || isW(el, 'bookmarkEnd')) {
+                toMove.push(el);
+                next = next.nextSibling;
+                continue;
+            }
+            break;
+        }
+        next = next.nextSibling;
+    }
+    for (const bm of toMove) {
+        // Insert at the beginning of the target paragraph (after pPr if present)
+        const firstNonPPr = Array.from(target.childNodes).find((n) => !(n.nodeType === 1 && isW(n, 'pPr')));
+        if (bm.parentNode)
+            bm.parentNode.removeChild(bm);
+        target.insertBefore(bm, firstNonPPr ?? null);
+    }
+}
+// ── Public API ──────────────────────────────────────────────────────
+/**
+ * Reject all tracked changes in the document body, restoring the
+ * document to its pre-edit state.
+ *
+ * Mutates the Document in place (same convention as acceptChanges).
+ */
+export function rejectChanges(doc) {
+    const body = doc.getElementsByTagNameNS(W_NS, 'body').item(0);
+    if (!body) {
+        return { insertionsRemoved: 0, deletionsRestored: 0, movesReverted: 0, propertyChangesReverted: 0 };
+    }
+    // Phase A — Identify paragraphs to remove (entirely inserted paragraphs)
+    const paragraphsToRemove = new Set();
+    const allParagraphs = collectByLocalName(body, 'p');
+    for (const p of allParagraphs) {
+        // Paragraph-level insertion marker: w:p > w:pPr > w:rPr > w:ins
+        if (paragraphHasParaMarker(p, 'ins')) {
+            paragraphsToRemove.add(p);
+            continue;
+        }
+        // Paragraphs whose only content is inside w:ins or w:moveTo
+        if (paragraphHasOnlyInsertedContent(p)) {
+            paragraphsToRemove.add(p);
+        }
+    }
+    // Phase B — Preserve cross-paragraph bookmark boundaries
+    for (const p of paragraphsToRemove) {
+        relocateBookmarks(p, paragraphsToRemove);
+    }
+    // Phase C — Remove insertions and move destinations
+    const insertionsRemoved = removeAllByLocalName(body, 'ins');
+    const moveToRemoved = removeAllByLocalName(body, 'moveTo');
+    removeAllByLocalName(body, 'moveToRangeStart');
+    removeAllByLocalName(body, 'moveToRangeEnd');
+    removeAllByLocalName(body, 'moveFromRangeStart');
+    removeAllByLocalName(body, 'moveFromRangeEnd');
+    // Phase D — Unwrap deletions and convert w:delText → w:t
+    const deletionsRestored = unwrapAllByLocalName(body, 'del');
+    // Rename all w:delText elements to w:t so getParagraphText() sees them
+    const delTexts = collectByLocalName(body, 'delText');
+    for (const dt of delTexts) {
+        const parent = dt.parentNode;
+        if (!parent)
+            continue;
+        const t = doc.createElementNS(W_NS, 'w:t');
+        // Copy text content
+        if (dt.textContent) {
+            t.appendChild(doc.createTextNode(dt.textContent));
+        }
+        // Copy xml:space attribute if present
+        const xmlSpace = dt.getAttributeNS('http://www.w3.org/XML/1998/namespace', 'space')
+            ?? dt.getAttribute('xml:space');
+        if (xmlSpace) {
+            t.setAttributeNS('http://www.w3.org/XML/1998/namespace', 'xml:space', xmlSpace);
+        }
+        parent.replaceChild(t, dt);
+    }
+    // Phase E — Unwrap move sources (keep content at original position)
+    const moveFromUnwrapped = unwrapAllByLocalName(body, 'moveFrom');
+    // Phase F — Restore original properties from *PrChange records
+    let propertyChangesReverted = 0;
+    for (const localName of PR_CHANGE_LOCALS) {
+        const changes = collectByLocalName(body, localName);
+        // Sort deepest-first
+        changes.sort((a, b) => getDepth(b) - getDepth(a));
+        for (const change of changes) {
+            const parentProp = change.parentNode;
+            if (!parentProp)
+                continue;
+            const grandParent = parentProp.parentNode;
+            if (!grandParent)
+                continue;
+            // The *PrChange element contains the original properties.
+            // Extract the original property element (e.g. rPr inside rPrChange).
+            // The expected mapping:
+            //   rPrChange → child rPr = original run properties
+            //   pPrChange → child pPr = original paragraph properties
+            //   etc.
+            const expectedChildLocal = localName.replace('Change', '');
+            let originalProps = null;
+            for (let i = 0; i < change.childNodes.length; i++) {
+                const child = change.childNodes[i];
+                if (child.nodeType === 1 && isW(child, expectedChildLocal)) {
+                    originalProps = child;
+                    break;
+                }
+            }
+            if (originalProps) {
+                // Replace the current property element with the original
+                const restored = originalProps.cloneNode(true);
+                grandParent.replaceChild(restored, parentProp);
+            }
+            else {
+                // Original props were empty — remove the parent property element entirely
+                grandParent.removeChild(parentProp);
+            }
+            propertyChangesReverted++;
+        }
+    }
+    // Phase G — Cleanup
+    // Strip paragraph-level revision markers from w:pPr/w:rPr
+    for (const p of collectByLocalName(body, 'p')) {
+        for (let i = 0; i < p.childNodes.length; i++) {
+            const child = p.childNodes[i];
+            if (!isW(child, 'pPr'))
+                continue;
+            for (let j = 0; j < child.childNodes.length; j++) {
+                const pPrChild = child.childNodes[j];
+                if (!isW(pPrChild, 'rPr'))
+                    continue;
+                const toRemove = [];
+                for (let k = 0; k < pPrChild.childNodes.length; k++) {
+                    const rPrChild = pPrChild.childNodes[k];
+                    if (isW(rPrChild, 'ins') || isW(rPrChild, 'del')) {
+                        toRemove.push(rPrChild);
+                    }
+                }
+                for (const el of toRemove) {
+                    pPrChild.removeChild(el);
+                }
+            }
+        }
+    }
+    // Remove paragraphs collected in Phase A
+    for (const p of paragraphsToRemove) {
+        if (p.parentNode) {
+            p.parentNode.removeChild(p);
+        }
+    }
+    // Strip w:rsidDel attributes on remaining elements
+    const allElements = body.getElementsByTagNameNS(W_NS, '*');
+    for (let i = 0; i < allElements.length; i++) {
+        const el = allElements[i];
+        if (el.hasAttributeNS(W_NS, 'rsidDel')) {
+            el.removeAttributeNS(W_NS, 'rsidDel');
+        }
+        if (el.hasAttribute('w:rsidDel')) {
+            el.removeAttribute('w:rsidDel');
+        }
+    }
+    return {
+        insertionsRemoved,
+        deletionsRestored,
+        movesReverted: moveFromUnwrapped + moveToRemoved,
+        propertyChangesReverted,
+    };
+}
+//# sourceMappingURL=reject_changes.js.map

package/dist/primitives/reject_changes.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"reject_changes.js","sourceRoot":"","sources":["../../src/primitives/reject_changes.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,OAAO,EAAE,KAAK,EAAE,MAAM,iBAAiB,CAAC;AAExC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;AASxB,uEAAuE;AAEvE,SAAS,GAAG,CAAC,IAAU,EAAE,SAAiB;IACxC,OAAO,CACL,IAAI,CAAC,QAAQ,KAAK,CAAC;QAClB,IAAgB,CAAC,YAAY,KAAK,IAAI;QACtC,IAAgB,CAAC,SAAS,KAAK,SAAS,CAC1C,CAAC;AACJ,CAAC;AAED,SAAS,UAAU,CAAC,IAAU;IAC5B,OAAO,IAAI,CAAC,QAAQ,KAAK,CAAC,IAAK,IAAgB,CAAC,YAAY,KAAK,IAAI,CAAC;AACxE,CAAC;AAED,SAAS,QAAQ,CAAC,IAAU;IAC1B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,GAAG,GAAgB,IAAI,CAAC,UAAU,CAAC;IACvC,OAAO,GAAG,EAAE,CAAC;QACX,KAAK,EAAE,CAAC;QACR,GAAG,GAAG,GAAG,CAAC,UAAU,CAAC;IACvB,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,kBAAkB,CAAC,SAA6B,EAAE,SAAiB;IAC1E,OAAO,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,sBAAsB,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC,CAAC;AACvE,CAAC;AAED,SAAS,oBAAoB,CAAC,SAA6B,EAAE,SAAiB;IAC5E,MAAM,QAAQ,GAAG,kBAAkB,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC;IAC1D,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;QAC1B,IAAI,EAAE,CAAC,UAAU,EAAE,CAAC;YAClB,EAAE,CAAC,UAAU,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC;YAC9B,KAAK,EAAE,CAAC;QACV,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,oBAAoB,CAAC,SAA6B,EAAE,SAAiB;IAC5E,MAAM,QAAQ,GAAG,kBAAkB,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC;IAC1D,yDAAyD;IACzD,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC;IACnD,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;QAC1B,MAAM,MAAM,GAAG,EAAE,CAAC,UAAU,CAAC;QAC7B,IAAI,CAAC,MAAM;YAAE,SAAS;QACtB,OAAO,EAAE,CAAC,UAAU,EAAE,CAAC;YACrB,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;QACzC,CAAC;QACD,MAAM,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC;QACvB,KAAK,EAAE,CAAC;IACV,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;GAGG;AACH,SAAS,sBAAsB,CAAC,CAAU,EAAE,eAAuB;IACjE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7C,MAAM,KAAK,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;QAC/B,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC;YAAE,SAAS;QACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACjD,MAAM,QAAQ,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;YACtC,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC;gBAAE,SAAS;YACpC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACpD,MAAM,QAAQ,GAAG,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;gBACzC,IAAI,GAAG,CAAC,QAAQ,EAAE,eAAe,CAAC;oBAAE,OAAO,IAAI,CAAC;YAClD,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,IAAU;IAC7B,IAAI,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAChC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAChD,IAAI,WAAW,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;YAAE,OAAO,IAAI,CAAC;IACpD,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,uEAAuE;AACvE,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC;AACnD,8CAA8C;AAC9C,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC,CAAC;AACjD,8BAA8B;AAC9B,MAAM,mBAAmB,GAAG,IAAI,GAAG,CAAC;IAClC,oBAAoB,EAAE,kBAAkB;IACxC,kBAAkB,EAAE,gBAAgB;CACrC,CAAC,CAAC;AAEH;;;;GAIG;AACH,SAAS,+BAA+B,CAAC,CAAU;IACjD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7C,MAAM,KAAK,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;QAC/B,IAAI,KAAK,CAAC,QAAQ,KAAK,CAAC;YAAE,SAAS;QACnC,MAAM,EAAE,GAAG,KAAgB,CAAC;QAC5B,IAAI,EAAE,CAAC,YAAY,KAAK,IAAI;YAAE,SAAS;QACvC,MAAM,KAAK,GAAG,EAAE,CAAC,SAAS,CAAC;QAE3B,uDAAuD;QACvD,IAAI,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC;YAAE,OAAO,KAAK,CAAC;QAEzC,iDAAiD;QACjD,IAAI,KAAK,KAAK,KAAK,IAAI,mBAAmB,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,eAAe,CAAC,GAAG,CAAC,KAAK,CAAC;YAAE,SAAS;QAE9F,wEAAwE;QACxE,IAAI,KAAK,KAAK,GAAG,IAAI,WAAW,CAAC,EAAE,CAAC;YAAE,OAAO,KAAK,CAAC;IACrD,CAAC;IAED,0EAA0E;IAC1E,+EAA+E;IAC/E,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7C,MAAM,KAAK,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;QAC/B,IAAI,KAAK,CAAC,QAAQ,KAAK,CAAC,IAAI,UAAU,CAAC,KAAK,CAAC,IAAI,eAAe,CAAC,GAAG,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;YACtF,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,oDAAoD;AACpD,MAAM,gBAAgB,GAAG;IACvB,WAAW,EAAE,WAAW,EAAE,cAAc;IACxC,aAAa,EAAE,YAAY,EAAE,YAAY;CAC1C,CAAC;AAEF;;;GAGG;AACH,SAAS,iBAAiB,CAAC,CAAU,EAAE,kBAAgC;IACrE,MAAM,MAAM,GAAG,CAAC,CAAC,UAAU,CAAC;IAC5B,IAAI,CAAC,MAAM;QAAE,OAAO;IAEpB,sEAAsE;IACtE,IAAI,MAAM,GAAmB,IAAI,CAAC;IAClC,IAAI,OAAO,GAAgB,CAAC,CAAC,WAAW,CAAC;IACzC,OAAO,OAAO,EAAE,CAAC;QACf,IAAI,OAAO,CAAC,QAAQ,KAAK,CAAC,IAAI,GAAG,CAAC,OAAkB,EAAE,GAAG,CAAC,IAAI,CAAC,kBAAkB,CAAC,GAAG,CAAC,OAAkB,CAAC,EAAE,CAAC;YAC1G,MAAM,GAAG,OAAkB,CAAC;YAC5B,MAAM;QACR,CAAC;QACD,OAAO,GAAG,OAAO,CAAC,WAAW,CAAC;IAChC,CAAC;IACD,IAAI,CAAC,MAAM,EAAE,CAAC;QACZ,OAAO,GAAG,CAAC,CAAC,eAAe,CAAC;QAC5B,OAAO,OAAO,EAAE,CAAC;YACf,IAAI,OAAO,CAAC,QAAQ,KAAK,CAAC,IAAI,GAAG,CAAC,OAAkB,EAAE,GAAG,CAAC,IAAI,CAAC,kBAAkB,CAAC,GAAG,CAAC,OAAkB,CAAC,EAAE,CAAC;gBAC1G,MAAM,GAAG,OAAkB,CAAC;gBAC5B,MAAM;YACR,CAAC;YACD,OAAO,GAAG,OAAO,CAAC,eAAe,CAAC;QACpC,CAAC;IACH,CAAC;IACD,IAAI,CAAC,MAAM;QAAE,OAAO;IAEpB,oFAAoF;IACpF,MAAM,MAAM,GAAc,EAAE,CAAC;IAC7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7C,MAAM,KAAK,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;QAC/B,IAAI,KAAK,CAAC,QAAQ,KAAK,CAAC;YAAE,SAAS;QACnC,MAAM,EAAE,GAAG,KAAgB,CAAC;QAC5B,IAAI,GAAG,CAAC,EAAE,EAAE,eAAe,CAAC,IAAI,GAAG,CAAC,EAAE,EAAE,aAAa,CAAC,EAAE,CAAC;YACvD,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAClB,CAAC;IACH,CAAC;IAED,mFAAmF;IACnF,IAAI,IAAI,GAAgB,CAAC,CAAC,eAAe,CAAC;IAC1C,OAAO,IAAI,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,EAAE,GAAG,IAAe,CAAC;YAC3B,IAAI,GAAG,CAAC,EAAE,EAAE,eAAe,CAAC,IAAI,GAAG,CAAC,EAAE,EAAE,aAAa,CAAC,EAAE,CAAC;gBACvD,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;gBAChB,IAAI,GAAG,IAAI,CAAC,eAAe,CAAC;gBAC5B,SAAS;YACX,CAAC;YACD,MAAM;QACR,CAAC;QACD,IAAI,GAAG,IAAI,CAAC,eAAe,CAAC;IAC9B,CAAC;IACD,IAAI,IAAI,GAAgB,CAAC,CAAC,WAAW,CAAC;IACtC,OAAO,IAAI,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,EAAE,GAAG,IAAe,CAAC;YAC3B,IAAI,GAAG,CAAC,EAAE,EAAE,eAAe,CAAC,IAAI,GAAG,CAAC,EAAE,EAAE,aAAa,CAAC,EAAE,CAAC;gBACvD,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;gBAChB,IAAI,GAAG,IAAI,CAAC,WAAW,CAAC;gBACxB,SAAS;YACX,CAAC;YACD,MAAM;QACR,CAAC;QACD,IAAI,GAAG,IAAI,CAAC,WAAW,CAAC;IAC1B,CAAC;IAED,KAAK,MAAM,EAAE,IAAI,MAAM,EAAE,CAAC;QACxB,yEAAyE;QACzE,MAAM,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,IAAI,CACpD,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,KAAK,CAAC,IAAI,GAAG,CAAC,CAAY,EAAE,KAAK,CAAC,CAAC,CACvD,CAAC;QACF,IAAI,EAAE,CAAC,UAAU;YAAE,EAAE,CAAC,UAAU,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC;QACjD,MAAM,CAAC,YAAY,CAAC,EAAE,EAAE,WAAW,IAAI,IAAI,CAAC,CAAC;IAC/C,CAAC;AACH,CAAC;AAED,uEAAuE;AAEvE;;;;;GAKG;AACH,MAAM,UAAU,aAAa,CAAC,GAAa;IACzC,MAAM,IAAI,GAAG,GAAG,CAAC,sBAAsB,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC9D,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,EAAE,iBAAiB,EAAE,CAAC,EAAE,iBAAiB,EAAE,CAAC,EAAE,aAAa,EAAE,CAAC,EAAE,uBAAuB,EAAE,CAAC,EAAE,CAAC;IACtG,CAAC;IAED,yEAAyE;IACzE,MAAM,kBAAkB,GAAG,IAAI,GAAG,EAAW,CAAC;IAC9C,MAAM,aAAa,GAAG,kBAAkB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAEpD,KAAK,MAAM,CAAC,IAAI,aAAa,EAAE,CAAC;QAC9B,gEAAgE;QAChE,IAAI,sBAAsB,CAAC,CAAC,EAAE,KAAK,CAAC,EAAE,CAAC;YACrC,kBAAkB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YAC1B,SAAS;QACX,CAAC;QACD,4DAA4D;QAC5D,IAAI,+BAA+B,CAAC,CAAC,CAAC,EAAE,CAAC;YACvC,kBAAkB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QAC5B,CAAC;IACH,CAAC;IAED,yDAAyD;IACzD,KAAK,MAAM,CAAC,IAAI,kBAAkB,EAAE,CAAC;QACnC,iBAAiB,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAC;IAC3C,CAAC;IAED,oDAAoD;IACpD,MAAM,iBAAiB,GAAG,oBAAoB,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC5D,MAAM,aAAa,GAAG,oBAAoB,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;IAC3D,oBAAoB,CAAC,IAAI,EAAE,kBAAkB,CAAC,CAAC;IAC/C,oBAAoB,CAAC,IAAI,EAAE,gBAAgB,CAAC,CAAC;IAC7C,oBAAoB,CAAC,IAAI,EAAE,oBAAoB,CAAC,CAAC;IACjD,oBAAoB,CAAC,IAAI,EAAE,kBAAkB,CAAC,CAAC;IAE/C,yDAAyD;IACzD,MAAM,iBAAiB,GAAG,oBAAoB,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAE5D,uEAAuE;IACvE,MAAM,QAAQ,GAAG,kBAAkB,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;IACrD,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;QAC1B,MAAM,MAAM,GAAG,EAAE,CAAC,UAAU,CAAC;QAC7B,IAAI,CAAC,MAAM;YAAE,SAAS;QACtB,MAAM,CAAC,GAAG,GAAG,CAAC,eAAe,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC3C,oBAAoB;QACpB,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YACnB,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,cAAc,CAAC,EAAE,CAAC,WAAW,CAAC,CAAC,CAAC;QACpD,CAAC;QACD,sCAAsC;QACtC,MAAM,QAAQ,GAAG,EAAE,CAAC,cAAc,CAAC,sCAAsC,EAAE,OAAO,CAAC;eAC9E,EAAE,CAAC,YAAY,CAAC,WAAW,CAAC,CAAC;QAClC,IAAI,QAAQ,EAAE,CAAC;YACb,CAAC,CAAC,cAAc,CAAC,sCAAsC,EAAE,WAAW,EAAE,QAAQ,CAAC,CAAC;QAClF,CAAC;QACD,MAAM,CAAC,YAAY,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAC7B,CAAC;IAED,oEAAoE;IACpE,MAAM,iBAAiB,GAAG,oBAAoB,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;IAEjE,+DAA+D;IAC/D,IAAI,uBAAuB,GAAG,CAAC,CAAC;IAChC,KAAK,MAAM,SAAS,IAAI,gBAAgB,EAAE,CAAC;QACzC,MAAM,OAAO,GAAG,kBAAkB,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;QACpD,qBAAqB;QACrB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC;QAClD,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,MAAM,UAAU,GAAG,MAAM,CAAC,UAA4B,CAAC;YACvD,IAAI,CAAC,UAAU;gBAAE,SAAS;YAC1B,MAAM,WAAW,GAAG,UAAU,CAAC,UAAU,CAAC;YAC1C,IAAI,CAAC,WAAW;gBAAE,SAAS;YAE3B,0DAA0D;YAC1D,qEAAqE;YACrE,wBAAwB;YACxB,oDAAoD;YACpD,0DAA0D;YAC1D,SAAS;YACT,MAAM,kBAAkB,GAAG,SAAS,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;YAC3D,IAAI,aAAa,GAAmB,IAAI,CAAC;YACzC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAClD,MAAM,KAAK,GAAG,MAAM,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;gBACpC,IAAI,KAAK,CAAC,QAAQ,KAAK,CAAC,IAAI,GAAG,CAAC,KAAgB,EAAE,kBAAkB,CAAC,EAAE,CAAC;oBACtE,aAAa,GAAG,KAAgB,CAAC;oBACjC,MAAM;gBACR,CAAC;YACH,CAAC;YAED,IAAI,aAAa,EAAE,CAAC;gBAClB,yDAAyD;gBACzD,MAAM,QAAQ,GAAG,aAAa,CAAC,SAAS,CAAC,IAAI,CAAY,CAAC;gBAC1D,WAAW,CAAC,YAAY,CAAC,QAAQ,EAAE,UAAU,CAAC,CAAC;YACjD,CAAC;iBAAM,CAAC;gBACN,0EAA0E;gBAC1E,WAAW,CAAC,WAAW,CAAC,UAAU,CAAC,CAAC;YACtC,CAAC;YACD,uBAAuB,EAAE,CAAC;QAC5B,CAAC;IACH,CAAC;IAED,oBAAoB;IACpB,0DAA0D;IAC1D,KAAK,MAAM,CAAC,IAAI,kBAAkB,CAAC,IAAI,EAAE,GAAG,CAAC,EAAE,CAAC;QAC9C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,KAAK,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;YAC/B,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC;gBAAE,SAAS;YACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACjD,MAAM,QAAQ,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;gBACtC,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC;oBAAE,SAAS;gBACpC,MAAM,QAAQ,GAAc,EAAE,CAAC;gBAC/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBACpD,MAAM,QAAQ,GAAG,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAE,CAAC;oBACzC,IAAI,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC,EAAE,CAAC;wBACjD,QAAQ,CAAC,IAAI,CAAC,QAAmB,CAAC,CAAC;oBACrC,CAAC;gBACH,CAAC;gBACD,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;oBAC1B,QAAQ,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC;gBAC3B,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,yCAAyC;IACzC,KAAK,MAAM,CAAC,IAAI,kBAAkB,EAAE,CAAC;QACnC,IAAI,CAAC,CAAC,UAAU,EAAE,CAAC;YACjB,CAAC,CAAC,UAAU,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC;QAC9B,CAAC;IACH,CAAC;IAED,mDAAmD;IACnD,MAAM,WAAW,GAAG,IAAI,CAAC,sBAAsB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAC3D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC5C,MAAM,EAAE,GAAG,WAAW,CAAC,CAAC,CAAE,CAAC;QAC3B,IAAI,EAAE,CAAC,cAAc,CAAC,IAAI,EAAE,SAAS,CAAC,EAAE,CAAC;YACvC,EAAE,CAAC,iBAAiB,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;QACxC,CAAC;QACD,IAAI,EAAE,CAAC,YAAY,CAAC,WAAW,CAAC,EAAE,CAAC;YACjC,EAAE,CAAC,eAAe,CAAC,WAAW,CAAC,CAAC;QAClC,CAAC;IACH,CAAC;IAED,OAAO;QACL,iBAAiB;QACjB,iBAAiB;QACjB,aAAa,EAAE,iBAAiB,GAAG,aAAa;QAChD,uBAAuB;KACxB,CAAC;AACJ,CAAC"}

package/dist/primitives/relationships.d.ts

@@ -0,0 +1,7 @@
+export type RelsMap = Map<string, string>;
+/**
+ * Parse a document.xml.rels DOM and return a Map<rId, targetUrl> for external hyperlinks only.
+ * Returns an empty map when the rels document is null (e.g. file missing from the DOCX archive).
+ */
+export declare function parseDocumentRels(relsDoc: Document | null): RelsMap;
+//# sourceMappingURL=relationships.d.ts.map

package/dist/primitives/relationships.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"relationships.d.ts","sourceRoot":"","sources":["../../src/primitives/relationships.ts"],"names":[],"mappings":"AAIA,MAAM,MAAM,OAAO,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;AAE1C;;;GAGG;AACH,wBAAgB,iBAAiB,CAAC,OAAO,EAAE,QAAQ,GAAG,IAAI,GAAG,OAAO,CAiBnE"}

package/dist/primitives/relationships.js

@@ -0,0 +1,24 @@
+// Parser for word/_rels/document.xml.rels — extracts external hyperlink relationships.
+import { OOXML } from './namespaces.js';
+/**
+ * Parse a document.xml.rels DOM and return a Map<rId, targetUrl> for external hyperlinks only.
+ * Returns an empty map when the rels document is null (e.g. file missing from the DOCX archive).
+ */
+export function parseDocumentRels(relsDoc) {
+    const map = new Map();
+    if (!relsDoc)
+        return map;
+    const relationships = relsDoc.getElementsByTagName('Relationship');
+    for (let i = 0; i < relationships.length; i++) {
+        const rel = relationships.item(i);
+        const type = rel.getAttribute('Type');
+        const targetMode = rel.getAttribute('TargetMode');
+        const id = rel.getAttribute('Id');
+        const target = rel.getAttribute('Target');
+        if (type === OOXML.HYPERLINK_REL_TYPE && targetMode === 'External' && id && target) {
+            map.set(id, target);
+        }
+    }
+    return map;
+}
+//# sourceMappingURL=relationships.js.map

package/dist/primitives/relationships.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"relationships.js","sourceRoot":"","sources":["../../src/primitives/relationships.ts"],"names":[],"mappings":"AAAA,uFAAuF;AAEvF,OAAO,EAAE,KAAK,EAAE,MAAM,iBAAiB,CAAC;AAIxC;;;GAGG;AACH,MAAM,UAAU,iBAAiB,CAAC,OAAwB;IACxD,MAAM,GAAG,GAAY,IAAI,GAAG,EAAE,CAAC;IAC/B,IAAI,CAAC,OAAO;QAAE,OAAO,GAAG,CAAC;IAEzB,MAAM,aAAa,GAAG,OAAO,CAAC,oBAAoB,CAAC,cAAc,CAAC,CAAC;IACnE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;QACnC,MAAM,IAAI,GAAG,GAAG,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;QACtC,MAAM,UAAU,GAAG,GAAG,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC;QAClD,MAAM,EAAE,GAAG,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;QAClC,MAAM,MAAM,GAAG,GAAG,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;QAE1C,IAAI,IAAI,KAAK,KAAK,CAAC,kBAAkB,IAAI,UAAU,KAAK,UAAU,IAAI,EAAE,IAAI,MAAM,EAAE,CAAC;YACnF,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/primitives/semantic_tags.d.ts

@@ -0,0 +1,32 @@
+export declare const DEFINITION_TAG = "definition";
+export declare const HIGHLIGHT_TAG = "highlighting";
+export declare function hasDefinitionTags(text: string): boolean;
+export declare function hasHighlightTags(text: string): boolean;
+export declare function emitDefinitionTagsFromString(text: string): string;
+export declare function findInlineDefinitionSpan(text: string): {
+    term: string;
+    term_start: number;
+    term_end: number;
+} | null;
+export declare function stripDefinitionTags(text: string): string;
+export declare function stripHighlightTags(text: string): string;
+export declare function hasFormattingTags(text: string): boolean;
+export declare function stripFormattingTags(text: string): string;
+export declare function hasHyperlinkTags(text: string): boolean;
+export declare function stripHyperlinkTags(text: string): string;
+/**
+ * Detect definition spans on plain text, returning character offset ranges
+ * and the defined term. Used by the formatting tag emitter to interleave
+ * <definition> tags with formatting tags in a single consistent pass.
+ *
+ * Returns spans where:
+ * - `start` = index of the opening quote character
+ * - `end` = index one past the closing quote character
+ * - `term` = the defined term text (excluding quotes)
+ */
+export declare function detectDefinitionSpans(text: string): Array<{
+    start: number;
+    end: number;
+    term: string;
+}>;
+//# sourceMappingURL=semantic_tags.d.ts.map

package/dist/primitives/semantic_tags.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"semantic_tags.d.ts","sourceRoot":"","sources":["../../src/primitives/semantic_tags.ts"],"names":[],"mappings":"AASA,eAAO,MAAM,cAAc,eAAe,CAAC;AAC3C,eAAO,MAAM,aAAa,iBAAiB,CAAC;AA0B5C,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEvD;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEtD;AAED,wBAAgB,4BAA4B,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CA2BjE;AAED,wBAAgB,wBAAwB,CAAC,IAAI,EAAE,MAAM,GAAG;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,CAAC;IAAC,QAAQ,EAAE,MAAM,CAAA;CAAE,GAAG,IAAI,CAUpH;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAKxD;AAED,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAIvD;AAQD,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEvD;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAIxD;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEtD;AAED,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAEvD;AAED;;;;;;;;;GASG;AACH,wBAAgB,qBAAqB,CAAC,IAAI,EAAE,MAAM,GAAG,KAAK,CAAC;IAAE,KAAK,EAAE,MAAM,CAAC;IAAC,GAAG,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC,CAmBvG"}

package/dist/primitives/semantic_tags.js

@@ -0,0 +1,139 @@
+// Semantic tag emission + stripping for Safe-Docx TS.
+//
+// Parity goals (Python):
+// - Emit <definition> tags for explicit definitions, absorbing quotes:
+//     Input:  "Company" means ...
+//     Output: <definition>Company</definition> means ...
+//
+// Headers are represented via a dedicated column (not inline tags) in TOON output.
+export const DEFINITION_TAG = 'definition';
+export const HIGHLIGHT_TAG = 'highlighting';
+// Quote variants (subset of Python workflows/shared/regex/definitions.py).
+// Includes ASCII and common smart quotes.
+const QUOTES = `"'\u201c\u201d\u2018\u2019\u00ab\u00bb\u2039\u203a`;
+const QUOTE_CHARS_CLASS = escapeForCharClass(QUOTES);
+const QUOTE_CLASS = `[${QUOTE_CHARS_CLASS}]`;
+// Port of DEFINITION_VERB_PATTERN (simplified but compatible with the Python export).
+const DEF_VERB = '(?:' +
+    '(?:shall\\s+)?means?|' +
+    '(?:shall\\s+have|has|have)\\s+the\\s+meaning(?:\\s+(?:set\\s+forth|given|ascribed)\\s+in)?|' +
+    '(?:is|are)\\s+defined\\s+as|' +
+    'refers?\\s+to|' +
+    '(?:shall\\s+)?mean\\s+and\\s+include' +
+    ')';
+const INLINE_DEFINITION_RE = new RegExp(`${QUOTE_CLASS}([^${QUOTE_CHARS_CLASS}]+)${QUOTE_CLASS}\\s+${DEF_VERB}`, 'gi');
+const INLINE_DEFINITION_RE_ONE = new RegExp(`${QUOTE_CLASS}([^${QUOTE_CHARS_CLASS}]+)${QUOTE_CLASS}\\s+${DEF_VERB}`, 'i');
+function escapeForCharClass(s) {
+    // Escape characters that have meaning in a regex character class.
+    return s.replace(/[-\\\]^]/g, '\\$&');
+}
+export function hasDefinitionTags(text) {
+    return text.includes(`<${DEFINITION_TAG}>`) || text.includes(`</${DEFINITION_TAG}>`);
+}
+export function hasHighlightTags(text) {
+    return text.includes(`<${HIGHLIGHT_TAG}>`) || text.includes(`</${HIGHLIGHT_TAG}>`);
+}
+export function emitDefinitionTagsFromString(text) {
+    if (!text)
+        return text;
+    const matches = Array.from(text.matchAll(INLINE_DEFINITION_RE));
+    if (matches.length === 0)
+        return text;
+    let tagged = text;
+    // Replace from end to start to preserve offsets.
+    for (let i = matches.length - 1; i >= 0; i--) {
+        const m = matches[i];
+        const term = m[1] ?? '';
+        if (!term)
+            continue;
+        const matchStart = m.index ?? -1;
+        if (matchStart < 0)
+            continue;
+        // Find the opening quote at matchStart, and the closing quote right after the term.
+        // We want to replace the quoted term including both quotes.
+        const openQuoteIdx = matchStart;
+        const closeQuoteIdx = openQuoteIdx + 1 + term.length; // quote + term
+        if (closeQuoteIdx >= tagged.length)
+            continue;
+        const before = tagged.slice(0, openQuoteIdx);
+        const after = tagged.slice(closeQuoteIdx + 1); // +1 for closing quote
+        tagged = `${before}<${DEFINITION_TAG}>${term}</${DEFINITION_TAG}>${after}`;
+    }
+    return tagged;
+}
+export function findInlineDefinitionSpan(text) {
+    // Find the first explicit definition in text and return the defined-term span (excluding quotes).
+    const m = INLINE_DEFINITION_RE_ONE.exec(text);
+    if (!m || m.index == null)
+        return null;
+    const term = m[1] ?? '';
+    if (!term)
+        return null;
+    const openQuoteIdx = m.index;
+    const termStart = openQuoteIdx + 1;
+    const termEnd = termStart + term.length;
+    return { term, term_start: termStart, term_end: termEnd };
+}
+export function stripDefinitionTags(text) {
+    // SINGLE SOURCE OF TRUTH for stripping: replace tags with quotes.
+    return text
+        .replaceAll(new RegExp(`<${DEFINITION_TAG}>`, 'g'), '"')
+        .replaceAll(new RegExp(`</${DEFINITION_TAG}>`, 'g'), '"');
+}
+export function stripHighlightTags(text) {
+    return text
+        .replaceAll(new RegExp(`<${HIGHLIGHT_TAG}>`, 'g'), '')
+        .replaceAll(new RegExp(`</${HIGHLIGHT_TAG}>`, 'g'), '');
+}
+// ── Formatting tag helpers ───────────────────────────────────────────
+const FORMATTING_TAG_RE = /<\/?[biu]>/g;
+const HYPERLINK_OPEN_RE = /<a\s+href="[^"]*">/g;
+const HYPERLINK_CLOSE_RE = /<\/a>/g;
+export function hasFormattingTags(text) {
+    return FORMATTING_TAG_RE.test(text);
+}
+export function stripFormattingTags(text) {
+    // Reset lastIndex since these are global regexes.
+    FORMATTING_TAG_RE.lastIndex = 0;
+    return text.replace(FORMATTING_TAG_RE, '');
+}
+export function hasHyperlinkTags(text) {
+    return text.includes('<a ') || text.includes('</a>');
+}
+export function stripHyperlinkTags(text) {
+    return text.replace(HYPERLINK_OPEN_RE, '').replace(HYPERLINK_CLOSE_RE, '');
+}
+/**
+ * Detect definition spans on plain text, returning character offset ranges
+ * and the defined term. Used by the formatting tag emitter to interleave
+ * <definition> tags with formatting tags in a single consistent pass.
+ *
+ * Returns spans where:
+ * - `start` = index of the opening quote character
+ * - `end` = index one past the closing quote character
+ * - `term` = the defined term text (excluding quotes)
+ */
+export function detectDefinitionSpans(text) {
+    if (!text)
+        return [];
+    // Reset global regex state.
+    INLINE_DEFINITION_RE.lastIndex = 0;
+    const matches = Array.from(text.matchAll(INLINE_DEFINITION_RE));
+    const spans = [];
+    for (const m of matches) {
+        const term = m[1] ?? '';
+        if (!term)
+            continue;
+        const matchStart = m.index ?? -1;
+        if (matchStart < 0)
+            continue;
+        const openQuoteIdx = matchStart;
+        const termEnd = openQuoteIdx + 1 + term.length;
+        // Verify close quote exists.
+        if (termEnd < text.length) {
+            spans.push({ start: openQuoteIdx, end: termEnd + 1, term });
+        }
+    }
+    return spans;
+}
+//# sourceMappingURL=semantic_tags.js.map

package/dist/primitives/semantic_tags.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"semantic_tags.js","sourceRoot":"","sources":["../../src/primitives/semantic_tags.ts"],"names":[],"mappings":"AAAA,sDAAsD;AACtD,EAAE;AACF,yBAAyB;AACzB,uEAAuE;AACvE,kCAAkC;AAClC,yDAAyD;AACzD,EAAE;AACF,mFAAmF;AAEnF,MAAM,CAAC,MAAM,cAAc,GAAG,YAAY,CAAC;AAC3C,MAAM,CAAC,MAAM,aAAa,GAAG,cAAc,CAAC;AAE5C,2EAA2E;AAC3E,0CAA0C;AAC1C,MAAM,MAAM,GAAG,oDAAoD,CAAC;AACpE,MAAM,iBAAiB,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAC;AACrD,MAAM,WAAW,GAAG,IAAI,iBAAiB,GAAG,CAAC;AAE7C,sFAAsF;AACtF,MAAM,QAAQ,GACZ,KAAK;IACL,uBAAuB;IACvB,6FAA6F;IAC7F,8BAA8B;IAC9B,gBAAgB;IAChB,sCAAsC;IACtC,GAAG,CAAC;AAEN,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,GAAG,WAAW,MAAM,iBAAiB,MAAM,WAAW,OAAO,QAAQ,EAAE,EAAE,IAAI,CAAC,CAAC;AACvH,MAAM,wBAAwB,GAAG,IAAI,MAAM,CAAC,GAAG,WAAW,MAAM,iBAAiB,MAAM,WAAW,OAAO,QAAQ,EAAE,EAAE,GAAG,CAAC,CAAC;AAE1H,SAAS,kBAAkB,CAAC,CAAS;IACnC,kEAAkE;IAClE,OAAO,CAAC,CAAC,OAAO,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC;AACxC,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,IAAY;IAC5C,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,cAAc,GAAG,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,cAAc,GAAG,CAAC,CAAC;AACvF,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,aAAa,GAAG,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,aAAa,GAAG,CAAC,CAAC;AACrF,CAAC;AAED,MAAM,UAAU,4BAA4B,CAAC,IAAY;IACvD,IAAI,CAAC,IAAI;QAAE,OAAO,IAAI,CAAC;IACvB,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,oBAAoB,CAAC,CAAC,CAAC;IAChE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAEtC,IAAI,MAAM,GAAG,IAAI,CAAC;IAClB,iDAAiD;IACjD,KAAK,IAAI,CAAC,GAAG,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7C,MAAM,CAAC,GAAG,OAAO,CAAC,CAAC,CAAE,CAAC;QACtB,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACxB,IAAI,CAAC,IAAI;YAAE,SAAS;QAEpB,MAAM,UAAU,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC;QACjC,IAAI,UAAU,GAAG,CAAC;YAAE,SAAS;QAE7B,oFAAoF;QACpF,4DAA4D;QAC5D,MAAM,YAAY,GAAG,UAAU,CAAC;QAChC,MAAM,aAAa,GAAG,YAAY,GAAG,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,eAAe;QACrE,IAAI,aAAa,IAAI,MAAM,CAAC,MAAM;YAAE,SAAS;QAE7C,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;QAC7C,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,aAAa,GAAG,CAAC,CAAC,CAAC,CAAC,uBAAuB;QACtE,MAAM,GAAG,GAAG,MAAM,IAAI,cAAc,IAAI,IAAI,KAAK,cAAc,IAAI,KAAK,EAAE,CAAC;IAC7E,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,MAAM,UAAU,wBAAwB,CAAC,IAAY;IACnD,kGAAkG;IAClG,MAAM,CAAC,GAAG,wBAAwB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC9C,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,IAAI;QAAE,OAAO,IAAI,CAAC;IACvC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IACxB,IAAI,CAAC,IAAI;QAAE,OAAO,IAAI,CAAC;IACvB,MAAM,YAAY,GAAG,CAAC,CAAC,KAAK,CAAC;IAC7B,MAAM,SAAS,GAAG,YAAY,GAAG,CAAC,CAAC;IACnC,MAAM,OAAO,GAAG,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC;IACxC,OAAO,EAAE,IAAI,EAAE,UAAU,EAAE,SAAS,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC;AAC5D,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,IAAY;IAC9C,kEAAkE;IAClE,OAAO,IAAI;SACR,UAAU,CAAC,IAAI,MAAM,CAAC,IAAI,cAAc,GAAG,EAAE,GAAG,CAAC,EAAE,GAAG,CAAC;SACvD,UAAU,CAAC,IAAI,MAAM,CAAC,KAAK,cAAc,GAAG,EAAE,GAAG,CAAC,EAAE,GAAG,CAAC,CAAC;AAC9D,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,OAAO,IAAI;SACR,UAAU,CAAC,IAAI,MAAM,CAAC,IAAI,aAAa,GAAG,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC;SACrD,UAAU,CAAC,IAAI,MAAM,CAAC,KAAK,aAAa,GAAG,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC;AAC5D,CAAC;AAED,wEAAwE;AAExE,MAAM,iBAAiB,GAAG,aAAa,CAAC;AACxC,MAAM,iBAAiB,GAAG,qBAAqB,CAAC;AAChD,MAAM,kBAAkB,GAAG,QAAQ,CAAC;AAEpC,MAAM,UAAU,iBAAiB,CAAC,IAAY;IAC5C,OAAO,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AACtC,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,IAAY;IAC9C,kDAAkD;IAClD,iBAAiB,CAAC,SAAS,GAAG,CAAC,CAAC;IAChC,OAAO,IAAI,CAAC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC,CAAC;AAC7C,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,OAAO,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;AACvD,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,OAAO,IAAI,CAAC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC;AAC7E,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,qBAAqB,CAAC,IAAY;IAChD,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC;IACrB,4BAA4B;IAC5B,oBAAoB,CAAC,SAAS,GAAG,CAAC,CAAC;IACnC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,oBAAoB,CAAC,CAAC,CAAC;IAChE,MAAM,KAAK,GAAwD,EAAE,CAAC;IACtE,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;QACxB,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACxB,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,MAAM,UAAU,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC;QACjC,IAAI,UAAU,GAAG,CAAC;YAAE,SAAS;QAC7B,MAAM,YAAY,GAAG,UAAU,CAAC;QAChC,MAAM,OAAO,GAAG,YAAY,GAAG,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC;QAC/C,6BAA6B;QAC7B,IAAI,OAAO,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YAC1B,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,YAAY,EAAE,GAAG,EAAE,OAAO,GAAG,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC;QAC9D,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC"}

package/dist/primitives/simplify_redlines.d.ts

@@ -0,0 +1,19 @@
+/**
+ * simplify_redlines — merge adjacent same-author tracked-change wrappers.
+ *
+ * Consolidates consecutive `w:ins`, `w:del`, `w:moveFrom`, or `w:moveTo`
+ * wrappers that share the same `w:author` attribute and local name.
+ * Does not merge across different change types or non-whitespace separators.
+ */
+export type SimplifyRedlinesResult = {
+    wrappersConsolidated: number;
+};
+/**
+ * Simplify tracked-change wrappers across all paragraphs in the document
+ * body by merging adjacent same-type, same-author wrappers.
+ *
+ * This reduces visual clutter in redline documents without altering
+ * semantics (every run's tracked-change attribution is preserved).
+ */
+export declare function simplifyRedlines(doc: Document): SimplifyRedlinesResult;
+//# sourceMappingURL=simplify_redlines.d.ts.map

package/dist/primitives/simplify_redlines.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"simplify_redlines.d.ts","sourceRoot":"","sources":["../../src/primitives/simplify_redlines.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,MAAM,MAAM,sBAAsB,GAAG;IACnC,oBAAoB,EAAE,MAAM,CAAC;CAC9B,CAAC;AA2FF;;;;;;GAMG;AACH,wBAAgB,gBAAgB,CAAC,GAAG,EAAE,QAAQ,GAAG,sBAAsB,CAetE"}