@open-agreements/open-agreements 0.2.0

package/dist/core/recipe/bracket-normalizer.js

@@ -0,0 +1,354 @@
+import AdmZip from 'adm-zip';
+import { writeFileSync } from 'node:fs';
+import { DOMParser, XMLSerializer } from '@xmldom/xmldom';
+import { enumerateTextParts, getGeneralTextPartNames } from './ooxml-parts.js';
+import { BLANK_PLACEHOLDER } from '../fill-utils.js';
+const W_NS = 'http://schemas.openxmlformats.org/wordprocessingml/2006/main';
+/**
+ * Normalize residual bracket artifacts in generated recipe documents.
+ *
+ * Strategy:
+ * - Unbracket simple single-level clauses (`[text]` -> `text`)
+ * - Remove complex/nested segments with unresolved placeholders
+ * - Remove orphan bracket characters and punctuation artifacts
+ * - Drop paragraphs that collapse to punctuation-only content
+ */
+export async function normalizeBracketArtifacts(inputPath, outputPath, options) {
+    const zip = new AdmZip(inputPath);
+    const parser = new DOMParser();
+    const serializer = new XMLSerializer();
+    const parts = enumerateTextParts(zip);
+    const partNames = getGeneralTextPartNames(parts);
+    const stats = {
+        unbracketedSegments: 0,
+        removedSegments: 0,
+        removedParagraphs: 0,
+        normalizedParagraphs: 0,
+        declarativeRuleApplications: 0,
+        declarativeRuleMatchCounts: {},
+        declarativeRuleMutationCounts: {},
+        declarativeRuleExpectationFailures: [],
+        declarativeRuleExamples: [],
+    };
+    const rules = options?.rules ?? [];
+    const fieldValues = options?.fieldValues ?? {};
+    const blankPlaceholder = options?.blankPlaceholder ?? BLANK_PLACEHOLDER;
+    const hasDeclarativeRules = rules.length > 0;
+    for (const rule of rules) {
+        if (!(rule.id in stats.declarativeRuleMatchCounts)) {
+            stats.declarativeRuleMatchCounts[rule.id] = 0;
+            stats.declarativeRuleMutationCounts[rule.id] = 0;
+        }
+    }
+    for (const partName of partNames) {
+        const entry = zip.getEntry(partName);
+        if (!entry)
+            continue;
+        const xml = entry.getData().toString('utf-8');
+        const doc = parser.parseFromString(xml, 'text/xml');
+        const paragraphs = doc.getElementsByTagNameNS(W_NS, 'p');
+        const toRemove = [];
+        let lastHeading = '';
+        for (let i = 0; i < paragraphs.length; i++) {
+            const para = paragraphs[i];
+            const original = extractParagraphText(para);
+            if (!original.trim())
+                continue;
+            const trimmedOriginal = original.trim();
+            if (isHeadingLike(trimmedOriginal)) {
+                lastHeading = trimmedOriginal;
+            }
+            const declarativeResult = applyDeclarativeRulesToParagraph({
+                text: original,
+                heading: lastHeading,
+                rules,
+                fieldValues,
+                blankPlaceholder,
+            });
+            if (declarativeResult.applied) {
+                stats.declarativeRuleApplications += 1;
+            }
+            if (declarativeResult.rule_id) {
+                const ruleId = declarativeResult.rule_id;
+                stats.declarativeRuleMatchCounts[ruleId] =
+                    (stats.declarativeRuleMatchCounts[ruleId] ?? 0) + 1;
+                if (declarativeResult.text !== original) {
+                    stats.declarativeRuleMutationCounts[ruleId] =
+                        (stats.declarativeRuleMutationCounts[ruleId] ?? 0) + 1;
+                }
+                const existingExamples = stats.declarativeRuleExamples.filter((example) => example.rule_id === ruleId).length;
+                if (existingExamples < 2) {
+                    stats.declarativeRuleExamples.push({
+                        rule_id: ruleId,
+                        heading: lastHeading,
+                        before: original,
+                        after: declarativeResult.text,
+                    });
+                }
+            }
+            // If declarative rules are provided, prefer targeted mutation only.
+            // This avoids broad paragraph rewrites that can shift inline styling.
+            if (hasDeclarativeRules) {
+                if (declarativeResult.text !== original) {
+                    setParagraphText(para, declarativeResult.text);
+                    stats.normalizedParagraphs += 1;
+                }
+                continue;
+            }
+            const normalized = normalizeParagraphText(declarativeResult.text, stats);
+            if (normalized === null) {
+                toRemove.push(para);
+                stats.removedParagraphs += 1;
+                continue;
+            }
+            if (normalized !== original) {
+                setParagraphText(para, normalized);
+                stats.normalizedParagraphs += 1;
+            }
+        }
+        for (const para of toRemove) {
+            para.parentNode?.removeChild(para);
+        }
+        zip.updateFile(partName, Buffer.from(serializer.serializeToString(doc), 'utf-8'));
+    }
+    for (const rule of rules) {
+        if (rule.expected_min_matches === undefined)
+            continue;
+        const actualMatches = stats.declarativeRuleMatchCounts[rule.id] ?? 0;
+        if (actualMatches < rule.expected_min_matches) {
+            stats.declarativeRuleExpectationFailures.push(`${rule.id}: expected at least ${rule.expected_min_matches} match(es), found ${actualMatches}`);
+        }
+    }
+    const outZip = new AdmZip();
+    for (const entry of zip.getEntries()) {
+        outZip.addFile(entry.entryName, entry.getData());
+    }
+    writeFileSync(outputPath, outZip.toBuffer());
+    return stats;
+}
+function isHeadingLike(text) {
+    return !text.startsWith('.') && !text.includes('[') && text.replaceAll('.', '').trim().length > 0;
+}
+function applyDeclarativeRulesToParagraph(params) {
+    const { text, heading, rules, fieldValues, blankPlaceholder } = params;
+    let mutated = text;
+    let matchedRule = false;
+    let matchedRuleId;
+    for (const rule of rules) {
+        if (!matchesHeading(rule, heading))
+            continue;
+        if (!mutated.includes(rule.paragraph_contains))
+            continue;
+        if (rule.paragraph_end_contains && !mutated.includes(rule.paragraph_end_contains))
+            continue;
+        matchedRule = true;
+        matchedRuleId = rule.id;
+        for (const [token, template] of Object.entries(rule.replacements ?? {})) {
+            const resolved = resolveTemplateValue(template, fieldValues, blankPlaceholder);
+            mutated = mutated.split(token).join(resolved);
+        }
+        if (rule.trim_unmatched_trailing_bracket) {
+            const openCount = (mutated.match(/\[/g) ?? []).length;
+            const closeCount = (mutated.match(/\]/g) ?? []).length;
+            if (closeCount > openCount) {
+                mutated = mutated.replace(/\]+$/g, '');
+            }
+        }
+        mutated = mutated
+            .replace(/\s{2,}/g, ' ')
+            .replace(/\s+\./g, '.')
+            .replace(/^\.\s*/, '')
+            .trim();
+        break;
+    }
+    // Lightweight fallback cleanup for declarative mode:
+    // 1) Preserve bracket-prefixed heading labels by removing only leading '['.
+    // 2) Trim unmatched trailing ']' artifacts that remain after option pruning.
+    mutated = normalizeBracketPrefixedHeading(mutated);
+    mutated = trimUnmatchedTrailingBrackets(mutated);
+    return { text: mutated, applied: matchedRule || mutated !== text, rule_id: matchedRuleId };
+}
+function matchesHeading(rule, heading) {
+    if (rule.ignore_heading)
+        return true;
+    if (heading.includes(rule.section_heading))
+        return true;
+    for (const alias of rule.section_heading_any ?? []) {
+        if (heading.includes(alias))
+            return true;
+    }
+    return false;
+}
+function resolveTemplateValue(template, fieldValues, blankPlaceholder) {
+    return template.replace(/\{([a-zA-Z0-9_]+)\}/g, (_full, key) => {
+        const value = fieldValues[key];
+        if (typeof value === 'boolean')
+            return value ? 'true' : 'false';
+        if (typeof value === 'string' && value.trim().length > 0)
+            return value;
+        return blankPlaceholder;
+    });
+}
+function normalizeBracketPrefixedHeading(text) {
+    const trimmed = text.trim();
+    if (!trimmed.startsWith('['))
+        return text;
+    if (trimmed.includes(']'))
+        return text;
+    if (trimmed.length > 90)
+        return text;
+    if (trimmed.includes('_'))
+        return text;
+    if (/[.:;!?].+[.:;!?]/.test(trimmed))
+        return text;
+    if (!/^\[[A-Z][A-Za-z0-9 ,&()'’/.-]+$/.test(trimmed))
+        return text;
+    return trimmed.slice(1).trim();
+}
+function trimUnmatchedTrailingBrackets(text) {
+    let out = text;
+    const openCount = (out.match(/\[/g) ?? []).length;
+    let closeCount = (out.match(/\]/g) ?? []).length;
+    while (closeCount > openCount && /\]\s*$/.test(out)) {
+        out = out.replace(/\]\s*$/, '').trimEnd();
+        closeCount -= 1;
+    }
+    return out;
+}
+function normalizeParagraphText(text, stats) {
+    let out = text;
+    // First pass: handle complete same-paragraph bracket segments.
+    const segments = extractTopLevelSegments(out);
+    if (segments.length > 0) {
+        const ordered = [...segments].sort((a, b) => b.start - a.start);
+        for (const seg of ordered) {
+            const replacement = shouldRemoveSegment(seg.inner) ? '' : seg.inner.trim();
+            out = `${out.slice(0, seg.start)}${replacement}${out.slice(seg.end + 1)}`;
+            if (replacement === '') {
+                stats.removedSegments += 1;
+            }
+            else {
+                stats.unbracketedSegments += 1;
+            }
+        }
+    }
+    // Second pass: remove orphan bracket chars from cross-paragraph constructs.
+    out = out.replaceAll('[', '').replaceAll(']', '');
+    // Punctuation/spacing normalization for artifacts created by bracket deletion.
+    out = out
+        .replace(/^\s*[.;:,]\s*/, '')
+        .replace(/\s+,/g, ',')
+        .replace(/\s+;/g, ';')
+        .replace(/\s+\./g, '.')
+        .replace(/\bfrom,\s+counsel\b/gi, 'from counsel')
+        .replace(/\(\s+/g, '(')
+        .replace(/\s+\)/g, ')')
+        .replace(/\s{2,}/g, ' ')
+        .trim();
+    // Prefer a single deterministic costs clause when both alternatives are present.
+    if (out.includes('If any action at law or in equity') &&
+        out.includes('Each party will bear its own costs in respect of any disputes arising under this Agreement.')) {
+        out = 'Each party will bear its own costs in respect of any disputes arising under this Agreement.';
+    }
+    // Drop known degenerate artifacts produced when optional bracket blocks are stripped.
+    if (/^A minimum of\s+Shares must be sold at the Initial Closing\.?$/i.test(out)) {
+        return null;
+    }
+    if (/^As of the Initial Closing, the authorized size of the Board of Directors shall be, and the Board of Directors shall be comprised of\.?$/i.test(out)) {
+        return null;
+    }
+    // Drop degenerate paragraphs that are now effectively empty.
+    if (!out || /^[.,;:]+$/.test(out)) {
+        return null;
+    }
+    return out;
+}
+function shouldRemoveSegment(inner) {
+    const normalized = inner.trim();
+    if (!normalized)
+        return true;
+    // Nested bracket segments are usually unresolved alternatives; drop them.
+    if (normalized.includes('[') || normalized.includes(']'))
+        return true;
+    // Placeholder blanks and underscore markers should be removed.
+    if (normalized.includes('_'))
+        return true;
+    // Known placeholder-only tokens.
+    if (/^(date|specify|state|initial|applicable|audited|unaudited)$/i.test(normalized))
+        return true;
+    if (/^\(?[ivx]+\)?$/i.test(normalized))
+        return true;
+    if (/^\d+$/.test(normalized))
+        return true;
+    return false;
+}
+function extractTopLevelSegments(text) {
+    const out = [];
+    let depth = 0;
+    let start = -1;
+    for (let i = 0; i < text.length; i++) {
+        const ch = text[i];
+        if (ch === '[') {
+            if (depth === 0)
+                start = i;
+            depth += 1;
+        }
+        else if (ch === ']') {
+            if (depth > 0)
+                depth -= 1;
+            if (depth === 0 && start !== -1) {
+                const raw = text.slice(start, i + 1);
+                out.push({ start, end: i, raw, inner: raw.slice(1, -1) });
+                start = -1;
+            }
+        }
+    }
+    return out;
+}
+function extractParagraphText(para) {
+    const tElements = para.getElementsByTagNameNS(W_NS, 't');
+    const parts = [];
+    for (let i = 0; i < tElements.length; i++) {
+        parts.push(tElements[i].textContent ?? '');
+    }
+    return parts.join('');
+}
+function setParagraphText(para, text) {
+    const doc = para.ownerDocument;
+    const tElements = para.getElementsByTagNameNS(W_NS, 't');
+    if (tElements.length === 0) {
+        const run = doc.createElementNS(W_NS, 'w:r');
+        const t = doc.createElementNS(W_NS, 'w:t');
+        t.textContent = text;
+        run.appendChild(t);
+        para.appendChild(run);
+        return;
+    }
+    // Preserve existing run-level formatting by distributing normalized text
+    // across current text nodes using original node-length quotas.
+    const originalLengths = [];
+    for (let i = 0; i < tElements.length; i++) {
+        originalLengths.push((tElements[i].textContent ?? '').length);
+    }
+    let cursor = 0;
+    const total = text.length;
+    for (let i = 0; i < tElements.length; i++) {
+        const quota = i === tElements.length - 1
+            ? total - cursor
+            : Math.min(originalLengths[i], Math.max(0, total - cursor));
+        if (quota > 0) {
+            tElements[i].textContent = text.slice(cursor, cursor + quota);
+            cursor += quota;
+        }
+        else {
+            tElements[i].textContent = '';
+        }
+    }
+    // If the paragraph grew beyond original capacity, append the tail to
+    // the last node to avoid creating new runs and style surprises.
+    if (cursor < total) {
+        const last = tElements[tElements.length - 1];
+        last.textContent = (last.textContent ?? '') + text.slice(cursor);
+    }
+}
+//# sourceMappingURL=bracket-normalizer.js.map

package/dist/core/recipe/bracket-normalizer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"bracket-normalizer.js","sourceRoot":"","sources":["../../../src/core/recipe/bracket-normalizer.ts"],"names":[],"mappings":"AAAA,OAAO,MAAM,MAAM,SAAS,CAAC;AAC7B,OAAO,EAAE,aAAa,EAAE,MAAM,SAAS,CAAC;AACxC,OAAO,EAAE,SAAS,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE1D,OAAO,EAAE,kBAAkB,EAAE,uBAAuB,EAAE,MAAM,kBAAkB,CAAC;AAC/E,OAAO,EAAE,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAErD,MAAM,IAAI,GAAG,8DAA8D,CAAC;AAgD5E;;;;;;;;GAQG;AACH,MAAM,CAAC,KAAK,UAAU,yBAAyB,CAC7C,SAAiB,EACjB,UAAkB,EAClB,OAAqC;IAErC,MAAM,GAAG,GAAG,IAAI,MAAM,CAAC,SAAS,CAAC,CAAC;IAClC,MAAM,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;IAC/B,MAAM,UAAU,GAAG,IAAI,aAAa,EAAE,CAAC;IACvC,MAAM,KAAK,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAC;IACtC,MAAM,SAAS,GAAG,uBAAuB,CAAC,KAAK,CAAC,CAAC;IAEjD,MAAM,KAAK,GAA8B;QACvC,mBAAmB,EAAE,CAAC;QACtB,eAAe,EAAE,CAAC;QAClB,iBAAiB,EAAE,CAAC;QACpB,oBAAoB,EAAE,CAAC;QACvB,2BAA2B,EAAE,CAAC;QAC9B,0BAA0B,EAAE,EAAE;QAC9B,6BAA6B,EAAE,EAAE;QACjC,kCAAkC,EAAE,EAAE;QACtC,uBAAuB,EAAE,EAAE;KAC5B,CAAC;IACF,MAAM,KAAK,GAAG,OAAO,EAAE,KAAK,IAAI,EAAE,CAAC;IACnC,MAAM,WAAW,GAAG,OAAO,EAAE,WAAW,IAAI,EAAE,CAAC;IAC/C,MAAM,gBAAgB,GAAG,OAAO,EAAE,gBAAgB,IAAI,iBAAiB,CAAC;IACxE,MAAM,mBAAmB,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;IAC7C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,CAAC,IAAI,CAAC,EAAE,IAAI,KAAK,CAAC,0BAA0B,CAAC,EAAE,CAAC;YACnD,KAAK,CAAC,0BAA0B,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC;YAC9C,KAAK,CAAC,6BAA6B,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC;QACnD,CAAC;IACH,CAAC;IAED,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;QACjC,MAAM,KAAK,GAAG,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;QACrC,IAAI,CAAC,KAAK;YAAE,SAAS;QAErB,MAAM,GAAG,GAAG,KAAK,CAAC,OAAO,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;QAC9C,MAAM,GAAG,GAAa,MAAM,CAAC,eAAe,CAAC,GAAG,EAAE,UAAU,CAAC,CAAC;QAC9D,MAAM,UAAU,GAAG,GAAG,CAAC,sBAAsB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;QACzD,MAAM,QAAQ,GAAc,EAAE,CAAC;QAC/B,IAAI,WAAW,GAAG,EAAE,CAAC;QAErB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3C,MAAM,IAAI,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;YAC3B,MAAM,QAAQ,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAC;YAC5C,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE;gBAAE,SAAS;YAE/B,MAAM,eAAe,GAAG,QAAQ,CAAC,IAAI,EAAE,CAAC;YACxC,IAAI,aAAa,CAAC,eAAe,CAAC,EAAE,CAAC;gBACnC,WAAW,GAAG,eAAe,CAAC;YAChC,CAAC;YAED,MAAM,iBAAiB,GAAG,gCAAgC,CAAC;gBACzD,IAAI,EAAE,QAAQ;gBACd,OAAO,EAAE,WAAW;gBACpB,KAAK;gBACL,WAAW;gBACX,gBAAgB;aACjB,CAAC,CAAC;YACH,IAAI,iBAAiB,CAAC,OAAO,EAAE,CAAC;gBAC9B,KAAK,CAAC,2BAA2B,IAAI,CAAC,CAAC;YACzC,CAAC;YACD,IAAI,iBAAiB,CAAC,OAAO,EAAE,CAAC;gBAC9B,MAAM,MAAM,GAAG,iBAAiB,CAAC,OAAO,CAAC;gBACzC,KAAK,CAAC,0BAA0B,CAAC,MAAM,CAAC;oBACtC,CAAC,KAAK,CAAC,0BAA0B,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;gBACtD,IAAI,iBAAiB,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;oBACxC,KAAK,CAAC,6BAA6B,CAAC,MAAM,CAAC;wBACzC,CAAC,KAAK,CAAC,6BAA6B,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;gBAC3D,CAAC;gBACD,MAAM,gBAAgB,GAAG,KAAK,CAAC,uBAAuB,CAAC,MAAM,CAC3D,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,OAAO,KAAK,MAAM,CACxC,CAAC,MAAM,CAAC;gBACT,IAAI,gBAAgB,GAAG,CAAC,EAAE,CAAC;oBACzB,KAAK,CAAC,uBAAuB,CAAC,IAAI,CAAC;wBACjC,OAAO,EAAE,MAAM;wBACf,OAAO,EAAE,WAAW;wBACpB,MAAM,EAAE,QAAQ;wBAChB,KAAK,EAAE,iBAAiB,CAAC,IAAI;qBAC9B,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAED,oEAAoE;YACpE,sEAAsE;YACtE,IAAI,mBAAmB,EAAE,CAAC;gBACxB,IAAI,iBAAiB,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;oBACxC,gBAAgB,CAAC,IAAI,EAAE,iBAAiB,CAAC,IAAI,CAAC,CAAC;oBAC/C,KAAK,CAAC,oBAAoB,IAAI,CAAC,CAAC;gBAClC,CAAC;gBACD,SAAS;YACX,CAAC;YAED,MAAM,UAAU,GAAG,sBAAsB,CAAC,iBAAiB,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;YACzE,IAAI,UAAU,KAAK,IAAI,EAAE,CAAC;gBACxB,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACpB,KAAK,CAAC,iBAAiB,IAAI,CAAC,CAAC;gBAC7B,SAAS;YACX,CAAC;YAED,IAAI,UAAU,KAAK,QAAQ,EAAE,CAAC;gBAC5B,gBAAgB,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;gBACnC,KAAK,CAAC,oBAAoB,IAAI,CAAC,CAAC;YAClC,CAAC;QACH,CAAC;QAED,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC5B,IAAI,CAAC,UAAU,EAAE,WAAW,CAAC,IAAI,CAAC,CAAC;QACrC,CAAC;QAED,GAAG,CAAC,UAAU,CAAC,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,iBAAiB,CAAC,GAAG,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;IACpF,CAAC;IAED,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,CAAC,oBAAoB,KAAK,SAAS;YAAE,SAAS;QACtD,MAAM,aAAa,GAAG,KAAK,CAAC,0BAA0B,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC;QACrE,IAAI,aAAa,GAAG,IAAI,CAAC,oBAAoB,EAAE,CAAC;YAC9C,KAAK,CAAC,kCAAkC,CAAC,IAAI,CAC3C,GAAG,IAAI,CAAC,EAAE,uBAAuB,IAAI,CAAC,oBAAoB,qBAAqB,aAAa,EAAE,CAC/F,CAAC;QACJ,CAAC;IACH,CAAC;IAED,MAAM,MAAM,GAAG,IAAI,MAAM,EAAE,CAAC;IAC5B,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,UAAU,EAAE,EAAE,CAAC;QACrC,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,SAAS,EAAE,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;IACnD,CAAC;IACD,aAAa,CAAC,UAAU,EAAE,MAAM,CAAC,QAAQ,EAAE,CAAC,CAAC;IAE7C,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,aAAa,CAAC,IAAY;IACjC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC;AACpG,CAAC;AAED,SAAS,gCAAgC,CAAC,MAMzC;IACC,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,WAAW,EAAE,gBAAgB,EAAE,GAAG,MAAM,CAAC;IACvE,IAAI,OAAO,GAAG,IAAI,CAAC;IACnB,IAAI,WAAW,GAAG,KAAK,CAAC;IACxB,IAAI,aAAiC,CAAC;IAEtC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,cAAc,CAAC,IAAI,EAAE,OAAO,CAAC;YAAE,SAAS;QAC7C,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,kBAAkB,CAAC;YAAE,SAAS;QACzD,IAAI,IAAI,CAAC,sBAAsB,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,sBAAsB,CAAC;YAAE,SAAS;QAC5F,WAAW,GAAG,IAAI,CAAC;QACnB,aAAa,GAAG,IAAI,CAAC,EAAE,CAAC;QAExB,KAAK,MAAM,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,YAAY,IAAI,EAAE,CAAC,EAAE,CAAC;YACxE,MAAM,QAAQ,GAAG,oBAAoB,CAAC,QAAQ,EAAE,WAAW,EAAE,gBAAgB,CAAC,CAAC;YAC/E,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAChD,CAAC;QAED,IAAI,IAAI,CAAC,+BAA+B,EAAE,CAAC;YACzC,MAAM,SAAS,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;YACtD,MAAM,UAAU,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;YACvD,IAAI,UAAU,GAAG,SAAS,EAAE,CAAC;gBAC3B,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;YACzC,CAAC;QACH,CAAC;QAED,OAAO,GAAG,OAAO;aACd,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;aACvB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;aACtB,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC;aACrB,IAAI,EAAE,CAAC;QAEV,MAAM;IACR,CAAC;IAED,qDAAqD;IACrD,4EAA4E;IAC5E,6EAA6E;IAC7E,OAAO,GAAG,+BAA+B,CAAC,OAAO,CAAC,CAAC;IACnD,OAAO,GAAG,6BAA6B,CAAC,OAAO,CAAC,CAAC;IAEjD,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,WAAW,IAAI,OAAO,KAAK,IAAI,EAAE,OAAO,EAAE,aAAa,EAAE,CAAC;AAC7F,CAAC;AAED,SAAS,cAAc,CAAC,IAAuC,EAAE,OAAe;IAC9E,IAAI,IAAI,CAAC,cAAc;QAAE,OAAO,IAAI,CAAC;IACrC,IAAI,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,eAAe,CAAC;QAAE,OAAO,IAAI,CAAC;IACxD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,mBAAmB,IAAI,EAAE,EAAE,CAAC;QACnD,IAAI,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAC;YAAE,OAAO,IAAI,CAAC;IAC3C,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,oBAAoB,CAC3B,QAAgB,EAChB,WAA6C,EAC7C,gBAAwB;IAExB,OAAO,QAAQ,CAAC,OAAO,CAAC,sBAAsB,EAAE,CAAC,KAAK,EAAE,GAAW,EAAE,EAAE;QACrE,MAAM,KAAK,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC;QAC/B,IAAI,OAAO,KAAK,KAAK,SAAS;YAAE,OAAO,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC;QAChE,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,KAAK,CAAC;QACvE,OAAO,gBAAgB,CAAC;IAC1B,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,+BAA+B,CAAC,IAAY;IACnD,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC5B,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAC1C,IAAI,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACvC,IAAI,OAAO,CAAC,MAAM,GAAG,EAAE;QAAE,OAAO,IAAI,CAAC;IACrC,IAAI,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACvC,IAAI,kBAAkB,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,IAAI,CAAC;IAClD,IAAI,CAAC,iCAAiC,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,IAAI,CAAC;IAClE,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;AACjC,CAAC;AAED,SAAS,6BAA6B,CAAC,IAAY;IACjD,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,MAAM,SAAS,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;IAClD,IAAI,UAAU,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;IACjD,OAAO,UAAU,GAAG,SAAS,IAAI,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QACpD,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;QAC1C,UAAU,IAAI,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,sBAAsB,CAAC,IAAY,EAAE,KAAgC;IAC5E,IAAI,GAAG,GAAG,IAAI,CAAC;IAEf,+DAA+D;IAC/D,MAAM,QAAQ,GAAG,uBAAuB,CAAC,GAAG,CAAC,CAAC;IAC9C,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACxB,MAAM,OAAO,GAAG,CAAC,GAAG,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;QAChE,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;YAC1B,MAAM,WAAW,GAAG,mBAAmB,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;YAC3E,GAAG,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,KAAK,CAAC,GAAG,WAAW,GAAG,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC;YAC1E,IAAI,WAAW,KAAK,EAAE,EAAE,CAAC;gBACvB,KAAK,CAAC,eAAe,IAAI,CAAC,CAAC;YAC7B,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,mBAAmB,IAAI,CAAC,CAAC;YACjC,CAAC;QACH,CAAC;IACH,CAAC;IAED,4EAA4E;IAC5E,GAAG,GAAG,GAAG,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;IAElD,+EAA+E;IAC/E,GAAG,GAAG,GAAG;SACN,OAAO,CAAC,eAAe,EAAE,EAAE,CAAC;SAC5B,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,uBAAuB,EAAE,cAAc,CAAC;SAChD,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,IAAI,EAAE,CAAC;IAEV,iFAAiF;IACjF,IACE,GAAG,CAAC,QAAQ,CAAC,mCAAmC,CAAC;QACjD,GAAG,CAAC,QAAQ,CAAC,6FAA6F,CAAC,EAC3G,CAAC;QACD,GAAG,GAAG,6FAA6F,CAAC;IACtG,CAAC;IAED,sFAAsF;IACtF,IAAI,iEAAiE,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QAChF,OAAO,IAAI,CAAC;IACd,CAAC;IACD,IACE,2IAA2I,CAAC,IAAI,CAC9I,GAAG,CACJ,EACD,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,6DAA6D;IAC7D,IAAI,CAAC,GAAG,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QAClC,OAAO,IAAI,CAAC;IACd,CAAC;IAED,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,mBAAmB,CAAC,KAAa;IACxC,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAChC,IAAI,CAAC,UAAU;QAAE,OAAO,IAAI,CAAC;IAE7B,0EAA0E;IAC1E,IAAI,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAEtE,+DAA+D;IAC/D,IAAI,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAE1C,iCAAiC;IACjC,IAAI,8DAA8D,CAAC,IAAI,CAAC,UAAU,CAAC;QAAE,OAAO,IAAI,CAAC;IACjG,IAAI,iBAAiB,CAAC,IAAI,CAAC,UAAU,CAAC;QAAE,OAAO,IAAI,CAAC;IACpD,IAAI,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC;QAAE,OAAO,IAAI,CAAC;IAE1C,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,uBAAuB,CAAC,IAAY;IAC3C,MAAM,GAAG,GAAc,EAAE,CAAC;IAC1B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC;IAEf,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACnB,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,IAAI,KAAK,KAAK,CAAC;gBAAE,KAAK,GAAG,CAAC,CAAC;YAC3B,KAAK,IAAI,CAAC,CAAC;QACb,CAAC;aAAM,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACtB,IAAI,KAAK,GAAG,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC1B,IAAI,KAAK,KAAK,CAAC,IAAI,KAAK,KAAK,CAAC,CAAC,EAAE,CAAC;gBAChC,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;gBACrC,GAAG,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;gBAC1D,KAAK,GAAG,CAAC,CAAC,CAAC;YACb,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,oBAAoB,CAAC,IAAa;IACzC,MAAM,SAAS,GAAG,IAAI,CAAC,sBAAsB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IACzD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC1C,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC;IAC7C,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACxB,CAAC;AAED,SAAS,gBAAgB,CAAC,IAAa,EAAE,IAAY;IACnD,MAAM,GAAG,GAAG,IAAI,CAAC,aAAyB,CAAC;IAC3C,MAAM,SAAS,GAAG,IAAI,CAAC,sBAAsB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAEzD,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,MAAM,GAAG,GAAG,GAAG,CAAC,eAAe,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC7C,MAAM,CAAC,GAAG,GAAG,CAAC,eAAe,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC3C,CAAC,CAAC,WAAW,GAAG,IAAI,CAAC;QACrB,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC;QACnB,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;QACtB,OAAO;IACT,CAAC;IAED,yEAAyE;IACzE,+DAA+D;IAC/D,MAAM,eAAe,GAAa,EAAE,CAAC;IACrC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC1C,eAAe,CAAC,IAAI,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC;IAChE,CAAC;IAED,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC;IAE1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC1C,MAAM,KAAK,GAAG,CAAC,KAAK,SAAS,CAAC,MAAM,GAAG,CAAC;YACtC,CAAC,CAAC,KAAK,GAAG,MAAM;YAChB,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,eAAe,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC;QAC9D,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACd,SAAS,CAAC,CAAC,CAAC,CAAC,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,KAAK,CAAC,CAAC;YAC9D,MAAM,IAAI,KAAK,CAAC;QAClB,CAAC;aAAM,CAAC;YACN,SAAS,CAAC,CAAC,CAAC,CAAC,WAAW,GAAG,EAAE,CAAC;QAChC,CAAC;IACH,CAAC;IAED,qEAAqE;IACrE,gEAAgE;IAChE,IAAI,MAAM,GAAG,KAAK,EAAE,CAAC;QACnB,MAAM,IAAI,GAAG,SAAS,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC7C,IAAI,CAAC,WAAW,GAAG,CAAC,IAAI,CAAC,WAAW,IAAI,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACnE,CAAC;AACH,CAAC"}

package/dist/core/recipe/cleaner.d.ts

@@ -0,0 +1,27 @@
+import type { CleanConfig, GuidanceOutput } from '../metadata.js';
+export interface CleanResult {
+    outputPath: string;
+    guidance?: GuidanceOutput;
+}
+export interface CleanOptions {
+    extractGuidance?: boolean;
+    /** Pre-computed hashes for guidance staleness detection */
+    sourceHash?: string;
+    configHash?: string;
+}
+/**
+ * Clean a DOCX document by removing footnotes, pattern-matched paragraphs,
+ * and clearing specified parts.
+ * Operates at the OOXML level to preserve formatting of retained content.
+ *
+ * Processes all general OOXML text parts (document, headers, footers, endnotes).
+ * Footnotes.xml is handled separately with its separator/continuationSeparator logic.
+ *
+ * When options.extractGuidance is true, captures the text content of all
+ * removed elements before deletion and returns them as structured guidance data.
+ *
+ * Note: Helpers use xmldom's own `Document`/`Element`/`Node` types to avoid
+ * incompatibilities with the global DOM lib declarations.
+ */
+export declare function cleanDocument(inputPath: string, outputPath: string, config: CleanConfig, options?: CleanOptions): Promise<CleanResult>;
+//# sourceMappingURL=cleaner.d.ts.map

package/dist/core/recipe/cleaner.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cleaner.d.ts","sourceRoot":"","sources":["../../../src/core/recipe/cleaner.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,WAAW,EAAiB,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAKjF,MAAM,WAAW,WAAW;IAC1B,UAAU,EAAE,MAAM,CAAC;IACnB,QAAQ,CAAC,EAAE,cAAc,CAAC;CAC3B;AAED,MAAM,WAAW,YAAY;IAC3B,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,2DAA2D;IAC3D,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED;;;;;;;;;;;;;GAaG;AACH,wBAAsB,aAAa,CACjC,SAAS,EAAE,MAAM,EACjB,UAAU,EAAE,MAAM,EAClB,MAAM,EAAE,WAAW,EACnB,OAAO,CAAC,EAAE,YAAY,GACrB,OAAO,CAAC,WAAW,CAAC,CAmItB"}