wikimem 0.8.0 → 0.8.2

package/dist/processors/xlsx.js

@@ -1,13 +1,14 @@
 /**
  * Excel/spreadsheet (.xlsx, .xls) processor.
  * Uses xlsx (SheetJS) for extraction, with a raw XML fallback.
+ * Features: multiple sheets as sections, formula result values,
+ * bold/italic header detection, cell metadata.
  */
 import { readFileSync } from 'node:fs';
 import { basename, extname } from 'node:path';
 export async function processXlsx(filePath) {
     const ext = extname(filePath).toLowerCase();
     const title = basename(filePath, ext);
-    // Try SheetJS first (best quality), fall back to raw XML
     let content;
     let sheetCount = 0;
     try {
@@ -32,74 +33,142 @@ export async function processXlsx(filePath) {
     };
 }
 async function extractWithSheetJS(filePath) {
-    // Dynamic import — xlsx is an optional dependency
     const XLSX = await import('xlsx');
     const buffer = readFileSync(filePath);
-    const workbook = XLSX.read(buffer, { type: 'buffer' });
+    const workbook = XLSX.read(buffer, {
+        type: 'buffer',
+        cellFormula: true,
+        cellDates: true,
+        cellStyles: true,
+        cellNF: true,
+    });
     const sections = [];
-    for (const sheetName of workbook.SheetNames) {
+    const totalSheets = workbook.SheetNames.length;
+    for (let sheetIdx = 0; sheetIdx < totalSheets; sheetIdx++) {
+        const sheetName = workbook.SheetNames[sheetIdx];
+        if (!sheetName)
+            continue;
         const sheet = workbook.Sheets[sheetName];
         if (!sheet)
             continue;
-        // Convert sheet to array of arrays
-        const data = XLSX.utils.sheet_to_json(sheet, { header: 1 });
+        // Get sheet dimensions
+        const ref = sheet['!ref'];
+        let rowCount = 0;
+        let colCount = 0;
+        if (ref) {
+            const range = XLSX.utils.decode_range(ref);
+            rowCount = range.e.r - range.s.r + 1;
+            colCount = range.e.c - range.s.c + 1;
+        }
+        // Convert sheet to array of arrays (formula results rendered, not formula text)
+        const data = XLSX.utils.sheet_to_json(sheet, {
+            header: 1,
+            rawNumbers: false,
+            defval: '',
+        });
         if (data.length === 0)
             continue;
-        const table = arrayToMarkdownTable(data);
-        if (table) {
-            sections.push(`### ${sheetName}\n\n${table}`);
+        // Analyze cells for formatting and formulas
+        const cellInfoMap = analyzeCells(sheet, XLSX);
+        const headerRow = detectHeaderRow(data, cellInfoMap, XLSX);
+        // Build section header with metadata
+        let section = '';
+        if (totalSheets > 1) {
+            section += `### Sheet ${sheetIdx + 1}: ${sheetName}\n\n`;
+        }
+        else {
+            section += `### ${sheetName}\n\n`;
+        }
+        // Sheet metadata line
+        const metaParts = [];
+        metaParts.push(`_${rowCount} rows × ${colCount} columns_`);
+        const merges = sheet['!merges'];
+        if (merges && merges.length > 0) {
+            metaParts.push(`_${merges.length} merged region(s)_`);
+        }
+        const formulaCount = Object.values(cellInfoMap).filter((c) => c.formula).length;
+        if (formulaCount > 0) {
+            metaParts.push(`_${formulaCount} formula(s)_`);
         }
+        section += metaParts.join(' · ') + '\n\n';
+        // Build the table with formatting hints
+        const table = buildEnhancedTable(data, cellInfoMap, headerRow, XLSX);
+        if (table)
+            section += table;
+        // Formula summary (show formulas alongside their computed values)
+        const formulaSummary = buildFormulaSummary(cellInfoMap);
+        if (formulaSummary) {
+            section += '\n\n' + formulaSummary;
+        }
+        sections.push(section);
+    }
+    // Chart info
+    const chartInfo = extractChartInfo(workbook);
+    if (chartInfo) {
+        sections.push(`### Charts\n\n${chartInfo}`);
     }
     return {
         content: sections.join('\n\n---\n\n'),
         sheetCount: workbook.SheetNames.length,
     };
 }
-function extractFromRawXml(filePath) {
-    // .xlsx is a zip file — try to find sharedStrings.xml for text content
-    const buffer = readFileSync(filePath);
-    const content = buffer.toString('latin1');
-    const textParts = [];
-    // Look for <t> elements (shared strings in xlsx XML)
-    const textRegex = /<t[^>]*>([\s\S]*?)<\/t>/g;
-    let match;
-    while ((match = textRegex.exec(content)) !== null) {
-        if (match[1] && match[1].trim()) {
-            textParts.push(match[1].trim());
-        }
+/**
+ * Analyze cells for formatting (bold, italic) and formulas.
+ */
+function analyzeCells(sheet, XLSX) {
+    const cellMap = {};
+    for (const [key, val] of Object.entries(sheet)) {
+        if (key.startsWith('!'))
+            continue;
+        const cell = val;
+        const displayValue = cell.w ?? (cell.v !== null && cell.v !== undefined ? String(cell.v) : '');
+        cellMap[key] = {
+            address: key,
+            value: String(displayValue),
+            formula: cell.f ? `=${cell.f}` : undefined,
+            isBold: !!(cell.s?.font?.bold),
+            isItalic: !!(cell.s?.font?.italic),
+            type: cell.t ?? 's',
+        };
     }
-    // Also look for <v> elements (cell values)
-    const valueRegex = /<v>([\s\S]*?)<\/v>/g;
-    while ((match = valueRegex.exec(content)) !== null) {
-        if (match[1] && match[1].trim()) {
-            textParts.push(match[1].trim());
+    // Suppress unused import warning — XLSX is used by callers for decode
+    void XLSX;
+    return cellMap;
+}
+/**
+ * Detect which row is the header row.
+ * Heuristics: first row with all string values, or first row with bold cells.
+ */
+function detectHeaderRow(data, cellInfoMap, XLSX) {
+    if (data.length === 0)
+        return 0;
+    // Check if first row has bold cells
+    const firstRow = data[0];
+    if (Array.isArray(firstRow)) {
+        let boldCount = 0;
+        for (let col = 0; col < firstRow.length; col++) {
+            const addr = XLSX.utils.encode_cell({ r: 0, c: col });
+            if (cellInfoMap[addr]?.isBold)
+                boldCount++;
         }
+        if (boldCount > 0 && boldCount >= firstRow.length / 2)
+            return 0;
     }
-    // Count sheets
-    const sheetMatches = content.match(/<sheet /g);
-    const sheetCount = sheetMatches ? sheetMatches.length : 1;
-    if (textParts.length === 0) {
-        return { content: '', sheetCount };
-    }
-    // Present as a simple list since we can't reconstruct table structure
-    const uniqueParts = [...new Set(textParts)].slice(0, 500);
-    return {
-        content: `**Extracted cell values:**\n\n${uniqueParts.join(' | ')}`,
-        sheetCount,
-    };
+    // Default: first row is header
+    return 0;
 }
-function arrayToMarkdownTable(data) {
-    if (data.length === 0)
-        return '';
+/**
+ * Build markdown table with formatting hints.
+ * Bold cells get **wrapped**, italic cells get _wrapped_.
+ */
+function buildEnhancedTable(data, cellInfoMap, headerRow, XLSX) {
     // Filter out completely empty rows
     const rows = data.filter((row) => Array.isArray(row) && row.some((cell) => cell !== null && cell !== undefined && String(cell).trim() !== ''));
     if (rows.length === 0)
         return '';
-    // Determine max columns
     const maxCols = Math.max(...rows.map((row) => (Array.isArray(row) ? row.length : 0)));
     if (maxCols === 0)
         return '';
-    // Build markdown table
     const lines = [];
     for (let i = 0; i < Math.min(rows.length, 100); i++) {
         const row = rows[i];
@@ -108,12 +177,23 @@ function arrayToMarkdownTable(data) {
         const cells = [];
         for (let j = 0; j < maxCols; j++) {
             const cell = row[j];
-            const cellStr = cell !== null && cell !== undefined ? String(cell).replace(/\|/g, '\\|').replace(/\n/g, ' ') : '';
+            let cellStr = cell !== null && cell !== undefined ? String(cell).replace(/\|/g, '\\|').replace(/\n/g, ' ') : '';
+            // Apply formatting hints from cell style
+            const addr = XLSX.utils.encode_cell({ r: i, c: j });
+            const info = cellInfoMap[addr];
+            if (info) {
+                if (info.isBold && cellStr.length > 0) {
+                    cellStr = `**${cellStr}**`;
+                }
+                else if (info.isItalic && cellStr.length > 0) {
+                    cellStr = `_${cellStr}_`;
+                }
+            }
             cells.push(cellStr);
         }
         lines.push(`| ${cells.join(' | ')} |`);
-        // Add header separator after first row
-        if (i === 0) {
+        // Header separator after the detected header row
+        if (i === headerRow) {
             lines.push(`| ${cells.map(() => '---').join(' | ')} |`);
         }
     }
@@ -122,6 +202,62 @@ function arrayToMarkdownTable(data) {
     }
     return lines.join('\n');
 }
+/**
+ * Build a formula summary section showing formula → computed value.
+ */
+function buildFormulaSummary(cellInfoMap) {
+    const formulas = Object.values(cellInfoMap).filter((c) => c.formula);
+    if (formulas.length === 0)
+        return '';
+    const lines = ['**Formulas:**', ''];
+    const shown = formulas.slice(0, 15);
+    for (const f of shown) {
+        lines.push(`- \`${f.address}\`: \`${f.formula}\` → **${f.value}**`);
+    }
+    if (formulas.length > 15) {
+        lines.push(`- _...and ${formulas.length - 15} more formulas_`);
+    }
+    return lines.join('\n');
+}
+function extractChartInfo(workbook) {
+    const sheets = workbook.Sheets;
+    const chartNotes = [];
+    for (const [name, sheet] of Object.entries(sheets)) {
+        const type = sheet['!type'];
+        if (type === 'chart') {
+            chartNotes.push(`- **${name}**: Chart sheet detected`);
+        }
+    }
+    return chartNotes.length > 0 ? chartNotes.join('\n') : '';
+}
+function extractFromRawXml(filePath) {
+    const buffer = readFileSync(filePath);
+    const content = buffer.toString('latin1');
+    const textParts = [];
+    const textRegex = /<t[^>]*>([\s\S]*?)<\/t>/g;
+    let match;
+    while ((match = textRegex.exec(content)) !== null) {
+        if (match[1] && match[1].trim()) {
+            textParts.push(match[1].trim());
+        }
+    }
+    const valueRegex = /<v>([\s\S]*?)<\/v>/g;
+    while ((match = valueRegex.exec(content)) !== null) {
+        if (match[1] && match[1].trim()) {
+            textParts.push(match[1].trim());
+        }
+    }
+    const sheetMatches = content.match(/<sheet /g);
+    const sheetCount = sheetMatches ? sheetMatches.length : 1;
+    if (textParts.length === 0) {
+        return { content: '', sheetCount };
+    }
+    const uniqueParts = [...new Set(textParts)].slice(0, 500);
+    return {
+        content: `**Extracted cell values:**\n\n${uniqueParts.join(' | ')}`,
+        sheetCount,
+    };
+}
 function buildMarkdown(title, filePath, content, sheetCount) {
     return `# ${title}
 

package/dist/processors/xlsx.js.map

@@ -1 +1 @@
-{"version":3,"file":"xlsx.js","sourceRoot":"","sources":["../../src/processors/xlsx.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAU9C,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,QAAgB;IAChD,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IAC5C,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;IAEtC,yDAAyD;IACzD,IAAI,OAAe,CAAC;IACpB,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,kBAAkB,CAAC,QAAQ,CAAC,CAAC;QAClD,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC;QACzB,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC;IACjC,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,MAAM,GAAG,iBAAiB,CAAC,QAAQ,CAAC,CAAC;QAC3C,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC;QACzB,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC;IACjC,CAAC;IAED,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;QACpB,OAAO,GAAG,yCAAyC,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC;IAC3E,CAAC;IAED,OAAO;QACL,KAAK;QACL,OAAO;QACP,QAAQ,EAAE,aAAa,CAAC,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,UAAU,CAAC;QAC7D,UAAU;QACV,UAAU,EAAE,QAAQ;KACrB,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,kBAAkB,CAAC,QAAgB;IAChD,kDAAkD;IAClD,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,MAAM,CAAC,CAAC;IAClC,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;IAEvD,MAAM,QAAQ,GAAa,EAAE,CAAC;IAE9B,KAAK,MAAM,SAAS,IAAI,QAAQ,CAAC,UAAU,EAAE,CAAC;QAC5C,MAAM,KAAK,GAAG,QAAQ,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QACzC,IAAI,CAAC,KAAK;YAAE,SAAS;QAErB,mCAAmC;QACnC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,aAAa,CAAW,KAAK,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC,CAAC;QACtE,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS;QAEhC,MAAM,KAAK,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAC;QACzC,IAAI,KAAK,EAAE,CAAC;YACV,QAAQ,CAAC,IAAI,CAAC,OAAO,SAAS,OAAO,KAAK,EAAE,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;IAED,OAAO;QACL,OAAO,EAAE,QAAQ,CAAC,IAAI,CAAC,aAAa,CAAC;QACrC,UAAU,EAAE,QAAQ,CAAC,UAAU,CAAC,MAAM;KACvC,CAAC;AACJ,CAAC;AAED,SAAS,iBAAiB,CAAC,QAAgB;IACzC,uEAAuE;IACvE,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;IAE1C,MAAM,SAAS,GAAa,EAAE,CAAC;IAE/B,qDAAqD;IACrD,MAAM,SAAS,GAAG,0BAA0B,CAAC;IAC7C,IAAI,KAA6B,CAAC;IAElC,OAAO,CAAC,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAClD,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;YAChC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAClC,CAAC;IACH,CAAC;IAED,2CAA2C;IAC3C,MAAM,UAAU,GAAG,qBAAqB,CAAC;IACzC,OAAO,CAAC,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACnD,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;YAChC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAClC,CAAC;IACH,CAAC;IAED,eAAe;IACf,MAAM,YAAY,GAAG,OAAO,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;IAC/C,MAAM,UAAU,GAAG,YAAY,CAAC,CAAC,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAE1D,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,UAAU,EAAE,CAAC;IACrC,CAAC;IAED,sEAAsE;IACtE,MAAM,WAAW,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IAC1D,OAAO;QACL,OAAO,EAAE,iCAAiC,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE;QACnE,UAAU;KACX,CAAC;AACJ,CAAC;AAED,SAAS,oBAAoB,CAAC,IAAiB;IAC7C,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEjC,mCAAmC;IACnC,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAC/B,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,SAAS,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAC5G,CAAC;IAEF,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEjC,wBAAwB;IACxB,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACtF,IAAI,OAAO,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE7B,uBAAuB;IACvB,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,EAAE,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QACpD,MAAM,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACpB,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC;YAAE,SAAS;QAElC,MAAM,KAAK,GAAG,EAAE,CAAC;QACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,EAAE,CAAC,EAAE,EAAE,CAAC;YACjC,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;YACpB,MAAM,OAAO,GAAG,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YAClH,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACtB,CAAC;QACD,KAAK,CAAC,IAAI,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAEvC,uCAAuC;QACvC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;YACZ,KAAK,CAAC,IAAI,CAAC,KAAK,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC1D,CAAC;IACH,CAAC;IAED,IAAI,IAAI,CAAC,MAAM,GAAG,GAAG,EAAE,CAAC;QACtB,KAAK,CAAC,IAAI,CAAC,eAAe,IAAI,CAAC,MAAM,GAAG,GAAG,yBAAyB,CAAC,CAAC;IACxE,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED,SAAS,aAAa,CAAC,KAAa,EAAE,QAAgB,EAAE,OAAe,EAAE,UAAkB;IACzF,OAAO,KAAK,KAAK;;iBAEF,QAAQ,CAAC,QAAQ,CAAC,KAAK,QAAQ;2BACrB,OAAO,CAAC,QAAQ,CAAC;gBAC5B,UAAU;mBACP,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;;;;EAIvD,OAAO;CACR,CAAC;AACF,CAAC"}
+{"version":3,"file":"xlsx.js","sourceRoot":"","sources":["../../src/processors/xlsx.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAU9C,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,QAAgB;IAChD,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IAC5C,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;IAEtC,IAAI,OAAe,CAAC;IACpB,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,kBAAkB,CAAC,QAAQ,CAAC,CAAC;QAClD,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC;QACzB,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC;IACjC,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,MAAM,GAAG,iBAAiB,CAAC,QAAQ,CAAC,CAAC;QAC3C,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC;QACzB,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC;IACjC,CAAC;IAED,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;QACpB,OAAO,GAAG,yCAAyC,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC;IAC3E,CAAC;IAED,OAAO;QACL,KAAK;QACL,OAAO;QACP,QAAQ,EAAE,aAAa,CAAC,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,UAAU,CAAC;QAC7D,UAAU;QACV,UAAU,EAAE,QAAQ;KACrB,CAAC;AACJ,CAAC;AAWD,KAAK,UAAU,kBAAkB,CAAC,QAAgB;IAChD,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,MAAM,CAAC,CAAC;IAClC,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE;QACjC,IAAI,EAAE,QAAQ;QACd,WAAW,EAAE,IAAI;QACjB,SAAS,EAAE,IAAI;QACf,UAAU,EAAE,IAAI;QAChB,MAAM,EAAE,IAAI;KACb,CAAC,CAAC;IAEH,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,WAAW,GAAG,QAAQ,CAAC,UAAU,CAAC,MAAM,CAAC;IAE/C,KAAK,IAAI,QAAQ,GAAG,CAAC,EAAE,QAAQ,GAAG,WAAW,EAAE,QAAQ,EAAE,EAAE,CAAC;QAC1D,MAAM,SAAS,GAAG,QAAQ,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC;QAChD,IAAI,CAAC,SAAS;YAAE,SAAS;QACzB,MAAM,KAAK,GAAG,QAAQ,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QACzC,IAAI,CAAC,KAAK;YAAE,SAAS;QAErB,uBAAuB;QACvB,MAAM,GAAG,GAAI,KAAgC,CAAC,MAAM,CAAC,CAAC;QACtD,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,IAAI,GAAG,EAAE,CAAC;YACR,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC;YAC3C,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACrC,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;QACvC,CAAC;QAED,gFAAgF;QAChF,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,aAAa,CAAW,KAAK,EAAE;YACrD,MAAM,EAAE,CAAC;YACT,UAAU,EAAE,KAAK;YACjB,MAAM,EAAE,EAAE;SACX,CAAC,CAAC;QACH,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS;QAEhC,4CAA4C;QAC5C,MAAM,WAAW,GAAG,YAAY,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAC9C,MAAM,SAAS,GAAG,eAAe,CAAC,IAAI,EAAE,WAAW,EAAE,IAAI,CAAC,CAAC;QAE3D,qCAAqC;QACrC,IAAI,OAAO,GAAG,EAAE,CAAC;QACjB,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;YACpB,OAAO,IAAI,aAAa,QAAQ,GAAG,CAAC,KAAK,SAAS,MAAM,CAAC;QAC3D,CAAC;aAAM,CAAC;YACN,OAAO,IAAI,OAAO,SAAS,MAAM,CAAC;QACpC,CAAC;QAED,sBAAsB;QACtB,MAAM,SAAS,GAAa,EAAE,CAAC;QAC/B,SAAS,CAAC,IAAI,CAAC,IAAI,QAAQ,WAAW,QAAQ,WAAW,CAAC,CAAC;QAE3D,MAAM,MAAM,GAAI,KAAmC,CAAC,SAAS,CAAC,CAAC;QAC/D,IAAI,MAAM,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAChC,SAAS,CAAC,IAAI,CAAC,IAAI,MAAM,CAAC,MAAM,oBAAoB,CAAC,CAAC;QACxD,CAAC;QAED,MAAM,YAAY,GAAG,MAAM,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC;QAChF,IAAI,YAAY,GAAG,CAAC,EAAE,CAAC;YACrB,SAAS,CAAC,IAAI,CAAC,IAAI,YAAY,cAAc,CAAC,CAAC;QACjD,CAAC;QAED,OAAO,IAAI,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,MAAM,CAAC;QAE1C,wCAAwC;QACxC,MAAM,KAAK,GAAG,kBAAkB,CAAC,IAAI,EAAE,WAAW,EAAE,SAAS,EAAE,IAAI,CAAC,CAAC;QACrE,IAAI,KAAK;YAAE,OAAO,IAAI,KAAK,CAAC;QAE5B,kEAAkE;QAClE,MAAM,cAAc,GAAG,mBAAmB,CAAC,WAAW,CAAC,CAAC;QACxD,IAAI,cAAc,EAAE,CAAC;YACnB,OAAO,IAAI,MAAM,GAAG,cAAc,CAAC;QACrC,CAAC;QAED,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACzB,CAAC;IAED,aAAa;IACb,MAAM,SAAS,GAAG,gBAAgB,CAAC,QAA8C,CAAC,CAAC;IACnF,IAAI,SAAS,EAAE,CAAC;QACd,QAAQ,CAAC,IAAI,CAAC,iBAAiB,SAAS,EAAE,CAAC,CAAC;IAC9C,CAAC;IAED,OAAO;QACL,OAAO,EAAE,QAAQ,CAAC,IAAI,CAAC,aAAa,CAAC;QACrC,UAAU,EAAE,QAAQ,CAAC,UAAU,CAAC,MAAM;KACvC,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CACnB,KAA8B,EAC9B,IAA2B;IAE3B,MAAM,OAAO,GAA6B,EAAE,CAAC;IAE7C,KAAK,MAAM,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;QAC/C,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAClC,MAAM,IAAI,GAAG,GAA+G,CAAC;QAE7H,MAAM,YAAY,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,IAAI,IAAI,CAAC,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QAE/F,OAAO,CAAC,GAAG,CAAC,GAAG;YACb,OAAO,EAAE,GAAG;YACZ,KAAK,EAAE,MAAM,CAAC,YAAY,CAAC;YAC3B,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS;YAC1C,MAAM,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC;YAC9B,QAAQ,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC;YAClC,IAAI,EAAE,IAAI,CAAC,CAAC,IAAI,GAAG;SACpB,CAAC;IACJ,CAAC;IAED,sEAAsE;IACtE,KAAK,IAAI,CAAC;IAEV,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;GAGG;AACH,SAAS,eAAe,CACtB,IAAiB,EACjB,WAAqC,EACrC,IAA2B;IAE3B,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAEhC,oCAAoC;IACpC,MAAM,QAAQ,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;IACzB,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC5B,IAAI,SAAS,GAAG,CAAC,CAAC;QAClB,KAAK,IAAI,GAAG,GAAG,CAAC,EAAE,GAAG,GAAG,QAAQ,CAAC,MAAM,EAAE,GAAG,EAAE,EAAE,CAAC;YAC/C,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC;YACtD,IAAI,WAAW,CAAC,IAAI,CAAC,EAAE,MAAM;gBAAE,SAAS,EAAE,CAAC;QAC7C,CAAC;QACD,IAAI,SAAS,GAAG,CAAC,IAAI,SAAS,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,CAAC,CAAC;IAClE,CAAC;IAED,+BAA+B;IAC/B,OAAO,CAAC,CAAC;AACX,CAAC;AAED;;;GAGG;AACH,SAAS,kBAAkB,CACzB,IAAiB,EACjB,WAAqC,EACrC,SAAiB,EACjB,IAA2B;IAE3B,mCAAmC;IACnC,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAC/B,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,SAAS,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAC5G,CAAC;IAEF,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEjC,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACtF,IAAI,OAAO,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE7B,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,EAAE,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QACpD,MAAM,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACpB,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC;YAAE,SAAS;QAElC,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,EAAE,CAAC,EAAE,EAAE,CAAC;YACjC,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;YACpB,IAAI,OAAO,GAAG,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YAEhH,yCAAyC;YACzC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;YACpD,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;YAC/B,IAAI,IAAI,EAAE,CAAC;gBACT,IAAI,IAAI,CAAC,MAAM,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACtC,OAAO,GAAG,KAAK,OAAO,IAAI,CAAC;gBAC7B,CAAC;qBAAM,IAAI,IAAI,CAAC,QAAQ,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC/C,OAAO,GAAG,IAAI,OAAO,GAAG,CAAC;gBAC3B,CAAC;YACH,CAAC;YAED,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACtB,CAAC;QACD,KAAK,CAAC,IAAI,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAEvC,iDAAiD;QACjD,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACpB,KAAK,CAAC,IAAI,CAAC,KAAK,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC1D,CAAC;IACH,CAAC;IAED,IAAI,IAAI,CAAC,MAAM,GAAG,GAAG,EAAE,CAAC;QACtB,KAAK,CAAC,IAAI,CAAC,eAAe,IAAI,CAAC,MAAM,GAAG,GAAG,yBAAyB,CAAC,CAAC;IACxE,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED;;GAEG;AACH,SAAS,mBAAmB,CAAC,WAAqC;IAChE,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;IACrE,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAErC,MAAM,KAAK,GAAa,CAAC,eAAe,EAAE,EAAE,CAAC,CAAC;IAC9C,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAEpC,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,OAAO,SAAS,CAAC,CAAC,OAAO,UAAU,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC;IACtE,CAAC;IAED,IAAI,QAAQ,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,aAAa,QAAQ,CAAC,MAAM,GAAG,EAAE,iBAAiB,CAAC,CAAC;IACjE,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED,SAAS,gBAAgB,CAAC,QAAiC;IACzD,MAAM,MAAM,GAAI,QAAgE,CAAC,MAAM,CAAC;IACxF,MAAM,UAAU,GAAa,EAAE,CAAC;IAEhC,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QACnD,MAAM,IAAI,GAAI,KAAgC,CAAC,OAAO,CAAC,CAAC;QACxD,IAAI,IAAI,KAAK,OAAO,EAAE,CAAC;YACrB,UAAU,CAAC,IAAI,CAAC,OAAO,IAAI,0BAA0B,CAAC,CAAC;QACzD,CAAC;IACH,CAAC;IAED,OAAO,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AAC5D,CAAC;AAED,SAAS,iBAAiB,CAAC,QAAgB;IACzC,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;IAE1C,MAAM,SAAS,GAAa,EAAE,CAAC;IAE/B,MAAM,SAAS,GAAG,0BAA0B,CAAC;IAC7C,IAAI,KAA6B,CAAC;IAElC,OAAO,CAAC,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAClD,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;YAChC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAClC,CAAC;IACH,CAAC;IAED,MAAM,UAAU,GAAG,qBAAqB,CAAC;IACzC,OAAO,CAAC,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACnD,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;YAChC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAClC,CAAC;IACH,CAAC;IAED,MAAM,YAAY,GAAG,OAAO,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;IAC/C,MAAM,UAAU,GAAG,YAAY,CAAC,CAAC,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAE1D,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,UAAU,EAAE,CAAC;IACrC,CAAC;IAED,MAAM,WAAW,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IAC1D,OAAO;QACL,OAAO,EAAE,iCAAiC,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE;QACnE,UAAU;KACX,CAAC;AACJ,CAAC;AAED,SAAS,aAAa,CAAC,KAAa,EAAE,QAAgB,EAAE,OAAe,EAAE,UAAkB;IACzF,OAAO,KAAK,KAAK;;iBAEF,QAAQ,CAAC,QAAQ,CAAC,KAAK,QAAQ;2BACrB,OAAO,CAAC,QAAQ,CAAC;gBAC5B,UAAU;mBACP,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;;;;EAIvD,OAAO;CACR,CAAC;AACF,CAAC"}

package/dist/templates/source-types.d.ts

@@ -0,0 +1,33 @@
+/**
+ * Source-type templates for wiki page generation.
+ * Each template defines how a specific type of source should be
+ * structured when compiled into wiki pages.
+ *
+ * Used by ingest.ts to generate appropriately structured content.
+ */
+export interface SourceTypeTemplate {
+    id: string;
+    name: string;
+    description: string;
+    /** Keywords that help detect this source type from content */
+    detectionKeywords: string[];
+    /** MIME type patterns */
+    mimePatterns?: string[];
+    /** System prompt addition for LLM compilation */
+    systemPromptAddition: string;
+    /** Suggested frontmatter fields beyond the defaults */
+    extraFrontmatterFields: string[];
+    /** Suggested page structure (headings) */
+    suggestedStructure: string[];
+}
+export declare const SOURCE_TYPE_TEMPLATES: SourceTypeTemplate[];
+/**
+ * Detect source type from content and metadata.
+ * Returns the best-matching template or null for generic.
+ */
+export declare function detectSourceType(content: string, filename?: string, mimeType?: string): SourceTypeTemplate | null;
+/**
+ * Get the system prompt addition for a detected source type.
+ */
+export declare function getSourceTypePrompt(template: SourceTypeTemplate | null): string;
+//# sourceMappingURL=source-types.d.ts.map

package/dist/templates/source-types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"source-types.d.ts","sourceRoot":"","sources":["../../src/templates/source-types.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,MAAM,WAAW,kBAAkB;IACjC,EAAE,EAAE,MAAM,CAAC;IACX,IAAI,EAAE,MAAM,CAAC;IACb,WAAW,EAAE,MAAM,CAAC;IACpB,8DAA8D;IAC9D,iBAAiB,EAAE,MAAM,EAAE,CAAC;IAC5B,yBAAyB;IACzB,YAAY,CAAC,EAAE,MAAM,EAAE,CAAC;IACxB,iDAAiD;IACjD,oBAAoB,EAAE,MAAM,CAAC;IAC7B,uDAAuD;IACvD,sBAAsB,EAAE,MAAM,EAAE,CAAC;IACjC,0CAA0C;IAC1C,kBAAkB,EAAE,MAAM,EAAE,CAAC;CAC9B;AAED,eAAO,MAAM,qBAAqB,EAAE,kBAAkB,EAqHrD,CAAC;AAEF;;;GAGG;AACH,wBAAgB,gBAAgB,CAC9B,OAAO,EAAE,MAAM,EACf,QAAQ,CAAC,EAAE,MAAM,EACjB,QAAQ,CAAC,EAAE,MAAM,GAChB,kBAAkB,GAAG,IAAI,CAyC3B;AAED;;GAEG;AACH,wBAAgB,mBAAmB,CAAC,QAAQ,EAAE,kBAAkB,GAAG,IAAI,GAAG,MAAM,CAG/E"}

package/dist/templates/source-types.js

@@ -0,0 +1,178 @@
+/**
+ * Source-type templates for wiki page generation.
+ * Each template defines how a specific type of source should be
+ * structured when compiled into wiki pages.
+ *
+ * Used by ingest.ts to generate appropriately structured content.
+ */
+export const SOURCE_TYPE_TEMPLATES = [
+    {
+        id: 'article',
+        name: 'Article / Blog Post',
+        description: 'Web articles, blog posts, news pieces',
+        detectionKeywords: ['article', 'blog', 'post', 'published', 'author', 'byline'],
+        mimePatterns: ['text/html'],
+        systemPromptAddition: `This source is a web article or blog post. Structure the wiki pages to capture:
+- The core thesis or argument
+- Key claims with supporting evidence
+- Author's perspective and potential biases
+- Practical takeaways and implications
+- How this connects to other knowledge in the wiki
+Use citation-style references when quoting the original.`,
+        extraFrontmatterFields: ['author', 'publication', 'publish_date', 'url'],
+        suggestedStructure: ['Summary', 'Key Arguments', 'Evidence & Claims', 'Implications', 'Related Concepts'],
+    },
+    {
+        id: 'paper',
+        name: 'Academic Paper / Research',
+        description: 'Research papers, whitepapers, academic publications',
+        detectionKeywords: ['abstract', 'methodology', 'findings', 'conclusion', 'references', 'doi', 'arxiv', 'et al'],
+        mimePatterns: ['application/pdf'],
+        systemPromptAddition: `This source is an academic paper or research document. Structure the wiki pages to capture:
+- Abstract / TL;DR in accessible language
+- Research question and hypothesis
+- Methodology overview (non-technical summary)
+- Key findings and results
+- Limitations acknowledged by authors
+- Practical implications and applications
+- How this advances the field
+Translate academic jargon into plain language while preserving accuracy.`,
+        extraFrontmatterFields: ['authors', 'institution', 'year', 'doi', 'conference', 'citation_count'],
+        suggestedStructure: ['Abstract', 'Research Question', 'Methodology', 'Key Findings', 'Limitations', 'Implications', 'References'],
+    },
+    {
+        id: 'tweet-thread',
+        name: 'Tweet Thread / Social Post',
+        description: 'Twitter/X threads, social media discussions',
+        detectionKeywords: ['thread', 'tweet', '🧵', '1/', 'RT', '@', 'x.com', 'twitter.com'],
+        systemPromptAddition: `This source is a social media thread (likely Twitter/X). Structure the wiki pages to capture:
+- The main argument condensed from the thread
+- Key insights (numbered, one per point)
+- Any data, links, or evidence shared
+- Community reactions or counter-arguments if visible
+- The author's credibility and context
+Threads are informal — extract the substance while noting the conversational tone.`,
+        extraFrontmatterFields: ['author_handle', 'platform', 'post_date', 'engagement'],
+        suggestedStructure: ['Main Argument', 'Key Points', 'Evidence Shared', 'Context & Credibility'],
+    },
+    {
+        id: 'podcast',
+        name: 'Podcast / Audio Content',
+        description: 'Podcast episodes, audio interviews, lectures',
+        detectionKeywords: ['podcast', 'episode', 'interview', 'host', 'guest', 'transcript', 'listen'],
+        mimePatterns: ['audio/'],
+        systemPromptAddition: `This source is a podcast episode or audio content (likely transcribed). Structure the wiki pages to capture:
+- Episode summary (who, what, why it matters)
+- Key quotes from each speaker (attributed)
+- Main topics discussed (as separate sections)
+- Actionable advice or takeaways
+- Disagreements or debates between speakers
+- Resources/links mentioned
+Attribute quotes to speakers. Separate opinion from fact.`,
+        extraFrontmatterFields: ['host', 'guests', 'podcast_name', 'episode_number', 'duration', 'air_date'],
+        suggestedStructure: ['Episode Overview', 'Speakers', 'Key Topics', 'Notable Quotes', 'Takeaways', 'Resources Mentioned'],
+    },
+    {
+        id: 'video',
+        name: 'Video / Lecture',
+        description: 'YouTube videos, lectures, presentations, tutorials',
+        detectionKeywords: ['video', 'youtube', 'lecture', 'presentation', 'tutorial', 'watch', 'timestamp'],
+        mimePatterns: ['video/'],
+        systemPromptAddition: `This source is a video or lecture (likely transcribed). Structure the wiki pages to capture:
+- Video summary with key timestamps
+- Main concepts taught or discussed
+- Visual elements described (diagrams, demos, code shown)
+- Step-by-step instructions if tutorial
+- Speaker's key arguments
+- Q&A highlights if present
+Note: visual content may be lost in transcription — flag where visuals were important.`,
+        extraFrontmatterFields: ['creator', 'channel', 'duration', 'publish_date', 'url', 'platform'],
+        suggestedStructure: ['Overview', 'Key Concepts', 'Timestamps', 'Visual Notes', 'Takeaways'],
+    },
+    {
+        id: 'book',
+        name: 'Book / Long-form',
+        description: 'Books, book chapters, long-form essays, reports',
+        detectionKeywords: ['chapter', 'book', 'isbn', 'publisher', 'edition', 'foreword', 'preface'],
+        systemPromptAddition: `This source is a book or long-form document. Structure the wiki pages to capture:
+- Book summary (thesis, scope, intended audience)
+- Chapter-by-chapter key takeaways
+- Core frameworks or models introduced
+- Memorable examples and case studies
+- Practical applications
+- Author's background and perspective
+- How this connects to other books/ideas in the wiki
+Create separate wiki pages for major concepts introduced by the book.`,
+        extraFrontmatterFields: ['author', 'publisher', 'year', 'isbn', 'genre', 'pages'],
+        suggestedStructure: ['Book Summary', 'Core Thesis', 'Key Frameworks', 'Chapter Notes', 'Practical Applications', 'Related Works'],
+    },
+    {
+        id: 'notes',
+        name: 'Raw Notes / Meeting Notes',
+        description: 'Personal notes, meeting notes, brainstorms, voice memos',
+        detectionKeywords: ['notes', 'meeting', 'action items', 'TODO', 'decision', 'brainstorm', 'idea'],
+        systemPromptAddition: `This source is raw notes (meeting notes, personal notes, brainstorm). Structure the wiki pages to capture:
+- Clean summary of what was discussed/noted
+- Action items extracted (with owners if mentioned)
+- Decisions made
+- Open questions or unresolved items
+- Key ideas worth developing further
+- Connections to existing wiki topics
+Raw notes are messy — your job is to extract structure and meaning. Link to existing wiki pages where concepts overlap.`,
+        extraFrontmatterFields: ['meeting_date', 'participants', 'context'],
+        suggestedStructure: ['Summary', 'Key Points', 'Action Items', 'Decisions', 'Open Questions', 'Related Topics'],
+    },
+];
+/**
+ * Detect source type from content and metadata.
+ * Returns the best-matching template or null for generic.
+ */
+export function detectSourceType(content, filename, mimeType) {
+    const lowerContent = content.toLowerCase().slice(0, 2000); // Only check first 2K chars
+    const lowerFilename = (filename ?? '').toLowerCase();
+    let bestMatch = null;
+    let bestScore = 0;
+    for (const template of SOURCE_TYPE_TEMPLATES) {
+        let score = 0;
+        // Check keywords in content
+        for (const keyword of template.detectionKeywords) {
+            if (lowerContent.includes(keyword.toLowerCase()))
+                score += 1;
+        }
+        // Check MIME type
+        if (mimeType && template.mimePatterns) {
+            for (const pattern of template.mimePatterns) {
+                if (mimeType.startsWith(pattern))
+                    score += 3;
+            }
+        }
+        // Check filename hints
+        if (lowerFilename.includes('paper') || lowerFilename.includes('arxiv')) {
+            if (template.id === 'paper')
+                score += 2;
+        }
+        if (lowerFilename.includes('notes') || lowerFilename.includes('meeting')) {
+            if (template.id === 'notes')
+                score += 2;
+        }
+        if (lowerFilename.includes('transcript') || lowerFilename.includes('podcast')) {
+            if (template.id === 'podcast')
+                score += 2;
+        }
+        if (score > bestScore) {
+            bestScore = score;
+            bestMatch = template;
+        }
+    }
+    // Require minimum confidence (at least 2 keyword matches)
+    return bestScore >= 2 ? bestMatch : null;
+}
+/**
+ * Get the system prompt addition for a detected source type.
+ */
+export function getSourceTypePrompt(template) {
+    if (!template)
+        return '';
+    return `\n\nSOURCE TYPE DETECTED: ${template.name}\n${template.systemPromptAddition}`;
+}
+//# sourceMappingURL=source-types.js.map

package/dist/templates/source-types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"source-types.js","sourceRoot":"","sources":["../../src/templates/source-types.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAkBH,MAAM,CAAC,MAAM,qBAAqB,GAAyB;IACzD;QACE,EAAE,EAAE,SAAS;QACb,IAAI,EAAE,qBAAqB;QAC3B,WAAW,EAAE,uCAAuC;QACpD,iBAAiB,EAAE,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,WAAW,EAAE,QAAQ,EAAE,QAAQ,CAAC;QAC/E,YAAY,EAAE,CAAC,WAAW,CAAC;QAC3B,oBAAoB,EAAE;;;;;;yDAM+B;QACrD,sBAAsB,EAAE,CAAC,QAAQ,EAAE,aAAa,EAAE,cAAc,EAAE,KAAK,CAAC;QACxE,kBAAkB,EAAE,CAAC,SAAS,EAAE,eAAe,EAAE,mBAAmB,EAAE,cAAc,EAAE,kBAAkB,CAAC;KAC1G;IACD;QACE,EAAE,EAAE,OAAO;QACX,IAAI,EAAE,2BAA2B;QACjC,WAAW,EAAE,qDAAqD;QAClE,iBAAiB,EAAE,CAAC,UAAU,EAAE,aAAa,EAAE,UAAU,EAAE,YAAY,EAAE,YAAY,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,CAAC;QAC/G,YAAY,EAAE,CAAC,iBAAiB,CAAC;QACjC,oBAAoB,EAAE;;;;;;;;yEAQ+C;QACrE,sBAAsB,EAAE,CAAC,SAAS,EAAE,aAAa,EAAE,MAAM,EAAE,KAAK,EAAE,YAAY,EAAE,gBAAgB,CAAC;QACjG,kBAAkB,EAAE,CAAC,UAAU,EAAE,mBAAmB,EAAE,aAAa,EAAE,cAAc,EAAE,aAAa,EAAE,cAAc,EAAE,YAAY,CAAC;KAClI;IACD;QACE,EAAE,EAAE,cAAc;QAClB,IAAI,EAAE,4BAA4B;QAClC,WAAW,EAAE,6CAA6C;QAC1D,iBAAiB,EAAE,CAAC,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,OAAO,EAAE,aAAa,CAAC;QACrF,oBAAoB,EAAE;;;;;;mFAMyD;QAC/E,sBAAsB,EAAE,CAAC,eAAe,EAAE,UAAU,EAAE,WAAW,EAAE,YAAY,CAAC;QAChF,kBAAkB,EAAE,CAAC,eAAe,EAAE,YAAY,EAAE,iBAAiB,EAAE,uBAAuB,CAAC;KAChG;IACD;QACE,EAAE,EAAE,SAAS;QACb,IAAI,EAAE,yBAAyB;QAC/B,WAAW,EAAE,8CAA8C;QAC3D,iBAAiB,EAAE,CAAC,SAAS,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,EAAE,OAAO,EAAE,YAAY,EAAE,QAAQ,CAAC;QAC/F,YAAY,EAAE,CAAC,QAAQ,CAAC;QACxB,oBAAoB,EAAE;;;;;;;0DAOgC;QACtD,sBAAsB,EAAE,CAAC,MAAM,EAAE,QAAQ,EAAE,cAAc,EAAE,gBAAgB,EAAE,UAAU,EAAE,UAAU,CAAC;QACpG,kBAAkB,EAAE,CAAC,kBAAkB,EAAE,UAAU,EAAE,YAAY,EAAE,gBAAgB,EAAE,WAAW,EAAE,qBAAqB,CAAC;KACzH;IACD;QACE,EAAE,EAAE,OAAO;QACX,IAAI,EAAE,iBAAiB;QACvB,WAAW,EAAE,oDAAoD;QACjE,iBAAiB,EAAE,CAAC,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,cAAc,EAAE,UAAU,EAAE,OAAO,EAAE,WAAW,CAAC;QACpG,YAAY,EAAE,CAAC,QAAQ,CAAC;QACxB,oBAAoB,EAAE;;;;;;;uFAO6D;QACnF,sBAAsB,EAAE,CAAC,SAAS,EAAE,SAAS,EAAE,UAAU,EAAE,cAAc,EAAE,KAAK,EAAE,UAAU,CAAC;QAC7F,kBAAkB,EAAE,CAAC,UAAU,EAAE,cAAc,EAAE,YAAY,EAAE,cAAc,EAAE,WAAW,CAAC;KAC5F;IACD;QACE,EAAE,EAAE,MAAM;QACV,IAAI,EAAE,kBAAkB;QACxB,WAAW,EAAE,iDAAiD;QAC9D,iBAAiB,EAAE,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,WAAW,EAAE,SAAS,EAAE,UAAU,EAAE,SAAS,CAAC;QAC7F,oBAAoB,EAAE;;;;;;;;sEAQ4C;QAClE,sBAAsB,EAAE,CAAC,QAAQ,EAAE,WAAW,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,CAAC;QACjF,kBAAkB,EAAE,CAAC,cAAc,EAAE,aAAa,EAAE,gBAAgB,EAAE,eAAe,EAAE,wBAAwB,EAAE,eAAe,CAAC;KAClI;IACD;QACE,EAAE,EAAE,OAAO;QACX,IAAI,EAAE,2BAA2B;QACjC,WAAW,EAAE,yDAAyD;QACtE,iBAAiB,EAAE,CAAC,OAAO,EAAE,SAAS,EAAE,cAAc,EAAE,MAAM,EAAE,UAAU,EAAE,YAAY,EAAE,MAAM,CAAC;QACjG,oBAAoB,EAAE;;;;;;;wHAO8F;QACpH,sBAAsB,EAAE,CAAC,cAAc,EAAE,cAAc,EAAE,SAAS,CAAC;QACnE,kBAAkB,EAAE,CAAC,SAAS,EAAE,YAAY,EAAE,cAAc,EAAE,WAAW,EAAE,gBAAgB,EAAE,gBAAgB,CAAC;KAC/G;CACF,CAAC;AAEF;;;GAGG;AACH,MAAM,UAAU,gBAAgB,CAC9B,OAAe,EACf,QAAiB,EACjB,QAAiB;IAEjB,MAAM,YAAY,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,4BAA4B;IACvF,MAAM,aAAa,GAAG,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IAErD,IAAI,SAAS,GAA8B,IAAI,CAAC;IAChD,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,MAAM,QAAQ,IAAI,qBAAqB,EAAE,CAAC;QAC7C,IAAI,KAAK,GAAG,CAAC,CAAC;QAEd,4BAA4B;QAC5B,KAAK,MAAM,OAAO,IAAI,QAAQ,CAAC,iBAAiB,EAAE,CAAC;YACjD,IAAI,YAAY,CAAC,QAAQ,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;QAC/D,CAAC;QAED,kBAAkB;QAClB,IAAI,QAAQ,IAAI,QAAQ,CAAC,YAAY,EAAE,CAAC;YACtC,KAAK,MAAM,OAAO,IAAI,QAAQ,CAAC,YAAY,EAAE,CAAC;gBAC5C,IAAI,QAAQ,CAAC,UAAU,CAAC,OAAO,CAAC;oBAAE,KAAK,IAAI,CAAC,CAAC;YAC/C,CAAC;QACH,CAAC;QAED,uBAAuB;QACvB,IAAI,aAAa,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,aAAa,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC;YACvE,IAAI,QAAQ,CAAC,EAAE,KAAK,OAAO;gBAAE,KAAK,IAAI,CAAC,CAAC;QAC1C,CAAC;QACD,IAAI,aAAa,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,aAAa,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC;YACzE,IAAI,QAAQ,CAAC,EAAE,KAAK,OAAO;gBAAE,KAAK,IAAI,CAAC,CAAC;QAC1C,CAAC;QACD,IAAI,aAAa,CAAC,QAAQ,CAAC,YAAY,CAAC,IAAI,aAAa,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC;YAC9E,IAAI,QAAQ,CAAC,EAAE,KAAK,SAAS;gBAAE,KAAK,IAAI,CAAC,CAAC;QAC5C,CAAC;QAED,IAAI,KAAK,GAAG,SAAS,EAAE,CAAC;YACtB,SAAS,GAAG,KAAK,CAAC;YAClB,SAAS,GAAG,QAAQ,CAAC;QACvB,CAAC;IACH,CAAC;IAED,0DAA0D;IAC1D,OAAO,SAAS,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AAC3C,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,mBAAmB,CAAC,QAAmC;IACrE,IAAI,CAAC,QAAQ;QAAE,OAAO,EAAE,CAAC;IACzB,OAAO,6BAA6B,QAAQ,CAAC,IAAI,KAAK,QAAQ,CAAC,oBAAoB,EAAE,CAAC;AACxF,CAAC"}