@cesarandreslopez/occ 0.3.0 → 0.4.0

package/dist/src/sheet/output.js

@@ -0,0 +1,158 @@
+import Table from 'cli-table3';
+import chalk from 'chalk';
+import { formatBytes, formatNumber } from '../utils.js';
+import { sectionHeader, stripAnsi, tableChars } from '../output/tabular.js';
+const colors = {
+    header: (value) => chalk.bold(value),
+    key: (value) => chalk.cyan(value),
+    value: (value) => value,
+    dim: (value) => chalk.dim(value),
+    ok: (value) => chalk.green(value),
+    warn: (value) => chalk.yellow(value),
+};
+const noColor = Object.fromEntries(Object.keys(colors).map(key => [key, (value) => value]));
+function palette(ci = false) {
+    return ci ? noColor : colors;
+}
+export function formatSheetPayloadJson(payload) {
+    return JSON.stringify(payload, null, 2);
+}
+export function formatSheetInspection(result, ci = false) {
+    const c = palette(ci);
+    const sections = [];
+    const overviewLines = [
+        `${c.key('File')}: ${c.value(result.workbook.file)}`,
+        `${c.key('Format')}: ${c.value('XLSX')}`,
+        `${c.key('Size')}: ${c.value(formatBytes(result.workbook.size))}`,
+        `${c.key('Sheets')}: ${c.value(`${formatNumber(result.workbook.sheetCount)} total (${formatNumber(result.workbook.visibleSheetCount)} visible, ${formatNumber(result.workbook.hiddenSheetCount)} hidden, ${formatNumber(result.workbook.veryHiddenSheetCount)} very hidden)`)}`,
+    ];
+    const propEntries = Object.entries(result.workbook.properties).filter(([, value]) => value);
+    if (propEntries.length > 0) {
+        overviewLines.push(`${c.key('Properties')}: ${propEntries.map(([key, value]) => `${key}=${value}`).join(' | ')}`);
+    }
+    if (result.workbook.definedNames.length > 0) {
+        overviewLines.push(`${c.key('Workbook Names')}: ${result.workbook.definedNames.map(name => `${name.name}=${name.ref}`).join(' | ')}`);
+    }
+    const riskFlags = Object.entries(result.workbook.riskFlags).filter(([, enabled]) => enabled).map(([key]) => key);
+    overviewLines.push(`${c.key('Risk Flags')}: ${riskFlags.length > 0 ? c.warn(riskFlags.join(', ')) : c.ok('none')}`);
+    sections.push(overviewLines.join('\n'));
+    if (result.sheets.length > 0) {
+        const inventory = new Table({
+            head: [
+                c.key('Sheet'),
+                c.key('Visibility'),
+                c.key('Range'),
+                c.key('Rows'),
+                c.key('Cols'),
+                c.key('Non-Empty'),
+                c.key('Formulae'),
+                c.key('Links'),
+                c.key('Comments'),
+                c.key('Tokens'),
+            ],
+            chars: tableChars(ci),
+            style: { head: [], border: [] },
+            colAligns: ['left', 'left', 'left', 'right', 'right', 'right', 'right', 'right', 'right', 'right'],
+        });
+        for (const sheet of result.sheets) {
+            inventory.push([
+                `${sheet.index}. ${sheet.name}`,
+                sheet.visibility,
+                sheet.usedRange ?? '',
+                formatNumber(sheet.totalRows),
+                formatNumber(sheet.totalCols),
+                formatNumber(sheet.nonEmptyCellCount),
+                formatNumber(sheet.formulaCellCount),
+                formatNumber(sheet.hyperlinkCellCount),
+                formatNumber(sheet.commentCellCount),
+                formatNumber(sheet.fullTokenEstimate),
+            ]);
+        }
+        const rendered = inventory.toString();
+        const width = stripAnsi(rendered.split('\n')[0] ?? '').length;
+        sections.push(`${c.header(sectionHeader('Sheet Inventory', width, ci))}\n${rendered}`);
+    }
+    for (const sheet of result.sheets) {
+        sections.push(formatSheetProfile(sheet, ci));
+    }
+    return sections.join('\n\n') + '\n';
+}
+function formatSheetProfile(sheet, ci = false) {
+    const c = palette(ci);
+    const lines = [];
+    const title = `Sheet: ${sheet.name} (${sheet.visibility})`;
+    const summary = [
+        `${c.key('Range')}: ${sheet.usedRange ?? '(empty)'}`,
+        `${c.key('Grid')}: ${formatNumber(sheet.totalRows)} rows x ${formatNumber(sheet.totalCols)} cols (${formatNumber(sheet.rectangularRangeCellCount)} cells)`,
+        `${c.key('Signals')}: ${formatNumber(sheet.formulaCellCount)} formulae | ${formatNumber(sheet.commentCellCount)} comments | ${formatNumber(sheet.hyperlinkCellCount)} hyperlinks | ${formatNumber(sheet.mergedRangeCount)} merges`,
+        `${c.key('Layout')}: ${formatNumber(sheet.hiddenRowCount)} hidden rows | ${formatNumber(sheet.hiddenColumnCount)} hidden cols | autofilter=${sheet.autoFilterRef ?? 'none'} | protected=${sheet.protected ? 'yes' : 'no'}`,
+        `${c.key('Header')}: ${sheet.headerSelection.rowNumber ?? 'none'} (${sheet.headerSelection.mode})`,
+        `${c.key('Token Estimate')}: sample=${formatNumber(sheet.sampleTokenEstimate)} | full=${formatNumber(sheet.fullTokenEstimate)}`,
+    ];
+    if (sheet.definedNames.length > 0) {
+        summary.push(`${c.key('Defined Names')}: ${sheet.definedNames.map(name => `${name.name}=${name.ref}`).join(' | ')}`);
+    }
+    lines.push(summary.join('\n'));
+    if (sheet.schema.columns.length > 0) {
+        lines.push(formatSchemaTable(sheet.schema.columns, sheet.schema.truncated, ci));
+    }
+    else {
+        lines.push(c.dim('No schema columns found.'));
+    }
+    if (sheet.sample.rows.length > 0) {
+        lines.push(formatSampleTable(sheet.sample.rows, sheet.sample.truncatedRows, sheet.sample.truncatedColumns, ci));
+    }
+    else {
+        lines.push(c.dim('No sample rows found.'));
+    }
+    const width = Math.max(...lines.map(line => stripAnsi(line.split('\n')[0] ?? '').length), title.length, 40);
+    return `${c.header(sectionHeader(title, width, ci))}\n${lines.join('\n\n')}`;
+}
+function formatSchemaTable(columns, truncated, ci = false) {
+    const c = palette(ci);
+    const table = new Table({
+        head: [
+            c.key('Col'),
+            c.key('Name'),
+            c.key('Type'),
+            c.key('Non-Empty'),
+            c.key('Coverage'),
+            c.key('Examples'),
+        ],
+        chars: tableChars(ci),
+        style: { head: [], border: [] },
+        colAligns: ['left', 'left', 'left', 'right', 'right', 'left'],
+    });
+    for (const column of columns) {
+        table.push([
+            `${column.letter} (${column.index})`,
+            column.name,
+            column.dominantType,
+            formatNumber(column.nonEmptyCount),
+            `${Math.round(column.nonEmptyRatio * 100)}%`,
+            column.examples.join(' | '),
+        ]);
+    }
+    const suffix = truncated ? '\n(truncated columns)' : '';
+    return `${palette(ci).header('Schema')}\n${table.toString()}${suffix}`;
+}
+function formatSampleTable(rows, truncatedRows, truncatedColumns, ci = false) {
+    const c = palette(ci);
+    const columnNames = Object.keys(rows[0]?.values ?? {});
+    const table = new Table({
+        head: [c.key('Row'), ...columnNames.map(name => c.key(name))],
+        chars: tableChars(ci),
+        style: { head: [], border: [] },
+    });
+    for (const row of rows) {
+        table.push([String(row.rowNumber), ...columnNames.map(name => row.values[name] ?? '')]);
+    }
+    const notes = [];
+    if (truncatedRows)
+        notes.push('truncated rows');
+    if (truncatedColumns)
+        notes.push('truncated columns');
+    const suffix = notes.length > 0 ? `\n(${notes.join(', ')})` : '';
+    return `${palette(ci).header('Sample')}\n${table.toString()}${suffix}`;
+}
+//# sourceMappingURL=output.js.map

package/dist/src/sheet/output.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"output.js","sourceRoot":"","sources":["../../../src/sheet/output.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,MAAM,YAAY,CAAC;AAC/B,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AACxD,OAAO,EAAE,aAAa,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,sBAAsB,CAAC;AAc5E,MAAM,MAAM,GAAgB;IAC1B,MAAM,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;IACpC,GAAG,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;IACjC,KAAK,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK;IACvB,GAAG,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC;IAChC,EAAE,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC;IACjC,IAAI,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC;CACrC,CAAC;AAEF,MAAM,OAAO,GAAgB,MAAM,CAAC,WAAW,CAC7C,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC,KAAa,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CACtC,CAAC;AAE5B,SAAS,OAAO,CAAC,EAAE,GAAG,KAAK;IACzB,OAAO,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC;AAC/B,CAAC;AAED,MAAM,UAAU,sBAAsB,CAAC,OAA4B;IACjE,OAAO,IAAI,CAAC,SAAS,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;AAC1C,CAAC;AAED,MAAM,UAAU,qBAAqB,CAAC,MAA6B,EAAE,EAAE,GAAG,KAAK;IAC7E,MAAM,CAAC,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC;IACtB,MAAM,QAAQ,GAAa,EAAE,CAAC;IAE9B,MAAM,aAAa,GAAG;QACpB,GAAG,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE;QACpD,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE;QACxC,GAAG,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE;QACjE,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,GAAG,YAAY,CAAC,MAAM,CAAC,QAAQ,CAAC,UAAU,CAAC,WAAW,YAAY,CAAC,MAAM,CAAC,QAAQ,CAAC,iBAAiB,CAAC,aAAa,YAAY,CAAC,MAAM,CAAC,QAAQ,CAAC,gBAAgB,CAAC,YAAY,YAAY,CAAC,MAAM,CAAC,QAAQ,CAAC,oBAAoB,CAAC,eAAe,CAAC,EAAE;KAChR,CAAC;IACF,MAAM,WAAW,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC;IAC5F,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC3B,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,KAAK,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,IAAI,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACpH,CAAC;IACD,IAAI,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5C,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,gBAAgB,CAAC,KAAK,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACxI,CAAC;IACD,MAAM,SAAS,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;IACjH,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,KAAK,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IAEpH,QAAQ,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAExC,IAAI,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC7B,MAAM,SAAS,GAAG,IAAI,KAAK,CAAC;YAC1B,IAAI,EAAE;gBACJ,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC;gBACd,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC;gBACnB,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC;gBACd,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC;gBACb,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC;gBACb,CAAC,CAAC,GAAG,CAAC,WAAW,CAAC;gBAClB,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC;gBACjB,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC;gBACd,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC;gBACjB,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC;aAChB;YACD,KAAK,EAAE,UAAU,CAAC,EAAE,CAAC;YACrB,KAAK,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE;YAC/B,SAAS,EAAE,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,CAAC;SACnG,CAAC,CAAC;QAEH,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;YAClC,SAAS,CAAC,IAAI,CAAC;gBACb,GAAG,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,IAAI,EAAE;gBAC/B,KAAK,CAAC,UAAU;gBAChB,KAAK,CAAC,SAAS,IAAI,EAAE;gBACrB,YAAY,CAAC,KAAK,CAAC,SAAS,CAAC;gBAC7B,YAAY,CAAC,KAAK,CAAC,SAAS,CAAC;gBAC7B,YAAY,CAAC,KAAK,CAAC,iBAAiB,CAAC;gBACrC,YAAY,CAAC,KAAK,CAAC,gBAAgB,CAAC;gBACpC,YAAY,CAAC,KAAK,CAAC,kBAAkB,CAAC;gBACtC,YAAY,CAAC,KAAK,CAAC,gBAAgB,CAAC;gBACpC,YAAY,CAAC,KAAK,CAAC,iBAAiB,CAAC;aACtC,CAAC,CAAC;QACL,CAAC;QAED,MAAM,QAAQ,GAAG,SAAS,CAAC,QAAQ,EAAE,CAAC;QACtC,MAAM,KAAK,GAAG,SAAS,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;QAC9D,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,iBAAiB,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC,CAAC;IACzF,CAAC;IAED,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;QAClC,QAAQ,CAAC,IAAI,CAAC,kBAAkB,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,CAAC;IAC/C,CAAC;IAED,OAAO,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;AACtC,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAmB,EAAE,EAAE,GAAG,KAAK;IACzD,MAAM,CAAC,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC;IACtB,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,MAAM,KAAK,GAAG,UAAU,KAAK,CAAC,IAAI,KAAK,KAAK,CAAC,UAAU,GAAG,CAAC;IAC3D,MAAM,OAAO,GAAG;QACd,GAAG,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,SAAS,IAAI,SAAS,EAAE;QACpD,GAAG,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,YAAY,CAAC,KAAK,CAAC,SAAS,CAAC,WAAW,YAAY,CAAC,KAAK,CAAC,SAAS,CAAC,UAAU,YAAY,CAAC,KAAK,CAAC,yBAAyB,CAAC,SAAS;QAC1J,GAAG,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,KAAK,YAAY,CAAC,KAAK,CAAC,gBAAgB,CAAC,eAAe,YAAY,CAAC,KAAK,CAAC,gBAAgB,CAAC,eAAe,YAAY,CAAC,KAAK,CAAC,kBAAkB,CAAC,iBAAiB,YAAY,CAAC,KAAK,CAAC,gBAAgB,CAAC,SAAS;QAClO,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,YAAY,CAAC,KAAK,CAAC,cAAc,CAAC,kBAAkB,YAAY,CAAC,KAAK,CAAC,iBAAiB,CAAC,6BAA6B,KAAK,CAAC,aAAa,IAAI,MAAM,gBAAgB,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE;QAC1N,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,KAAK,CAAC,eAAe,CAAC,SAAS,IAAI,MAAM,KAAK,KAAK,CAAC,eAAe,CAAC,IAAI,GAAG;QAClG,GAAG,CAAC,CAAC,GAAG,CAAC,gBAAgB,CAAC,YAAY,YAAY,CAAC,KAAK,CAAC,mBAAmB,CAAC,WAAW,YAAY,CAAC,KAAK,CAAC,iBAAiB,CAAC,EAAE;KAChI,CAAC;IACF,IAAI,KAAK,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,eAAe,CAAC,KAAK,KAAK,CAAC,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACvH,CAAC;IACD,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAE/B,IAAI,KAAK,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpC,KAAK,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,MAAM,CAAC,OAAO,EAAE,KAAK,CAAC,MAAM,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC,CAAC;IAClF,CAAC;SAAM,CAAC;QACN,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,0BAA0B,CAAC,CAAC,CAAC;IAChD,CAAC;IAED,IAAI,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACjC,KAAK,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,MAAM,CAAC,aAAa,EAAE,KAAK,CAAC,MAAM,CAAC,gBAAgB,EAAE,EAAE,CAAC,CAAC,CAAC;IAClH,CAAC;SAAM,CAAC;QACN,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,uBAAuB,CAAC,CAAC,CAAC;IAC7C,CAAC;IAED,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,KAAK,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IAC5G,OAAO,GAAG,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;AAC/E,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAwB,EAAE,SAAkB,EAAE,EAAE,GAAG,KAAK;IACjF,MAAM,CAAC,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC;IACtB,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC;QACtB,IAAI,EAAE;YACJ,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC;YACZ,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC;YACb,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC;YACb,CAAC,CAAC,GAAG,CAAC,WAAW,CAAC;YAClB,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC;YACjB,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC;SAClB;QACD,KAAK,EAAE,UAAU,CAAC,EAAE,CAAC;QACrB,KAAK,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE;QAC/B,SAAS,EAAE,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,CAAC;KAC9D,CAAC,CAAC;IAEH,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,KAAK,CAAC,IAAI,CAAC;YACT,GAAG,MAAM,CAAC,MAAM,KAAK,MAAM,CAAC,KAAK,GAAG;YACpC,MAAM,CAAC,IAAI;YACX,MAAM,CAAC,YAAY;YACnB,YAAY,CAAC,MAAM,CAAC,aAAa,CAAC;YAClC,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,aAAa,GAAG,GAAG,CAAC,GAAG;YAC5C,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;SAC5B,CAAC,CAAC;IACL,CAAC;IAED,MAAM,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,uBAAuB,CAAC,CAAC,CAAC,EAAE,CAAC;IACxD,OAAO,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,KAAK,CAAC,QAAQ,EAAE,GAAG,MAAM,EAAE,CAAC;AACzE,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAiB,EAAE,aAAsB,EAAE,gBAAyB,EAAE,EAAE,GAAG,KAAK;IACzG,MAAM,CAAC,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC;IACtB,MAAM,WAAW,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,EAAE,CAAC,CAAC;IACvD,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC;QACtB,IAAI,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC;QAC7D,KAAK,EAAE,UAAU,CAAC,EAAE,CAAC;QACrB,KAAK,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE;KAChC,CAAC,CAAC;IAEH,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;IAC1F,CAAC;IAED,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,aAAa;QAAE,KAAK,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;IAChD,IAAI,gBAAgB;QAAE,KAAK,CAAC,IAAI,CAAC,mBAAmB,CAAC,CAAC;IACtD,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;IACjE,OAAO,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,KAAK,CAAC,QAAQ,EAAE,GAAG,MAAM,EAAE,CAAC;AACzE,CAAC"}

package/dist/src/sheet/types.d.ts

@@ -0,0 +1,117 @@
+export type SheetVisibility = 'visible' | 'hidden' | 'very_hidden';
+export type HeaderSelectionMode = 'auto' | 'none' | 'explicit';
+export type ColumnValueType = 'string' | 'number' | 'boolean' | 'date' | 'error' | 'blank_stub' | 'unknown';
+export interface InspectSheetOptions {
+    sheet?: string;
+    sampleRows: number;
+    headerRow: 'auto' | 'none' | number;
+    maxColumns: number;
+}
+export interface DefinedNameInfo {
+    name: string;
+    ref: string;
+    scope: 'workbook' | 'sheet';
+    sheetIndex?: number;
+    sheetName?: string;
+    external: boolean;
+}
+export interface WorkbookProperties {
+    title?: string;
+    subject?: string;
+    author?: string;
+    company?: string;
+    manager?: string;
+    createdDate?: string;
+    modifiedDate?: string;
+}
+export interface WorkbookRiskFlags {
+    hiddenSheets: boolean;
+    formulas: boolean;
+    comments: boolean;
+    hyperlinks: boolean;
+    mergedCells: boolean;
+    protectedSheets: boolean;
+    externalFormulaRefs: boolean;
+}
+export interface CellTypeCounts {
+    string: number;
+    number: number;
+    boolean: number;
+    date: number;
+    error: number;
+    blankStub: number;
+}
+export interface ColumnProfile {
+    index: number;
+    letter: string;
+    name: string;
+    dominantType: ColumnValueType;
+    nonEmptyCount: number;
+    nonEmptyRatio: number;
+    examples: string[];
+}
+export interface SampleRow {
+    rowNumber: number;
+    values: Record<string, string>;
+}
+export interface SheetProfile {
+    index: number;
+    name: string;
+    visibility: SheetVisibility;
+    usedRange: string | null;
+    totalRows: number;
+    totalCols: number;
+    rectangularRangeCellCount: number;
+    nonEmptyCellCount: number;
+    cellTypeCounts: CellTypeCounts;
+    formulaCellCount: number;
+    commentCellCount: number;
+    hyperlinkCellCount: number;
+    mergedRangeCount: number;
+    hiddenRowCount: number;
+    hiddenColumnCount: number;
+    autoFilterRef: string | null;
+    protected: boolean;
+    definedNames: DefinedNameInfo[];
+    externalFormulaRefCount: number;
+    headerSelection: {
+        requested: 'auto' | 'none' | number;
+        mode: HeaderSelectionMode;
+        rowNumber: number | null;
+    };
+    schema: {
+        truncated: boolean;
+        columns: ColumnProfile[];
+    };
+    sample: {
+        truncatedRows: boolean;
+        truncatedColumns: boolean;
+        rows: SampleRow[];
+    };
+    sampleTokenEstimate: number;
+    fullTokenEstimate: number;
+    estimateMethod: 'full_scan';
+}
+export interface WorkbookInspection {
+    file: string;
+    format: 'xlsx';
+    size: number;
+    properties: WorkbookProperties;
+    customPropertyCount: number;
+    customProperties: Record<string, unknown>;
+    sheetCount: number;
+    visibleSheetCount: number;
+    hiddenSheetCount: number;
+    veryHiddenSheetCount: number;
+    definedNames: DefinedNameInfo[];
+    riskFlags: WorkbookRiskFlags;
+}
+export interface SheetInspectionResult {
+    workbook: WorkbookInspection;
+    sheets: SheetProfile[];
+}
+export interface SheetInspectPayload {
+    file: string;
+    query: Record<string, unknown>;
+    results: SheetInspectionResult;
+}

package/dist/src/sheet/types.js

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=types.js.map

package/dist/src/sheet/types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.js","sourceRoot":"","sources":["../../../src/sheet/types.ts"],"names":[],"mappings":""}

package/dist/src/slide/command.d.ts

@@ -0,0 +1,2 @@
+import { Command } from 'commander';
+export declare function registerSlideCommands(program: Command): void;

package/dist/src/slide/command.js

@@ -0,0 +1,54 @@
+import { writeFile } from 'node:fs/promises';
+import path from 'node:path';
+import { inspectPresentation } from './inspect.js';
+import { formatSlideInspection, formatSlidePayloadJson } from './output.js';
+import { createInspectPayload } from '../inspect/shared.js';
+import { writeStream } from '../utils.js';
+function getOptions(command) {
+    return command.optsWithGlobals();
+}
+function parsePositiveInt(value, fallback, label) {
+    if (!value)
+        return fallback;
+    const parsed = Number.parseInt(value, 10);
+    if (!Number.isFinite(parsed) || parsed <= 0) {
+        throw new Error(`Invalid ${label}: "${value}"`);
+    }
+    return parsed;
+}
+async function emit(output, options) {
+    if (options.output) {
+        await writeFile(options.output, output);
+        return;
+    }
+    await writeStream(process.stdout, output);
+}
+export function registerSlideCommands(program) {
+    const slide = program.command('slide').description('Inspect presentation structure and content before deep reading');
+    const inspect = slide.command('inspect <file>').description('Inspect PPTX or ODP presentation metadata, slides, and content preview');
+    inspect
+        .option('--format <type>', 'output format: tabular or json', 'tabular')
+        .option('-o, --output <file>', 'write output to file')
+        .option('--ci', 'ASCII-only output, no colors')
+        .option('--sample-slides <n>', 'preview slide count', '3')
+        .option('--slide <number>', 'inspect specific slide (1-based index)');
+    inspect.action(async (file, _options, command) => {
+        const options = getOptions(command);
+        const resolved = path.resolve(file);
+        const inspectOptions = {
+            sampleSlides: parsePositiveInt(options.sampleSlides, 3, 'sample slide count'),
+            slide: options.slide ? parsePositiveInt(options.slide, 1, 'slide number') : undefined,
+        };
+        const result = await inspectPresentation(resolved, inspectOptions);
+        const payload = createInspectPayload(resolved, {
+            command: 'slide.inspect',
+            sampleSlides: inspectOptions.sampleSlides,
+            slide: inspectOptions.slide,
+        }, result);
+        const output = options.format === 'json'
+            ? formatSlidePayloadJson(payload)
+            : formatSlideInspection(result, options.ci);
+        await emit(output.endsWith('\n') ? output : `${output}\n`, options);
+    });
+}
+//# sourceMappingURL=command.js.map

package/dist/src/slide/command.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"command.js","sourceRoot":"","sources":["../../../src/slide/command.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAC7C,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B,OAAO,EAAE,mBAAmB,EAAE,MAAM,cAAc,CAAC;AACnD,OAAO,EAAE,qBAAqB,EAAE,sBAAsB,EAAE,MAAM,aAAa,CAAC;AAC5E,OAAO,EAAE,oBAAoB,EAAE,MAAM,sBAAsB,CAAC;AAC5D,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAW1C,SAAS,UAAU,CAAC,OAAgB;IAClC,OAAO,OAAO,CAAC,eAAe,EAAyB,CAAC;AAC1D,CAAC;AAED,SAAS,gBAAgB,CAAC,KAAyB,EAAE,QAAgB,EAAE,KAAa;IAClF,IAAI,CAAC,KAAK;QAAE,OAAO,QAAQ,CAAC;IAC5B,MAAM,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IAC1C,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,MAAM,IAAI,CAAC,EAAE,CAAC;QAC5C,MAAM,IAAI,KAAK,CAAC,WAAW,KAAK,MAAM,KAAK,GAAG,CAAC,CAAC;IAClD,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,KAAK,UAAU,IAAI,CAAC,MAAc,EAAE,OAA4B;IAC9D,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;QACnB,MAAM,SAAS,CAAC,OAAO,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QACxC,OAAO;IACT,CAAC;IACD,MAAM,WAAW,CAAC,OAAO,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;AAC5C,CAAC;AAED,MAAM,UAAU,qBAAqB,CAAC,OAAgB;IACpD,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,WAAW,CAAC,gEAAgE,CAAC,CAAC;IACrH,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC,WAAW,CAAC,wEAAwE,CAAC,CAAC;IAEtI,OAAO;SACJ,MAAM,CAAC,iBAAiB,EAAE,gCAAgC,EAAE,SAAS,CAAC;SACtE,MAAM,CAAC,qBAAqB,EAAE,sBAAsB,CAAC;SACrD,MAAM,CAAC,MAAM,EAAE,8BAA8B,CAAC;SAC9C,MAAM,CAAC,qBAAqB,EAAE,qBAAqB,EAAE,GAAG,CAAC;SACzD,MAAM,CAAC,kBAAkB,EAAE,wCAAwC,CAAC,CAAC;IAExE,OAAO,CAAC,MAAM,CAAC,KAAK,EAAE,IAAY,EAAE,QAA6B,EAAE,OAAgB,EAAE,EAAE;QACrF,MAAM,OAAO,GAAG,UAAU,CAAC,OAAO,CAAC,CAAC;QACpC,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACpC,MAAM,cAAc,GAAwB;YAC1C,YAAY,EAAE,gBAAgB,CAAC,OAAO,CAAC,YAAY,EAAE,CAAC,EAAE,oBAAoB,CAAC;YAC7E,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,gBAAgB,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC,EAAE,cAAc,CAAC,CAAC,CAAC,CAAC,SAAS;SACtF,CAAC;QACF,MAAM,MAAM,GAAG,MAAM,mBAAmB,CAAC,QAAQ,EAAE,cAAc,CAAC,CAAC;QACnE,MAAM,OAAO,GAAG,oBAAoB,CAAgB,QAAQ,EAAE;YAC5D,OAAO,EAAE,eAAe;YACxB,YAAY,EAAE,cAAc,CAAC,YAAY;YACzC,KAAK,EAAE,cAAc,CAAC,KAAK;SAC5B,EAAE,MAAM,CAAwB,CAAC;QAClC,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,KAAK,MAAM;YACtC,CAAC,CAAC,sBAAsB,CAAC,OAAO,CAAC;YACjC,CAAC,CAAC,qBAAqB,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,CAAC,CAAC;QAC9C,MAAM,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,MAAM,IAAI,EAAE,OAAO,CAAC,CAAC;IACtE,CAAC,CAAC,CAAC;AACL,CAAC"}

package/dist/src/slide/inspect-odp.d.ts

@@ -0,0 +1,7 @@
+import type { DocumentProperties } from '../inspect/shared.js';
+import type { SlideRiskFlags, SlideProfile } from './types.js';
+export declare function inspectOdp(buffer: Buffer): Promise<{
+    properties: DocumentProperties;
+    riskFlags: SlideRiskFlags;
+    slideProfiles: SlideProfile[];
+}>;

package/dist/src/slide/inspect-odp.js

@@ -0,0 +1,105 @@
+import JSZip from 'jszip';
+import { countWords } from '../utils.js';
+import { asOptionalString, formatDateLike } from '../inspect/shared.js';
+function parseXmlText(xml, tag) {
+    const regex = new RegExp(`<${tag}[^>]*>([^<]*)</${tag}>`, 'i');
+    const match = xml.match(regex);
+    return match?.[1]?.trim() || undefined;
+}
+function extractTextFromElement(xml) {
+    const matches = xml.match(/<text:(?:p|span)[^>]*>([^<]*)<\/text:(?:p|span)>/g) || [];
+    const texts = [];
+    for (const m of matches) {
+        const text = m.replace(/<[^>]+>/g, '').trim();
+        if (text)
+            texts.push(text);
+    }
+    return texts.join(' ');
+}
+export async function inspectOdp(buffer) {
+    const zip = await JSZip.loadAsync(buffer);
+    // Parse metadata
+    const properties = {};
+    const metaXml = await zip.file('meta.xml')?.async('text');
+    if (metaXml) {
+        properties.title = asOptionalString(parseXmlText(metaXml, 'dc:title'));
+        properties.subject = asOptionalString(parseXmlText(metaXml, 'dc:subject'));
+        properties.author = asOptionalString(parseXmlText(metaXml, 'dc:creator') ?? parseXmlText(metaXml, 'meta:initial-creator'));
+        properties.keywords = asOptionalString(parseXmlText(metaXml, 'meta:keyword'));
+        properties.createdDate = formatDateLike(parseXmlText(metaXml, 'meta:creation-date'));
+        properties.modifiedDate = formatDateLike(parseXmlText(metaXml, 'dc:date'));
+    }
+    // Parse content.xml for slides
+    const contentXml = await zip.file('content.xml')?.async('text');
+    const slideProfiles = [];
+    const riskFlags = {
+        comments: false,
+        speakerNotes: false,
+        hyperlinks: false,
+        embeddedMedia: false,
+        animations: false,
+        macros: false,
+        charts: false,
+        tables: false,
+    };
+    if (contentXml) {
+        // Split by draw:page elements
+        const slideMatches = contentXml.match(/<draw:page\b[^>]*>[\s\S]*?<\/draw:page>/g) || [];
+        for (let i = 0; i < slideMatches.length; i++) {
+            const slideXml = slideMatches[i];
+            // Get title from draw:name attribute or first text frame
+            let title = null;
+            const nameMatch = slideXml.match(/draw:name="([^"]*)"/);
+            if (nameMatch && nameMatch[1]) {
+                title = nameMatch[1];
+            }
+            // Extract text
+            const text = extractTextFromElement(slideXml);
+            const words = countWords(text);
+            // Count tables and images
+            const tables = (slideXml.match(/<table:table\b/g) || []).length;
+            const images = (slideXml.match(/<draw:image\b/g) || []).length;
+            // Check for notes
+            let hasNotes = false;
+            let notePreview = null;
+            const notesMatch = slideXml.match(/<presentation:notes\b[^>]*>([\s\S]*?)<\/presentation:notes>/);
+            if (notesMatch) {
+                const notesText = extractTextFromElement(notesMatch[1]);
+                if (notesText.trim()) {
+                    hasNotes = true;
+                    notePreview = notesText.length > 200 ? notesText.slice(0, 200) + '...' : notesText;
+                }
+            }
+            slideProfiles.push({
+                index: i + 1,
+                title,
+                words,
+                hasNotes,
+                notePreview,
+                images,
+                tables,
+                charts: 0,
+            });
+        }
+        // Risk flags
+        if (/<office:annotation\b/.test(contentXml))
+            riskFlags.comments = true;
+        if (slideProfiles.some(p => p.hasNotes))
+            riskFlags.speakerNotes = true;
+        if (/<text:a\b[^>]*xlink:href=/.test(contentXml))
+            riskFlags.hyperlinks = true;
+        if (/<draw:image\b/.test(contentXml))
+            riskFlags.embeddedMedia = true;
+        if (/<presentation:transition\b/.test(contentXml) || /<anim:/.test(contentXml))
+            riskFlags.animations = true;
+        if (/<table:table\b/.test(contentXml))
+            riskFlags.tables = true;
+        if (/<draw:object\b/.test(contentXml))
+            riskFlags.charts = true;
+    }
+    // Check for macros
+    if (zip.file('Basic/') || zip.file('Scripts/'))
+        riskFlags.macros = true;
+    return { properties, riskFlags, slideProfiles };
+}
+//# sourceMappingURL=inspect-odp.js.map

package/dist/src/slide/inspect-odp.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"inspect-odp.js","sourceRoot":"","sources":["../../../src/slide/inspect-odp.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAIxE,SAAS,YAAY,CAAC,GAAW,EAAE,GAAW;IAC5C,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,IAAI,GAAG,kBAAkB,GAAG,GAAG,EAAE,GAAG,CAAC,CAAC;IAC/D,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;IAC/B,OAAO,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,SAAS,CAAC;AACzC,CAAC;AAED,SAAS,sBAAsB,CAAC,GAAW;IACzC,MAAM,OAAO,GAAG,GAAG,CAAC,KAAK,CAAC,mDAAmD,CAAC,IAAI,EAAE,CAAC;IACrF,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;QACxB,MAAM,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAC9C,IAAI,IAAI;YAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC7B,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,UAAU,CAC9B,MAAc;IAMd,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;IAE1C,iBAAiB;IACjB,MAAM,UAAU,GAAuB,EAAE,CAAC;IAC1C,MAAM,OAAO,GAAG,MAAM,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IAC1D,IAAI,OAAO,EAAE,CAAC;QACZ,UAAU,CAAC,KAAK,GAAG,gBAAgB,CAAC,YAAY,CAAC,OAAO,EAAE,UAAU,CAAC,CAAC,CAAC;QACvE,UAAU,CAAC,OAAO,GAAG,gBAAgB,CAAC,YAAY,CAAC,OAAO,EAAE,YAAY,CAAC,CAAC,CAAC;QAC3E,UAAU,CAAC,MAAM,GAAG,gBAAgB,CAAC,YAAY,CAAC,OAAO,EAAE,YAAY,CAAC,IAAI,YAAY,CAAC,OAAO,EAAE,sBAAsB,CAAC,CAAC,CAAC;QAC3H,UAAU,CAAC,QAAQ,GAAG,gBAAgB,CAAC,YAAY,CAAC,OAAO,EAAE,cAAc,CAAC,CAAC,CAAC;QAC9E,UAAU,CAAC,WAAW,GAAG,cAAc,CAAC,YAAY,CAAC,OAAO,EAAE,oBAAoB,CAAC,CAAC,CAAC;QACrF,UAAU,CAAC,YAAY,GAAG,cAAc,CAAC,YAAY,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC,CAAC;IAC7E,CAAC;IAED,+BAA+B;IAC/B,MAAM,UAAU,GAAG,MAAM,GAAG,CAAC,IAAI,CAAC,aAAa,CAAC,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IAChE,MAAM,aAAa,GAAmB,EAAE,CAAC;IAEzC,MAAM,SAAS,GAAmB;QAChC,QAAQ,EAAE,KAAK;QACf,YAAY,EAAE,KAAK;QACnB,UAAU,EAAE,KAAK;QACjB,aAAa,EAAE,KAAK;QACpB,UAAU,EAAE,KAAK;QACjB,MAAM,EAAE,KAAK;QACb,MAAM,EAAE,KAAK;QACb,MAAM,EAAE,KAAK;KACd,CAAC;IAEF,IAAI,UAAU,EAAE,CAAC;QACf,8BAA8B;QAC9B,MAAM,YAAY,GAAG,UAAU,CAAC,KAAK,CAAC,0CAA0C,CAAC,IAAI,EAAE,CAAC;QAExF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,QAAQ,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC;YAEjC,yDAAyD;YACzD,IAAI,KAAK,GAAkB,IAAI,CAAC;YAChC,MAAM,SAAS,GAAG,QAAQ,CAAC,KAAK,CAAC,qBAAqB,CAAC,CAAC;YACxD,IAAI,SAAS,IAAI,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC;gBAC9B,KAAK,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;YACvB,CAAC;YAED,eAAe;YACf,MAAM,IAAI,GAAG,sBAAsB,CAAC,QAAQ,CAAC,CAAC;YAC9C,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC;YAE/B,0BAA0B;YAC1B,MAAM,MAAM,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,iBAAiB,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;YAChE,MAAM,MAAM,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,gBAAgB,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;YAE/D,kBAAkB;YAClB,IAAI,QAAQ,GAAG,KAAK,CAAC;YACrB,IAAI,WAAW,GAAkB,IAAI,CAAC;YACtC,MAAM,UAAU,GAAG,QAAQ,CAAC,KAAK,CAAC,6DAA6D,CAAC,CAAC;YACjG,IAAI,UAAU,EAAE,CAAC;gBACf,MAAM,SAAS,GAAG,sBAAsB,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;gBACxD,IAAI,SAAS,CAAC,IAAI,EAAE,EAAE,CAAC;oBACrB,QAAQ,GAAG,IAAI,CAAC;oBAChB,WAAW,GAAG,SAAS,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC;gBACrF,CAAC;YACH,CAAC;YAED,aAAa,CAAC,IAAI,CAAC;gBACjB,KAAK,EAAE,CAAC,GAAG,CAAC;gBACZ,KAAK;gBACL,KAAK;gBACL,QAAQ;gBACR,WAAW;gBACX,MAAM;gBACN,MAAM;gBACN,MAAM,EAAE,CAAC;aACV,CAAC,CAAC;QACL,CAAC;QAED,aAAa;QACb,IAAI,sBAAsB,CAAC,IAAI,CAAC,UAAU,CAAC;YAAE,SAAS,CAAC,QAAQ,GAAG,IAAI,CAAC;QACvE,IAAI,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC;YAAE,SAAS,CAAC,YAAY,GAAG,IAAI,CAAC;QACvE,IAAI,2BAA2B,CAAC,IAAI,CAAC,UAAU,CAAC;YAAE,SAAS,CAAC,UAAU,GAAG,IAAI,CAAC;QAC9E,IAAI,eAAe,CAAC,IAAI,CAAC,UAAU,CAAC;YAAE,SAAS,CAAC,aAAa,GAAG,IAAI,CAAC;QACrE,IAAI,4BAA4B,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,QAAQ,CAAC,IAAI,CAAC,UAAU,CAAC;YAAE,SAAS,CAAC,UAAU,GAAG,IAAI,CAAC;QAC5G,IAAI,gBAAgB,CAAC,IAAI,CAAC,UAAU,CAAC;YAAE,SAAS,CAAC,MAAM,GAAG,IAAI,CAAC;QAC/D,IAAI,gBAAgB,CAAC,IAAI,CAAC,UAAU,CAAC;YAAE,SAAS,CAAC,MAAM,GAAG,IAAI,CAAC;IACjE,CAAC;IAED,mBAAmB;IACnB,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC;QAAE,SAAS,CAAC,MAAM,GAAG,IAAI,CAAC;IAExE,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,aAAa,EAAE,CAAC;AAClD,CAAC"}

package/dist/src/slide/inspect-pptx.d.ts

@@ -0,0 +1,11 @@
+import JSZip from 'jszip';
+import type { DocumentProperties } from '../inspect/shared.js';
+import type { SlideRiskFlags, SlideProfile } from './types.js';
+export declare function extractPptxProperties(zip: JSZip): Promise<DocumentProperties>;
+export declare function extractPptxRiskFlags(zip: JSZip, slideProfiles: SlideProfile[]): Promise<SlideRiskFlags>;
+export declare function extractPptxSlideProfiles(zip: JSZip): Promise<SlideProfile[]>;
+export declare function inspectPptx(buffer: Buffer): Promise<{
+    properties: DocumentProperties;
+    riskFlags: SlideRiskFlags;
+    slideProfiles: SlideProfile[];
+}>;

package/dist/src/slide/inspect-pptx.js

@@ -0,0 +1,157 @@
+import JSZip from 'jszip';
+import { countWords } from '../utils.js';
+import { asOptionalString, formatDateLike } from '../inspect/shared.js';
+function parseXmlText(xml, tag) {
+    const regex = new RegExp(`<${tag}[^>]*>([^<]*)</${tag}>`, 'i');
+    const match = xml.match(regex);
+    return match?.[1]?.trim() || undefined;
+}
+function extractAllText(xml) {
+    const matches = xml.match(/<a:t>([^<]*)<\/a:t>/g) || [];
+    return matches.map(m => {
+        const text = m.match(/<a:t>([^<]*)<\/a:t>/);
+        return text?.[1] || '';
+    }).join(' ').trim();
+}
+export async function extractPptxProperties(zip) {
+    const props = {};
+    const coreXml = await zip.file('docProps/core.xml')?.async('text');
+    if (coreXml) {
+        props.title = asOptionalString(parseXmlText(coreXml, 'dc:title'));
+        props.subject = asOptionalString(parseXmlText(coreXml, 'dc:subject'));
+        props.author = asOptionalString(parseXmlText(coreXml, 'dc:creator'));
+        props.keywords = asOptionalString(parseXmlText(coreXml, 'cp:keywords'));
+        props.lastModifiedBy = asOptionalString(parseXmlText(coreXml, 'cp:lastModifiedBy'));
+        props.createdDate = formatDateLike(parseXmlText(coreXml, 'dcterms:created'));
+        props.modifiedDate = formatDateLike(parseXmlText(coreXml, 'dcterms:modified'));
+    }
+    const appXml = await zip.file('docProps/app.xml')?.async('text');
+    if (appXml) {
+        props.company = asOptionalString(parseXmlText(appXml, 'Company'));
+    }
+    return props;
+}
+export async function extractPptxRiskFlags(zip, slideProfiles) {
+    const flags = {
+        comments: false,
+        speakerNotes: false,
+        hyperlinks: false,
+        embeddedMedia: false,
+        animations: false,
+        macros: false,
+        charts: false,
+        tables: false,
+    };
+    // Check for comments folder
+    const commentFiles = Object.keys(zip.files).filter(name => /^ppt\/comments\//.test(name));
+    if (commentFiles.length > 0)
+        flags.comments = true;
+    // Check for speaker notes
+    if (slideProfiles.some(p => p.hasNotes))
+        flags.speakerNotes = true;
+    // Check for media files
+    const mediaFiles = Object.keys(zip.files).filter(name => /^ppt\/media\//.test(name));
+    if (mediaFiles.length > 0)
+        flags.embeddedMedia = true;
+    // Check for macros
+    if (zip.file('ppt/vbaProject.bin'))
+        flags.macros = true;
+    // Check for charts
+    const chartFiles = Object.keys(zip.files).filter(name => /^ppt\/charts\//.test(name));
+    if (chartFiles.length > 0)
+        flags.charts = true;
+    // Aggregate from profiles
+    if (slideProfiles.some(p => p.tables > 0))
+        flags.tables = true;
+    // Check slides for animations and hyperlinks
+    const slideFiles = Object.keys(zip.files)
+        .filter(name => /^ppt\/slides\/slide\d+\.xml$/.test(name))
+        .sort((a, b) => {
+        const numA = parseInt(a.match(/slide(\d+)/)?.[1] || '0', 10);
+        const numB = parseInt(b.match(/slide(\d+)/)?.[1] || '0', 10);
+        return numA - numB;
+    });
+    for (const slideFile of slideFiles) {
+        const xml = await zip.file(slideFile)?.async('text');
+        if (!xml)
+            continue;
+        if (/<p:transition\b/.test(xml) || /<p:anim\b/.test(xml) || /<p:animMotion\b/.test(xml)) {
+            flags.animations = true;
+        }
+        if (/<a:hlinkClick\b/.test(xml) || /<a:hlinkMouseOver\b/.test(xml)) {
+            flags.hyperlinks = true;
+        }
+        if (flags.animations && flags.hyperlinks)
+            break;
+    }
+    return flags;
+}
+export async function extractPptxSlideProfiles(zip) {
+    const slideFiles = Object.keys(zip.files)
+        .filter(name => /^ppt\/slides\/slide\d+\.xml$/.test(name))
+        .sort((a, b) => {
+        const numA = parseInt(a.match(/slide(\d+)/)?.[1] || '0', 10);
+        const numB = parseInt(b.match(/slide(\d+)/)?.[1] || '0', 10);
+        return numA - numB;
+    });
+    const profiles = [];
+    for (let i = 0; i < slideFiles.length; i++) {
+        const slideXml = await zip.file(slideFiles[i])?.async('text');
+        if (!slideXml)
+            continue;
+        // Extract title
+        let title = null;
+        // Look for title placeholder
+        const titleMatch = slideXml.match(/<p:sp>[\s\S]*?<p:ph[^>]*type="(?:title|ctrTitle)"[^>]*\/>[\s\S]*?<\/p:sp>/i);
+        if (titleMatch) {
+            const titleText = extractAllText(titleMatch[0]);
+            if (titleText)
+                title = titleText;
+        }
+        // Extract all text for word count
+        const allText = extractAllText(slideXml);
+        const words = countWords(allText);
+        // Count tables
+        const tables = (slideXml.match(/<a:tbl\b/g) || []).length;
+        // Count images
+        const images = (slideXml.match(/<p:pic\b/g) || []).length + (slideXml.match(/<a:blip\b/g) || []).length;
+        // Count charts (from relationships)
+        const slideNum = parseInt(slideFiles[i].match(/slide(\d+)/)?.[1] || '0', 10);
+        const relsFile = `ppt/slides/_rels/slide${slideNum}.xml.rels`;
+        const relsXml = await zip.file(relsFile)?.async('text');
+        const charts = relsXml ? (relsXml.match(/chart/gi) || []).length : 0;
+        // Check for speaker notes
+        const notesFile = `ppt/notesSlides/notesSlide${slideNum}.xml`;
+        const notesXml = await zip.file(notesFile)?.async('text');
+        let hasNotes = false;
+        let notePreview = null;
+        if (notesXml) {
+            const notesText = extractAllText(notesXml);
+            // Filter out slide number placeholder text
+            const cleanNotes = notesText.replace(/\d+/g, '').trim();
+            if (cleanNotes.length > 0) {
+                hasNotes = true;
+                notePreview = notesText.length > 200 ? notesText.slice(0, 200) + '...' : notesText;
+            }
+        }
+        profiles.push({
+            index: i + 1,
+            title,
+            words,
+            hasNotes,
+            notePreview,
+            images,
+            tables,
+            charts,
+        });
+    }
+    return profiles;
+}
+export async function inspectPptx(buffer) {
+    const zip = await JSZip.loadAsync(buffer);
+    const properties = await extractPptxProperties(zip);
+    const slideProfiles = await extractPptxSlideProfiles(zip);
+    const riskFlags = await extractPptxRiskFlags(zip, slideProfiles);
+    return { properties, riskFlags, slideProfiles };
+}
+//# sourceMappingURL=inspect-pptx.js.map