@statewalker/content-extractors 0.2.2 → 0.3.0

package/dist/extractors/pdf-extractor.js

@@ -1,4 +1,5 @@
 import { collectBytes } from "../collect-bytes.js";
+import { DEFAULT_MAX_BYTES } from "../limits.js";
 import { extractText } from "unpdf";
 //#region src/extractors/pdf-extractor.ts
 /**
@@ -6,17 +7,22 @@ import { extractText } from "unpdf";
 * section with a `## Page N` header. This preserves page boundaries in
 * the indexed content, making search results more precise.
 */
-const pdfExtractor = async (content) => {
-	const bytes = await collectBytes(content);
-	const { text: pages } = await extractText(bytes);
-	const sections = [];
-	for (let i = 0; i < pages.length; i++) {
-		const pageText = pages[i]?.trim();
-		if (pageText) sections.push(`## Page ${i + 1}\n\n${pageText}`);
-	}
-	return sections.join("\n\n");
-};
+function createPdfExtractor(limits = {}) {
+	const { maxBytes = DEFAULT_MAX_BYTES } = limits;
+	return async (content) => {
+		const bytes = await collectBytes(content, { maxBytes });
+		const { text: pages } = await extractText(new Uint8Array(bytes));
+		const sections = [];
+		for (let i = 0; i < pages.length; i++) {
+			const pageText = pages[i]?.trim();
+			if (pageText) sections.push(`## Page ${i + 1}\n\n${pageText}`);
+		}
+		return sections.join("\n\n");
+	};
+}
+/** PDF extractor with the default limits. */
+const pdfExtractor = createPdfExtractor();
 //#endregion
-export { pdfExtractor };
+export { createPdfExtractor, pdfExtractor };
 
 //# sourceMappingURL=pdf-extractor.js.map

package/dist/extractors/pdf-extractor.js.map

@@ -1 +1 @@
-{"version":3,"file":"pdf-extractor.js","names":[],"sources":["../../src/extractors/pdf-extractor.ts"],"sourcesContent":["import { extractText } from \"unpdf\";\nimport { collectBytes } from \"../collect-bytes.js\";\nimport type { ContentExtractor } from \"../types.js\";\n\n/**\n * Extracts PDF content page-by-page, formatting each page as a markdown\n * section with a `## Page N` header. This preserves page boundaries in\n * the indexed content, making search results more precise.\n */\nexport const pdfExtractor: ContentExtractor = async (content) => {\n  const bytes = await collectBytes(content);\n  const { text: pages } = await extractText(bytes);\n\n  const sections: string[] = [];\n  for (let i = 0; i < pages.length; i++) {\n    const pageText = pages[i]?.trim();\n    if (pageText) {\n      sections.push(`## Page ${i + 1}\\n\\n${pageText}`);\n    }\n  }\n\n  return sections.join(\"\\n\\n\");\n};\n"],"mappings":";;;;;;;;AASA,MAAa,eAAiC,OAAO,YAAY;CAC/D,MAAM,QAAQ,MAAM,aAAa,OAAO;CACxC,MAAM,EAAE,MAAM,UAAU,MAAM,YAAY,KAAK;CAE/C,MAAM,WAAqB,CAAC;CAC5B,KAAK,IAAI,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK;EACrC,MAAM,WAAW,MAAM,EAAE,EAAE,KAAK;EAChC,IAAI,UACF,SAAS,KAAK,WAAW,IAAI,EAAE,MAAM,UAAU;CAEnD;CAEA,OAAO,SAAS,KAAK,MAAM;AAC7B"}
+{"version":3,"file":"pdf-extractor.js","names":[],"sources":["../../src/extractors/pdf-extractor.ts"],"sourcesContent":["import { extractText } from \"unpdf\";\nimport { collectBytes } from \"../collect-bytes.js\";\nimport { DEFAULT_MAX_BYTES, type ExtractorLimits } from \"../limits.js\";\nimport type { ContentExtractor } from \"../types.js\";\n\n/**\n * Extracts PDF content page-by-page, formatting each page as a markdown\n * section with a `## Page N` header. This preserves page boundaries in\n * the indexed content, making search results more precise.\n */\nexport function createPdfExtractor(limits: ExtractorLimits = {}): ContentExtractor {\n  const { maxBytes = DEFAULT_MAX_BYTES } = limits;\n  return async (content) => {\n    const bytes = await collectBytes(content, { maxBytes });\n    // A copy, always a plain Uint8Array: pdfjs rejects a Node Buffer, and it takes over (detaches)\n    // the buffer it is given, which would empty the caller's array.\n    const { text: pages } = await extractText(new Uint8Array(bytes));\n\n    const sections: string[] = [];\n    for (let i = 0; i < pages.length; i++) {\n      const pageText = pages[i]?.trim();\n      if (pageText) {\n        sections.push(`## Page ${i + 1}\\n\\n${pageText}`);\n      }\n    }\n\n    return sections.join(\"\\n\\n\");\n  };\n}\n\n/** PDF extractor with the default limits. */\nexport const pdfExtractor: ContentExtractor = createPdfExtractor();\n"],"mappings":";;;;;;;;;AAUA,SAAgB,mBAAmB,SAA0B,CAAC,GAAqB;CACjF,MAAM,EAAE,WAAW,sBAAsB;CACzC,OAAO,OAAO,YAAY;EACxB,MAAM,QAAQ,MAAM,aAAa,SAAS,EAAE,SAAS,CAAC;EAGtD,MAAM,EAAE,MAAM,UAAU,MAAM,YAAY,IAAI,WAAW,KAAK,CAAC;EAE/D,MAAM,WAAqB,CAAC;EAC5B,KAAK,IAAI,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK;GACrC,MAAM,WAAW,MAAM,EAAE,EAAE,KAAK;GAChC,IAAI,UACF,SAAS,KAAK,WAAW,IAAI,EAAE,MAAM,UAAU;EAEnD;EAEA,OAAO,SAAS,KAAK,MAAM;CAC7B;AACF;;AAGA,MAAa,eAAiC,mBAAmB"}

package/dist/extractors/text-extractor.d.ts

@@ -1,3 +1,4 @@
+import { ExtractorLimits } from "../limits.js";
 import { ContentExtractor } from "../types.js";
 //#region src/extractors/text-extractor.d.ts
 /**
@@ -5,6 +6,8 @@ import { ContentExtractor } from "../types.js";
  * decode bytes to a string. Exists as a distinct entry so `.txt` files
  * route through the same pipeline as every other format.
  */
+export declare function createTextExtractor(limits?: ExtractorLimits): ContentExtractor;
+/** Plain-text extractor with the default limits. */
 export declare const textExtractor: ContentExtractor;
 //#endregion
 //# sourceMappingURL=text-extractor.d.ts.map

package/dist/extractors/text-extractor.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"text-extractor.d.ts","names":[],"sources":["../../src/extractors/text-extractor.ts"],"mappings":";;;;;;;qBAQa,eAAe"}
+{"version":3,"file":"text-extractor.d.ts","names":[],"sources":["../../src/extractors/text-extractor.ts"],"mappings":";;;;;;;;wBASgB,oBAAoB,SAAQ,kBAAuB;;qBAMtD,eAAe"}

package/dist/extractors/text-extractor.js

@@ -1,14 +1,18 @@
 import { collectText } from "../collect-bytes.js";
+import { DEFAULT_MAX_BYTES } from "../limits.js";
 //#region src/extractors/text-extractor.ts
 /**
 * Identity extractor for plain text -- no conversion needed, just
 * decode bytes to a string. Exists as a distinct entry so `.txt` files
 * route through the same pipeline as every other format.
 */
-const textExtractor = async (content) => {
-	return collectText(content);
-};
+function createTextExtractor(limits = {}) {
+	const { maxBytes = DEFAULT_MAX_BYTES } = limits;
+	return (content) => collectText(content, { maxBytes });
+}
+/** Plain-text extractor with the default limits. */
+const textExtractor = createTextExtractor();
 //#endregion
-export { textExtractor };
+export { createTextExtractor, textExtractor };
 
 //# sourceMappingURL=text-extractor.js.map

package/dist/extractors/text-extractor.js.map

@@ -1 +1 @@
-{"version":3,"file":"text-extractor.js","names":[],"sources":["../../src/extractors/text-extractor.ts"],"sourcesContent":["import { collectText } from \"../collect-bytes.js\";\nimport type { ContentExtractor } from \"../types.js\";\n\n/**\n * Identity extractor for plain text -- no conversion needed, just\n * decode bytes to a string. Exists as a distinct entry so `.txt` files\n * route through the same pipeline as every other format.\n */\nexport const textExtractor: ContentExtractor = async (content) => {\n  return collectText(content);\n};\n"],"mappings":";;;;;;;AAQA,MAAa,gBAAkC,OAAO,YAAY;CAChE,OAAO,YAAY,OAAO;AAC5B"}
+{"version":3,"file":"text-extractor.js","names":[],"sources":["../../src/extractors/text-extractor.ts"],"sourcesContent":["import { collectText } from \"../collect-bytes.js\";\nimport { DEFAULT_MAX_BYTES, type ExtractorLimits } from \"../limits.js\";\nimport type { ContentExtractor } from \"../types.js\";\n\n/**\n * Identity extractor for plain text -- no conversion needed, just\n * decode bytes to a string. Exists as a distinct entry so `.txt` files\n * route through the same pipeline as every other format.\n */\nexport function createTextExtractor(limits: ExtractorLimits = {}): ContentExtractor {\n  const { maxBytes = DEFAULT_MAX_BYTES } = limits;\n  return (content) => collectText(content, { maxBytes });\n}\n\n/** Plain-text extractor with the default limits. */\nexport const textExtractor: ContentExtractor = createTextExtractor();\n"],"mappings":";;;;;;;;AASA,SAAgB,oBAAoB,SAA0B,CAAC,GAAqB;CAClF,MAAM,EAAE,WAAW,sBAAsB;CACzC,QAAQ,YAAY,YAAY,SAAS,EAAE,SAAS,CAAC;AACvD;;AAGA,MAAa,gBAAkC,oBAAoB"}

package/dist/extractors/xlsx-extractor.d.ts

@@ -1,11 +1,15 @@
+import { ZipExtractorLimits } from "../limits.js";
 import { ContentExtractor } from "../types.js";
 //#region src/extractors/xlsx-extractor.d.ts
 /**
  * Converts Excel spreadsheets into markdown tables so the content
  * is searchable and consumable by text-based pipelines. Each worksheet
  * becomes a separate section with a level-2 heading. Empty sheets are
- * skipped to avoid noise.
+ * skipped to avoid noise. The input size and the archive's declared
+ * uncompressed size are checked before exceljs inflates it.
  */
+export declare function createXlsxExtractor(limits?: ZipExtractorLimits): ContentExtractor;
+/** XLSX extractor with the default limits. */
 export declare const xlsxExtractor: ContentExtractor;
 //#endregion
 //# sourceMappingURL=xlsx-extractor.d.ts.map

package/dist/extractors/xlsx-extractor.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"xlsx-extractor.d.ts","names":[],"sources":["../../src/extractors/xlsx-extractor.ts"],"mappings":";;;;;;;;qBAUa,eAAe"}
+{"version":3,"file":"xlsx-extractor.d.ts","names":[],"sources":["../../src/extractors/xlsx-extractor.ts"],"mappings":";;;;;;;;;;wBAkCgB,oBAAoB,SAAQ,qBAA0B;;qBAqDzD,eAAe"}

package/dist/extractors/xlsx-extractor.js

@@ -1,49 +1,69 @@
 import { collectBytes } from "../collect-bytes.js";
+import { DEFAULT_MAX_BYTES, DEFAULT_MAX_UNCOMPRESSED_BYTES } from "../limits.js";
+import { assertZipUncompressedSize } from "../zip-size.js";
 import ExcelJS from "exceljs";
 //#region src/extractors/xlsx-extractor.ts
 /**
+* Makes a cell value safe inside a Markdown table row: `|` becomes `\|` so it does not start a
+* new column, and every line break (CR, LF, CRLF) becomes one space so it does not end the row.
+*/
+function escapeCell(value) {
+	return value.replace(/\|/g, "\\|").replace(/\r\n|\r|\n/g, " ");
+}
+/** Width of the widest row. A loop, not `Math.max(...)`, so large sheets do not overflow the stack. */
+function maxColumnCount(rows) {
+	let max = 0;
+	for (const row of rows) if (row.length > max) max = row.length;
+	return max;
+}
+/**
 * Converts Excel spreadsheets into markdown tables so the content
 * is searchable and consumable by text-based pipelines. Each worksheet
 * becomes a separate section with a level-2 heading. Empty sheets are
-* skipped to avoid noise.
+* skipped to avoid noise. The input size and the archive's declared
+* uncompressed size are checked before exceljs inflates it.
 */
-const xlsxExtractor = async (content) => {
-	const bytes = await collectBytes(content);
-	const workbook = new ExcelJS.Workbook();
-	const arrayBuffer = bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength);
-	await workbook.xlsx.load(arrayBuffer);
-	const sections = [];
-	workbook.eachSheet((sheet) => {
-		const rows = [];
-		sheet.eachRow((row) => {
-			const cells = [];
-			row.eachCell({ includeEmpty: true }, (cell) => {
-				try {
-					const v = cell.value;
-					cells.push(v == null ? "" : String(v));
-				} catch {
-					cells.push("");
-				}
+function createXlsxExtractor(limits = {}) {
+	const { maxBytes = DEFAULT_MAX_BYTES, maxUncompressedBytes = DEFAULT_MAX_UNCOMPRESSED_BYTES } = limits;
+	return async (content) => {
+		const bytes = await collectBytes(content, { maxBytes });
+		assertZipUncompressedSize(bytes, maxUncompressedBytes);
+		const workbook = new ExcelJS.Workbook();
+		const arrayBuffer = bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength);
+		await workbook.xlsx.load(arrayBuffer);
+		const sections = [];
+		workbook.eachSheet((sheet) => {
+			const rows = [];
+			sheet.eachRow((row) => {
+				const cells = [];
+				row.eachCell({ includeEmpty: true }, (cell) => {
+					try {
+						const v = cell.value;
+						cells.push(v == null ? "" : escapeCell(String(v)));
+					} catch {
+						cells.push("");
+					}
+				});
+				rows.push(cells);
 			});
-			rows.push(cells);
-		});
-		if (rows.length === 0) return;
-		const colCount = Math.max(...rows.map((r) => r.length));
-		const [headers, ...dataRows] = rows.map((r) => {
-			while (r.length < colCount) r.push("");
-			return r;
+			if (rows.length === 0) return;
+			const colCount = maxColumnCount(rows);
+			for (const r of rows) while (r.length < colCount) r.push("");
+			const [headers, ...dataRows] = rows;
+			if (!headers) return;
+			const table = [
+				`| ${headers.join(" | ")} |`,
+				`| ${headers.map(() => "---").join(" | ")} |`,
+				...dataRows.map((r) => `| ${r.join(" | ")} |`)
+			].join("\n");
+			sections.push(`## ${sheet.name}\n\n${table}`);
 		});
-		if (!headers) return;
-		const table = [
-			`| ${headers.join(" | ")} |`,
-			`| ${headers.map(() => "---").join(" | ")} |`,
-			...dataRows.map((r) => `| ${r.join(" | ")} |`)
-		].join("\n");
-		sections.push(`## ${sheet.name}\n\n${table}`);
-	});
-	return sections.join("\n\n");
-};
+		return sections.join("\n\n");
+	};
+}
+/** XLSX extractor with the default limits. */
+const xlsxExtractor = createXlsxExtractor();
 //#endregion
-export { xlsxExtractor };
+export { createXlsxExtractor, escapeCell, maxColumnCount, xlsxExtractor };
 
 //# sourceMappingURL=xlsx-extractor.js.map

package/dist/extractors/xlsx-extractor.js.map

@@ -1 +1 @@
-{"version":3,"file":"xlsx-extractor.js","names":[],"sources":["../../src/extractors/xlsx-extractor.ts"],"sourcesContent":["import ExcelJS from \"exceljs\";\nimport { collectBytes } from \"../collect-bytes.js\";\nimport type { ContentExtractor } from \"../types.js\";\n\n/**\n * Converts Excel spreadsheets into markdown tables so the content\n * is searchable and consumable by text-based pipelines. Each worksheet\n * becomes a separate section with a level-2 heading. Empty sheets are\n * skipped to avoid noise.\n */\nexport const xlsxExtractor: ContentExtractor = async (content) => {\n  const bytes = await collectBytes(content);\n  const workbook = new ExcelJS.Workbook();\n  const arrayBuffer = bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength);\n  await workbook.xlsx.load(arrayBuffer as ArrayBuffer);\n\n  const sections: string[] = [];\n\n  workbook.eachSheet((sheet) => {\n    const rows: string[][] = [];\n\n    sheet.eachRow((row) => {\n      const cells: string[] = [];\n      row.eachCell({ includeEmpty: true }, (cell) => {\n        try {\n          const v = cell.value;\n          cells.push(v == null ? \"\" : String(v));\n        } catch {\n          cells.push(\"\");\n        }\n      });\n      rows.push(cells);\n    });\n\n    if (rows.length === 0) return;\n\n    // Normalize column count to the widest row\n    const colCount = Math.max(...rows.map((r) => r.length));\n    const padded = rows.map((r) => {\n      while (r.length < colCount) r.push(\"\");\n      return r;\n    });\n\n    // First row as headers, rest as data\n    const [headers, ...dataRows] = padded;\n    if (!headers) return;\n\n    const headerLine = `| ${headers.join(\" | \")} |`;\n    const separator = `| ${headers.map(() => \"---\").join(\" | \")} |`;\n    const bodyLines = dataRows.map((r) => `| ${r.join(\" | \")} |`);\n\n    const table = [headerLine, separator, ...bodyLines].join(\"\\n\");\n    sections.push(`## ${sheet.name}\\n\\n${table}`);\n  });\n\n  return sections.join(\"\\n\\n\");\n};\n"],"mappings":";;;;;;;;;AAUA,MAAa,gBAAkC,OAAO,YAAY;CAChE,MAAM,QAAQ,MAAM,aAAa,OAAO;CACxC,MAAM,WAAW,IAAI,QAAQ,SAAS;CACtC,MAAM,cAAc,MAAM,OAAO,MAAM,MAAM,YAAY,MAAM,aAAa,MAAM,UAAU;CAC5F,MAAM,SAAS,KAAK,KAAK,WAA0B;CAEnD,MAAM,WAAqB,CAAC;CAE5B,SAAS,WAAW,UAAU;EAC5B,MAAM,OAAmB,CAAC;EAE1B,MAAM,SAAS,QAAQ;GACrB,MAAM,QAAkB,CAAC;GACzB,IAAI,SAAS,EAAE,cAAc,KAAK,IAAI,SAAS;IAC7C,IAAI;KACF,MAAM,IAAI,KAAK;KACf,MAAM,KAAK,KAAK,OAAO,KAAK,OAAO,CAAC,CAAC;IACvC,QAAQ;KACN,MAAM,KAAK,EAAE;IACf;GACF,CAAC;GACD,KAAK,KAAK,KAAK;EACjB,CAAC;EAED,IAAI,KAAK,WAAW,GAAG;EAGvB,MAAM,WAAW,KAAK,IAAI,GAAG,KAAK,KAAK,MAAM,EAAE,MAAM,CAAC;EAOtD,MAAM,CAAC,SAAS,GAAG,YANJ,KAAK,KAAK,MAAM;GAC7B,OAAO,EAAE,SAAS,UAAU,EAAE,KAAK,EAAE;GACrC,OAAO;EACT,CAGoC;EACpC,IAAI,CAAC,SAAS;EAMd,MAAM,QAAQ;GAAC,KAJS,QAAQ,KAAK,KAAK,EAAE;GAIjB,KAHJ,QAAQ,UAAU,KAAK,CAAC,CAAC,KAAK,KAAK,EAAE;GAGtB,GAFpB,SAAS,KAAK,MAAM,KAAK,EAAE,KAAK,KAAK,EAAE,GAER;EAAC,CAAC,CAAC,KAAK,IAAI;EAC7D,SAAS,KAAK,MAAM,MAAM,KAAK,MAAM,OAAO;CAC9C,CAAC;CAED,OAAO,SAAS,KAAK,MAAM;AAC7B"}
+{"version":3,"file":"xlsx-extractor.js","names":[],"sources":["../../src/extractors/xlsx-extractor.ts"],"sourcesContent":["import ExcelJS from \"exceljs\";\nimport { collectBytes } from \"../collect-bytes.js\";\nimport {\n  DEFAULT_MAX_BYTES,\n  DEFAULT_MAX_UNCOMPRESSED_BYTES,\n  type ZipExtractorLimits,\n} from \"../limits.js\";\nimport type { ContentExtractor } from \"../types.js\";\nimport { assertZipUncompressedSize } from \"../zip-size.js\";\n\n/**\n * Makes a cell value safe inside a Markdown table row: `|` becomes `\\|` so it does not start a\n * new column, and every line break (CR, LF, CRLF) becomes one space so it does not end the row.\n */\nexport function escapeCell(value: string): string {\n  return value.replace(/\\|/g, \"\\\\|\").replace(/\\r\\n|\\r|\\n/g, \" \");\n}\n\n/** Width of the widest row. A loop, not `Math.max(...)`, so large sheets do not overflow the stack. */\nexport function maxColumnCount(rows: readonly (readonly unknown[])[]): number {\n  let max = 0;\n  for (const row of rows) {\n    if (row.length > max) max = row.length;\n  }\n  return max;\n}\n\n/**\n * Converts Excel spreadsheets into markdown tables so the content\n * is searchable and consumable by text-based pipelines. Each worksheet\n * becomes a separate section with a level-2 heading. Empty sheets are\n * skipped to avoid noise. The input size and the archive's declared\n * uncompressed size are checked before exceljs inflates it.\n */\nexport function createXlsxExtractor(limits: ZipExtractorLimits = {}): ContentExtractor {\n  const { maxBytes = DEFAULT_MAX_BYTES, maxUncompressedBytes = DEFAULT_MAX_UNCOMPRESSED_BYTES } =\n    limits;\n  return async (content) => {\n    const bytes = await collectBytes(content, { maxBytes });\n    assertZipUncompressedSize(bytes, maxUncompressedBytes);\n    const workbook = new ExcelJS.Workbook();\n    const arrayBuffer = bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength);\n    await workbook.xlsx.load(arrayBuffer as ArrayBuffer);\n\n    const sections: string[] = [];\n\n    workbook.eachSheet((sheet) => {\n      const rows: string[][] = [];\n\n      sheet.eachRow((row) => {\n        const cells: string[] = [];\n        row.eachCell({ includeEmpty: true }, (cell) => {\n          try {\n            const v = cell.value;\n            cells.push(v == null ? \"\" : escapeCell(String(v)));\n          } catch {\n            cells.push(\"\");\n          }\n        });\n        rows.push(cells);\n      });\n\n      if (rows.length === 0) return;\n\n      // Normalize column count to the widest row\n      const colCount = maxColumnCount(rows);\n      for (const r of rows) {\n        while (r.length < colCount) r.push(\"\");\n      }\n\n      // First row as headers, rest as data\n      const [headers, ...dataRows] = rows;\n      if (!headers) return;\n\n      const headerLine = `| ${headers.join(\" | \")} |`;\n      const separator = `| ${headers.map(() => \"---\").join(\" | \")} |`;\n      const bodyLines = dataRows.map((r) => `| ${r.join(\" | \")} |`);\n\n      const table = [headerLine, separator, ...bodyLines].join(\"\\n\");\n      sections.push(`## ${sheet.name}\\n\\n${table}`);\n    });\n\n    return sections.join(\"\\n\\n\");\n  };\n}\n\n/** XLSX extractor with the default limits. */\nexport const xlsxExtractor: ContentExtractor = createXlsxExtractor();\n"],"mappings":";;;;;;;;;AAcA,SAAgB,WAAW,OAAuB;CAChD,OAAO,MAAM,QAAQ,OAAO,KAAK,CAAC,CAAC,QAAQ,eAAe,GAAG;AAC/D;;AAGA,SAAgB,eAAe,MAA+C;CAC5E,IAAI,MAAM;CACV,KAAK,MAAM,OAAO,MAChB,IAAI,IAAI,SAAS,KAAK,MAAM,IAAI;CAElC,OAAO;AACT;;;;;;;;AASA,SAAgB,oBAAoB,SAA6B,CAAC,GAAqB;CACrF,MAAM,EAAE,WAAW,mBAAmB,uBAAuB,mCAC3D;CACF,OAAO,OAAO,YAAY;EACxB,MAAM,QAAQ,MAAM,aAAa,SAAS,EAAE,SAAS,CAAC;EACtD,0BAA0B,OAAO,oBAAoB;EACrD,MAAM,WAAW,IAAI,QAAQ,SAAS;EACtC,MAAM,cAAc,MAAM,OAAO,MAAM,MAAM,YAAY,MAAM,aAAa,MAAM,UAAU;EAC5F,MAAM,SAAS,KAAK,KAAK,WAA0B;EAEnD,MAAM,WAAqB,CAAC;EAE5B,SAAS,WAAW,UAAU;GAC5B,MAAM,OAAmB,CAAC;GAE1B,MAAM,SAAS,QAAQ;IACrB,MAAM,QAAkB,CAAC;IACzB,IAAI,SAAS,EAAE,cAAc,KAAK,IAAI,SAAS;KAC7C,IAAI;MACF,MAAM,IAAI,KAAK;MACf,MAAM,KAAK,KAAK,OAAO,KAAK,WAAW,OAAO,CAAC,CAAC,CAAC;KACnD,QAAQ;MACN,MAAM,KAAK,EAAE;KACf;IACF,CAAC;IACD,KAAK,KAAK,KAAK;GACjB,CAAC;GAED,IAAI,KAAK,WAAW,GAAG;GAGvB,MAAM,WAAW,eAAe,IAAI;GACpC,KAAK,MAAM,KAAK,MACd,OAAO,EAAE,SAAS,UAAU,EAAE,KAAK,EAAE;GAIvC,MAAM,CAAC,SAAS,GAAG,YAAY;GAC/B,IAAI,CAAC,SAAS;GAMd,MAAM,QAAQ;IAAC,KAJS,QAAQ,KAAK,KAAK,EAAE;IAIjB,KAHJ,QAAQ,UAAU,KAAK,CAAC,CAAC,KAAK,KAAK,EAAE;IAGtB,GAFpB,SAAS,KAAK,MAAM,KAAK,EAAE,KAAK,KAAK,EAAE,GAER;GAAC,CAAC,CAAC,KAAK,IAAI;GAC7D,SAAS,KAAK,MAAM,MAAM,KAAK,MAAM,OAAO;EAC9C,CAAC;EAED,OAAO,SAAS,KAAK,MAAM;CAC7B;AACF;;AAGA,MAAa,gBAAkC,oBAAoB"}

package/dist/html-to-markdown.d.ts

@@ -3,7 +3,9 @@
  * Normalizes HTML into markdown so downstream consumers get a uniform
  * text format regardless of whether the source was HTML, DOCX (via
  * mammoth's HTML output), or any other HTML-producing extractor.
- * Uses Turndown with GFM tables and strikethrough support.
+ * Uses Turndown with GFM tables and strikethrough support. The content of
+ * `<script>`, `<style>`, `<noscript>` and `<template>` is dropped: it is
+ * code or markup, not document text.
  */
 export declare function htmlToMarkdown(html: string): string;
 //#endregion

package/dist/html-to-markdown.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"html-to-markdown.d.ts","names":[],"sources":["../src/html-to-markdown.ts"],"mappings":";;;;;;;wBAUgB,eAAe"}
+{"version":3,"file":"html-to-markdown.d.ts","names":[],"sources":["../src/html-to-markdown.ts"],"mappings":";;;;;;;;;wBAYgB,eAAe"}

package/dist/html-to-markdown.js

@@ -5,7 +5,9 @@ import TurndownService from "turndown";
 * Normalizes HTML into markdown so downstream consumers get a uniform
 * text format regardless of whether the source was HTML, DOCX (via
 * mammoth's HTML output), or any other HTML-producing extractor.
-* Uses Turndown with GFM tables and strikethrough support.
+* Uses Turndown with GFM tables and strikethrough support. The content of
+* `<script>`, `<style>`, `<noscript>` and `<template>` is dropped: it is
+* code or markup, not document text.
 */
 function htmlToMarkdown(html) {
 	const turndownService = new TurndownService({
@@ -17,6 +19,12 @@ function htmlToMarkdown(html) {
 		strongDelimiter: "**"
 	});
 	turndownService.use(gfm);
+	turndownService.remove([
+		"script",
+		"style",
+		"noscript",
+		"template"
+	]);
 	return turndownService.turndown(html);
 }
 //#endregion

package/dist/html-to-markdown.js.map

@@ -1 +1 @@
-{"version":3,"file":"html-to-markdown.js","names":[],"sources":["../src/html-to-markdown.ts"],"sourcesContent":["/// <reference path=\"./turndown-plugin-gfm.d.ts\" />\nimport { gfm } from \"@joplin/turndown-plugin-gfm\";\nimport TurndownService from \"turndown\";\n\n/**\n * Normalizes HTML into markdown so downstream consumers get a uniform\n * text format regardless of whether the source was HTML, DOCX (via\n * mammoth's HTML output), or any other HTML-producing extractor.\n * Uses Turndown with GFM tables and strikethrough support.\n */\nexport function htmlToMarkdown(html: string): string {\n  const turndownService = new TurndownService({\n    headingStyle: \"atx\",\n    hr: \"---\",\n    bulletListMarker: \"*\",\n    codeBlockStyle: \"fenced\",\n    emDelimiter: \"*\",\n    strongDelimiter: \"**\",\n  });\n  turndownService.use(gfm);\n  return turndownService.turndown(html);\n}\n"],"mappings":";;;;;;;;;AAUA,SAAgB,eAAe,MAAsB;CACnD,MAAM,kBAAkB,IAAI,gBAAgB;EAC1C,cAAc;EACd,IAAI;EACJ,kBAAkB;EAClB,gBAAgB;EAChB,aAAa;EACb,iBAAiB;CACnB,CAAC;CACD,gBAAgB,IAAI,GAAG;CACvB,OAAO,gBAAgB,SAAS,IAAI;AACtC"}
+{"version":3,"file":"html-to-markdown.js","names":[],"sources":["../src/html-to-markdown.ts"],"sourcesContent":["/// <reference path=\"./turndown-plugin-gfm.d.ts\" />\nimport { gfm } from \"@joplin/turndown-plugin-gfm\";\nimport TurndownService from \"turndown\";\n\n/**\n * Normalizes HTML into markdown so downstream consumers get a uniform\n * text format regardless of whether the source was HTML, DOCX (via\n * mammoth's HTML output), or any other HTML-producing extractor.\n * Uses Turndown with GFM tables and strikethrough support. The content of\n * `<script>`, `<style>`, `<noscript>` and `<template>` is dropped: it is\n * code or markup, not document text.\n */\nexport function htmlToMarkdown(html: string): string {\n  const turndownService = new TurndownService({\n    headingStyle: \"atx\",\n    hr: \"---\",\n    bulletListMarker: \"*\",\n    codeBlockStyle: \"fenced\",\n    emDelimiter: \"*\",\n    strongDelimiter: \"**\",\n  });\n  turndownService.use(gfm);\n  turndownService.remove([\"script\", \"style\", \"noscript\", \"template\"]);\n  return turndownService.turndown(html);\n}\n"],"mappings":";;;;;;;;;;;AAYA,SAAgB,eAAe,MAAsB;CACnD,MAAM,kBAAkB,IAAI,gBAAgB;EAC1C,cAAc;EACd,IAAI;EACJ,kBAAkB;EAClB,gBAAgB;EAChB,aAAa;EACb,iBAAiB;CACnB,CAAC;CACD,gBAAgB,IAAI,GAAG;CACvB,gBAAgB,OAAO;EAAC;EAAU;EAAS;EAAY;CAAU,CAAC;CAClE,OAAO,gBAAgB,SAAS,IAAI;AACtC"}

package/dist/index.d.ts

@@ -1,14 +1,16 @@
+import { DEFAULT_MAX_BYTES, DEFAULT_MAX_UNCOMPRESSED_BYTES, ExtractorLimits, ZipExtractorLimits } from "./limits.js";
 import { ContentExtractor, ContentNormalizer } from "./types.js";
-import { docxExtractor } from "./extractors/docx-extractor.js";
-import { htmlExtractor } from "./extractors/html-extractor.js";
-import { markdownExtractor } from "./extractors/markdown-extractor.js";
-import { pdfExtractor } from "./extractors/pdf-extractor.js";
-import { textExtractor } from "./extractors/text-extractor.js";
-import { xlsxExtractor } from "./extractors/xlsx-extractor.js";
+import { createDocxExtractor, docxExtractor } from "./extractors/docx-extractor.js";
+import { createHtmlExtractor, htmlExtractor } from "./extractors/html-extractor.js";
+import { createMarkdownExtractor, markdownExtractor } from "./extractors/markdown-extractor.js";
+import { createPdfExtractor, pdfExtractor } from "./extractors/pdf-extractor.js";
+import { createTextExtractor, textExtractor } from "./extractors/text-extractor.js";
+import { createXlsxExtractor, xlsxExtractor } from "./extractors/xlsx-extractor.js";
 import { ExtractorRegistry } from "./extractor-registry.js";
 import { createDefaultRegistry } from "./extractors/index.js";
-import { collectBytes, collectText } from "./collect-bytes.js";
+import { CollectOptions, collectBytes, collectText } from "./collect-bytes.js";
 import { htmlToMarkdown } from "./html-to-markdown.js";
-import { markdownToHtml } from "./markdown-to-html.js";
+import { MarkdownToHtmlOptions, markdownToHtml } from "./markdown-to-html.js";
 import { getMimeType } from "./mime-utils.js";
-export { ContentExtractor, ContentNormalizer, ExtractorRegistry, collectBytes, collectText, createDefaultRegistry, docxExtractor, getMimeType, htmlExtractor, htmlToMarkdown, markdownExtractor, markdownToHtml, pdfExtractor, textExtractor, xlsxExtractor };
+import { assertZipUncompressedSize, zipUncompressedSize } from "./zip-size.js";
+export { CollectOptions, ContentExtractor, ContentNormalizer, DEFAULT_MAX_BYTES, DEFAULT_MAX_UNCOMPRESSED_BYTES, ExtractorLimits, ExtractorRegistry, MarkdownToHtmlOptions, ZipExtractorLimits, assertZipUncompressedSize, collectBytes, collectText, createDefaultRegistry, createDocxExtractor, createHtmlExtractor, createMarkdownExtractor, createPdfExtractor, createTextExtractor, createXlsxExtractor, docxExtractor, getMimeType, htmlExtractor, htmlToMarkdown, markdownExtractor, markdownToHtml, pdfExtractor, textExtractor, xlsxExtractor, zipUncompressedSize };

package/dist/index.js

@@ -1,13 +1,15 @@
 import { collectBytes, collectText } from "./collect-bytes.js";
 import { htmlToMarkdown } from "./html-to-markdown.js";
-import { docxExtractor } from "./extractors/docx-extractor.js";
-import { htmlExtractor } from "./extractors/html-extractor.js";
-import { markdownExtractor } from "./extractors/markdown-extractor.js";
-import { pdfExtractor } from "./extractors/pdf-extractor.js";
-import { textExtractor } from "./extractors/text-extractor.js";
-import { xlsxExtractor } from "./extractors/xlsx-extractor.js";
+import { DEFAULT_MAX_BYTES, DEFAULT_MAX_UNCOMPRESSED_BYTES } from "./limits.js";
+import { assertZipUncompressedSize, zipUncompressedSize } from "./zip-size.js";
+import { createDocxExtractor, docxExtractor } from "./extractors/docx-extractor.js";
+import { createHtmlExtractor, htmlExtractor } from "./extractors/html-extractor.js";
+import { createMarkdownExtractor, markdownExtractor } from "./extractors/markdown-extractor.js";
+import { createPdfExtractor, pdfExtractor } from "./extractors/pdf-extractor.js";
+import { createTextExtractor, textExtractor } from "./extractors/text-extractor.js";
+import { createXlsxExtractor, xlsxExtractor } from "./extractors/xlsx-extractor.js";
 import { ExtractorRegistry } from "./extractor-registry.js";
 import { createDefaultRegistry } from "./extractors/index.js";
 import { markdownToHtml } from "./markdown-to-html.js";
 import { getMimeType } from "./mime-utils.js";
-export { ExtractorRegistry, collectBytes, collectText, createDefaultRegistry, docxExtractor, getMimeType, htmlExtractor, htmlToMarkdown, markdownExtractor, markdownToHtml, pdfExtractor, textExtractor, xlsxExtractor };
+export { DEFAULT_MAX_BYTES, DEFAULT_MAX_UNCOMPRESSED_BYTES, ExtractorRegistry, assertZipUncompressedSize, collectBytes, collectText, createDefaultRegistry, createDocxExtractor, createHtmlExtractor, createMarkdownExtractor, createPdfExtractor, createTextExtractor, createXlsxExtractor, docxExtractor, getMimeType, htmlExtractor, htmlToMarkdown, markdownExtractor, markdownToHtml, pdfExtractor, textExtractor, xlsxExtractor, zipUncompressedSize };

package/dist/limits.d.ts

@@ -0,0 +1,20 @@
+//#region src/limits.d.ts
+/** Default largest input an extractor reads: 100 MiB. */
+export declare const DEFAULT_MAX_BYTES: number;
+/** Default largest total uncompressed size of a DOCX/XLSX archive: 500 MiB. */
+export declare const DEFAULT_MAX_UNCOMPRESSED_BYTES: number;
+/** Limits every built-in extractor accepts. Pass `Infinity` to turn a limit off. */
+export interface ExtractorLimits {
+  /** Largest input read, in bytes. Default {@link DEFAULT_MAX_BYTES}. */
+  maxBytes?: number;
+}
+/** Limits of the ZIP-based extractors (DOCX, DOCM, XLSX). */
+export interface ZipExtractorLimits extends ExtractorLimits {
+  /**
+   * Largest sum of the uncompressed entry sizes declared in the archive's central directory,
+   * checked before the archive is inflated. Default {@link DEFAULT_MAX_UNCOMPRESSED_BYTES}.
+   */
+  maxUncompressedBytes?: number;
+}
+//#endregion
+//# sourceMappingURL=limits.d.ts.map

package/dist/limits.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"limits.d.ts","names":[],"sources":["../src/limits.ts"],"mappings":";;qBACa;;qBAGA;;iBAGI;;EAEf;;;iBAIe,2BAA2B;;;;;EAK1C"}

package/dist/limits.js

@@ -0,0 +1,9 @@
+//#region src/limits.ts
+/** Default largest input an extractor reads: 100 MiB. */
+const DEFAULT_MAX_BYTES = 104857600;
+/** Default largest total uncompressed size of a DOCX/XLSX archive: 500 MiB. */
+const DEFAULT_MAX_UNCOMPRESSED_BYTES = 524288e3;
+//#endregion
+export { DEFAULT_MAX_BYTES, DEFAULT_MAX_UNCOMPRESSED_BYTES };
+
+//# sourceMappingURL=limits.js.map

package/dist/limits.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"limits.js","names":[],"sources":["../src/limits.ts"],"sourcesContent":["/** Default largest input an extractor reads: 100 MiB. */\nexport const DEFAULT_MAX_BYTES = 100 * 1024 * 1024;\n\n/** Default largest total uncompressed size of a DOCX/XLSX archive: 500 MiB. */\nexport const DEFAULT_MAX_UNCOMPRESSED_BYTES = 500 * 1024 * 1024;\n\n/** Limits every built-in extractor accepts. Pass `Infinity` to turn a limit off. */\nexport interface ExtractorLimits {\n  /** Largest input read, in bytes. Default {@link DEFAULT_MAX_BYTES}. */\n  maxBytes?: number;\n}\n\n/** Limits of the ZIP-based extractors (DOCX, DOCM, XLSX). */\nexport interface ZipExtractorLimits extends ExtractorLimits {\n  /**\n   * Largest sum of the uncompressed entry sizes declared in the archive's central directory,\n   * checked before the archive is inflated. Default {@link DEFAULT_MAX_UNCOMPRESSED_BYTES}.\n   */\n  maxUncompressedBytes?: number;\n}\n"],"mappings":";;AACA,MAAa,oBAAoB;;AAGjC,MAAa,iCAAiC"}

package/dist/markdown-to-html.d.ts

@@ -1,10 +1,18 @@
 //#region src/markdown-to-html.d.ts
+export interface MarkdownToHtmlOptions {
+  /**
+   * Pass raw HTML in the Markdown through to the output. Off by default: raw HTML is escaped,
+   * so `<script>` or `<img onerror>` in an uploaded file cannot run when the output is shown.
+   * Turn it on only for trusted input, or sanitise the output before rendering it.
+   */
+  html?: boolean;
+}
 /**
  * Provides the reverse of `htmlToMarkdown` for cases where extracted
  * markdown needs to be rendered back to HTML (e.g. previews, email
- * output). Configured with pass-through raw HTML and auto-linking
- * so embedded markup and bare URLs survive the round-trip.
+ * output). Bare URLs are auto-linked. Raw HTML is escaped unless
+ * `options.html` is set.
  */
-export declare function markdownToHtml(markdown: string): string;
+export declare function markdownToHtml(markdown: string, options?: MarkdownToHtmlOptions): string;
 //#endregion
 //# sourceMappingURL=markdown-to-html.d.ts.map

package/dist/markdown-to-html.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"markdown-to-html.d.ts","names":[],"sources":["../src/markdown-to-html.ts"],"mappings":";;;;;;;wBAegB,eAAe"}
+{"version":3,"file":"markdown-to-html.d.ts","names":[],"sources":["../src/markdown-to-html.ts"],"mappings":";iBAMiB;;;;;;EAMf;;;;;;;;wBASc,eAAe,kBAAkB,UAAS"}

package/dist/markdown-to-html.js

@@ -1,20 +1,27 @@
 import MarkdownIt from "markdown-it";
 //#region src/markdown-to-html.ts
-const md = new MarkdownIt({
-	html: true,
+const options = {
 	linkify: true,
 	typographer: true,
 	breaks: false
+};
+const escaping = new MarkdownIt({
+	...options,
+	html: false
+});
+const passThrough = new MarkdownIt({
+	...options,
+	html: true
 });
 /**
 * Provides the reverse of `htmlToMarkdown` for cases where extracted
 * markdown needs to be rendered back to HTML (e.g. previews, email
-* output). Configured with pass-through raw HTML and auto-linking
-* so embedded markup and bare URLs survive the round-trip.
+* output). Bare URLs are auto-linked. Raw HTML is escaped unless
+* `options.html` is set.
 */
-function markdownToHtml(markdown) {
+function markdownToHtml(markdown, options = {}) {
 	if (!markdown || markdown.trim() === "") return "";
-	return md.render(markdown);
+	return (options.html ? passThrough : escaping).render(markdown);
 }
 //#endregion
 export { markdownToHtml };

package/dist/markdown-to-html.js.map

@@ -1 +1 @@
-{"version":3,"file":"markdown-to-html.js","names":[],"sources":["../src/markdown-to-html.ts"],"sourcesContent":["import MarkdownIt from \"markdown-it\";\n\nconst md = new MarkdownIt({\n  html: true,\n  linkify: true,\n  typographer: true,\n  breaks: false,\n});\n\n/**\n * Provides the reverse of `htmlToMarkdown` for cases where extracted\n * markdown needs to be rendered back to HTML (e.g. previews, email\n * output). Configured with pass-through raw HTML and auto-linking\n * so embedded markup and bare URLs survive the round-trip.\n */\nexport function markdownToHtml(markdown: string): string {\n  if (!markdown || markdown.trim() === \"\") {\n    return \"\";\n  }\n  return md.render(markdown);\n}\n"],"mappings":";;AAEA,MAAM,KAAK,IAAI,WAAW;CACxB,MAAM;CACN,SAAS;CACT,aAAa;CACb,QAAQ;AACV,CAAC;;;;;;;AAQD,SAAgB,eAAe,UAA0B;CACvD,IAAI,CAAC,YAAY,SAAS,KAAK,MAAM,IACnC,OAAO;CAET,OAAO,GAAG,OAAO,QAAQ;AAC3B"}
+{"version":3,"file":"markdown-to-html.js","names":[],"sources":["../src/markdown-to-html.ts"],"sourcesContent":["import MarkdownIt from \"markdown-it\";\n\nconst options = { linkify: true, typographer: true, breaks: false } as const;\nconst escaping = new MarkdownIt({ ...options, html: false });\nconst passThrough = new MarkdownIt({ ...options, html: true });\n\nexport interface MarkdownToHtmlOptions {\n  /**\n   * Pass raw HTML in the Markdown through to the output. Off by default: raw HTML is escaped,\n   * so `<script>` or `<img onerror>` in an uploaded file cannot run when the output is shown.\n   * Turn it on only for trusted input, or sanitise the output before rendering it.\n   */\n  html?: boolean;\n}\n\n/**\n * Provides the reverse of `htmlToMarkdown` for cases where extracted\n * markdown needs to be rendered back to HTML (e.g. previews, email\n * output). Bare URLs are auto-linked. Raw HTML is escaped unless\n * `options.html` is set.\n */\nexport function markdownToHtml(markdown: string, options: MarkdownToHtmlOptions = {}): string {\n  if (!markdown || markdown.trim() === \"\") {\n    return \"\";\n  }\n  return (options.html ? passThrough : escaping).render(markdown);\n}\n"],"mappings":";;AAEA,MAAM,UAAU;CAAE,SAAS;CAAM,aAAa;CAAM,QAAQ;AAAM;AAClE,MAAM,WAAW,IAAI,WAAW;CAAE,GAAG;CAAS,MAAM;AAAM,CAAC;AAC3D,MAAM,cAAc,IAAI,WAAW;CAAE,GAAG;CAAS,MAAM;AAAK,CAAC;;;;;;;AAiB7D,SAAgB,eAAe,UAAkB,UAAiC,CAAC,GAAW;CAC5F,IAAI,CAAC,YAAY,SAAS,KAAK,MAAM,IACnC,OAAO;CAET,QAAQ,QAAQ,OAAO,cAAc,SAAA,CAAU,OAAO,QAAQ;AAChE"}

package/dist/mime-utils.js

@@ -18,9 +18,10 @@ const MIME_MAP = {
 * Content-Type header and need a MIME type for the registry fallback path.
 */
 function getMimeType(path) {
-	const dotIndex = path.lastIndexOf(".");
+	const name = path.slice(path.lastIndexOf("/") + 1);
+	const dotIndex = name.lastIndexOf(".");
 	if (dotIndex === -1) return void 0;
-	const ext = path.slice(dotIndex).toLowerCase();
+	const ext = name.slice(dotIndex).toLowerCase();
 	return MIME_MAP[ext];
 }
 //#endregion

package/dist/mime-utils.js.map

@@ -1 +1 @@
-{"version":3,"file":"mime-utils.js","names":[],"sources":["../src/mime-utils.ts"],"sourcesContent":["const MIME_MAP: Record<string, string> = {\n  \".md\": \"text/markdown\",\n  \".markdown\": \"text/markdown\",\n  \".txt\": \"text/plain\",\n  \".html\": \"text/html\",\n  \".htm\": \"text/html\",\n  \".json\": \"application/json\",\n  \".pdf\": \"application/pdf\",\n  \".docx\": \"application/vnd.openxmlformats-officedocument.wordprocessingml.document\",\n  \".docm\": \"application/vnd.ms-word.document.macroEnabled.12\",\n  \".xlsx\": \"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet\",\n};\n\n/**\n * Lightweight alternative to a full MIME-sniffing library -- covers only\n * the extensions this package can actually extract, keeping the dependency\n * footprint small. Used by callers who have a filename but no\n * Content-Type header and need a MIME type for the registry fallback path.\n */\nexport function getMimeType(path: string): string | undefined {\n  const dotIndex = path.lastIndexOf(\".\");\n  if (dotIndex === -1) return undefined;\n  const ext = path.slice(dotIndex).toLowerCase();\n  return MIME_MAP[ext];\n}\n"],"mappings":";AAAA,MAAM,WAAmC;CACvC,OAAO;CACP,aAAa;CACb,QAAQ;CACR,SAAS;CACT,QAAQ;CACR,SAAS;CACT,QAAQ;CACR,SAAS;CACT,SAAS;CACT,SAAS;AACX;;;;;;;AAQA,SAAgB,YAAY,MAAkC;CAC5D,MAAM,WAAW,KAAK,YAAY,GAAG;CACrC,IAAI,aAAa,IAAI,OAAO,KAAA;CAC5B,MAAM,MAAM,KAAK,MAAM,QAAQ,CAAC,CAAC,YAAY;CAC7C,OAAO,SAAS;AAClB"}
+{"version":3,"file":"mime-utils.js","names":[],"sources":["../src/mime-utils.ts"],"sourcesContent":["const MIME_MAP: Record<string, string> = {\n  \".md\": \"text/markdown\",\n  \".markdown\": \"text/markdown\",\n  \".txt\": \"text/plain\",\n  \".html\": \"text/html\",\n  \".htm\": \"text/html\",\n  \".json\": \"application/json\",\n  \".pdf\": \"application/pdf\",\n  \".docx\": \"application/vnd.openxmlformats-officedocument.wordprocessingml.document\",\n  \".docm\": \"application/vnd.ms-word.document.macroEnabled.12\",\n  \".xlsx\": \"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet\",\n};\n\n/**\n * Lightweight alternative to a full MIME-sniffing library -- covers only\n * the extensions this package can actually extract, keeping the dependency\n * footprint small. Used by callers who have a filename but no\n * Content-Type header and need a MIME type for the registry fallback path.\n */\nexport function getMimeType(path: string): string | undefined {\n  const name = path.slice(path.lastIndexOf(\"/\") + 1);\n  const dotIndex = name.lastIndexOf(\".\");\n  if (dotIndex === -1) return undefined;\n  const ext = name.slice(dotIndex).toLowerCase();\n  return MIME_MAP[ext];\n}\n"],"mappings":";AAAA,MAAM,WAAmC;CACvC,OAAO;CACP,aAAa;CACb,QAAQ;CACR,SAAS;CACT,QAAQ;CACR,SAAS;CACT,QAAQ;CACR,SAAS;CACT,SAAS;CACT,SAAS;AACX;;;;;;;AAQA,SAAgB,YAAY,MAAkC;CAC5D,MAAM,OAAO,KAAK,MAAM,KAAK,YAAY,GAAG,IAAI,CAAC;CACjD,MAAM,WAAW,KAAK,YAAY,GAAG;CACrC,IAAI,aAAa,IAAI,OAAO,KAAA;CAC5B,MAAM,MAAM,KAAK,MAAM,QAAQ,CAAC,CAAC,YAAY;CAC7C,OAAO,SAAS;AAClB"}

package/dist/zip-size.d.ts

@@ -0,0 +1,16 @@
+//#region src/zip-size.d.ts
+/**
+ * Sums the uncompressed sizes of all entries declared in a ZIP archive's central directory,
+ * without inflating anything. Supports ZIP64. Throws when the bytes are not a readable ZIP.
+ *
+ * The sizes are what the archive claims; a parser that inflates the data still has to check
+ * them, which JSZip (used by mammoth and exceljs) does.
+ */
+export declare function zipUncompressedSize(bytes: Uint8Array): number;
+/**
+ * Throws when the uncompressed size declared by a ZIP archive exceeds `maxBytes`, so a small
+ * "zip bomb" is refused before a parser inflates it.
+ */
+export declare function assertZipUncompressedSize(bytes: Uint8Array, maxBytes: number): void;
+//#endregion
+//# sourceMappingURL=zip-size.d.ts.map

package/dist/zip-size.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"zip-size.d.ts","names":[],"sources":["../src/zip-size.ts"],"mappings":";;;;;;;;wBAkBgB,oBAAoB,OAAO;;;;;wBAuF3B,0BAA0B,OAAO,YAAY"}

package/dist/zip-size.js

@@ -0,0 +1,90 @@
+//#region src/zip-size.ts
+const EOCD_SIG = 101010256;
+const ZIP64_LOCATOR_SIG = 117853008;
+const ZIP64_EOCD_SIG = 101075792;
+const CD_ENTRY_SIG = 33639248;
+const EOCD_SIZE = 22;
+const MAX_COMMENT = 65535;
+function fail(message) {
+	throw new Error(`Invalid ZIP archive: ${message}`);
+}
+/**
+* Sums the uncompressed sizes of all entries declared in a ZIP archive's central directory,
+* without inflating anything. Supports ZIP64. Throws when the bytes are not a readable ZIP.
+*
+* The sizes are what the archive claims; a parser that inflates the data still has to check
+* them, which JSZip (used by mammoth and exceljs) does.
+*/
+function zipUncompressedSize(bytes) {
+	const view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength);
+	const len = bytes.byteLength;
+	let eocd = -1;
+	const stop = Math.max(0, len - EOCD_SIZE - MAX_COMMENT);
+	for (let i = len - EOCD_SIZE; i >= stop; i--) if (view.getUint32(i, true) === EOCD_SIG) {
+		eocd = i;
+		break;
+	}
+	if (eocd < 0) fail("not a ZIP file (no end of central directory record)");
+	let cdEnd = eocd;
+	let entries = view.getUint16(eocd + 10, true);
+	let cdSize = view.getUint32(eocd + 12, true);
+	let cdOffset = view.getUint32(eocd + 16, true);
+	if (entries === 65535 || cdSize === 4294967295 || cdOffset === 4294967295) {
+		const locator = eocd - 20;
+		if (locator >= 0 && view.getUint32(locator, true) === ZIP64_LOCATOR_SIG) {
+			const z64 = Number(view.getBigUint64(locator + 8, true));
+			if (z64 + 56 > len || view.getUint32(z64, true) !== ZIP64_EOCD_SIG) fail("bad ZIP64 end of central directory record");
+			cdEnd = z64;
+			entries = Number(view.getBigUint64(z64 + 32, true));
+			cdSize = Number(view.getBigUint64(z64 + 40, true));
+			cdOffset = Number(view.getBigUint64(z64 + 48, true));
+		}
+	}
+	if (cdOffset + 4 > len || entries > 0 && view.getUint32(cdOffset, true) !== CD_ENTRY_SIG) {
+		const shifted = cdEnd - cdSize;
+		if (shifted >= 0 && shifted + 4 <= len && view.getUint32(shifted, true) === CD_ENTRY_SIG) cdOffset = shifted;
+	}
+	if (cdOffset + cdSize > len) fail("central directory lies outside the file");
+	let total = 0;
+	let pos = cdOffset;
+	for (let n = 0; n < entries; n++) {
+		if (pos + 46 > len || view.getUint32(pos, true) !== CD_ENTRY_SIG) fail("truncated or corrupt central directory");
+		const size32 = view.getUint32(pos + 24, true);
+		const nameLen = view.getUint16(pos + 28, true);
+		const extraLen = view.getUint16(pos + 30, true);
+		const commentLen = view.getUint16(pos + 32, true);
+		let size = size32;
+		if (size32 === 4294967295) size = readZip64Size(view, pos + 46 + nameLen, extraLen);
+		total += size;
+		pos += 46 + nameLen + extraLen + commentLen;
+	}
+	return total;
+}
+/** Reads the 64-bit uncompressed size from an entry's ZIP64 extended information extra field. */
+function readZip64Size(view, start, length) {
+	const end = start + length;
+	if (end > view.byteLength) fail("truncated or corrupt central directory");
+	let p = start;
+	while (p + 4 <= end) {
+		const id = view.getUint16(p, true);
+		const size = view.getUint16(p + 2, true);
+		if (id === 1) {
+			if (size < 8 || p + 12 > end) fail("bad ZIP64 extra field");
+			return Number(view.getBigUint64(p + 4, true));
+		}
+		p += 4 + size;
+	}
+	fail("missing ZIP64 extra field");
+}
+/**
+* Throws when the uncompressed size declared by a ZIP archive exceeds `maxBytes`, so a small
+* "zip bomb" is refused before a parser inflates it.
+*/
+function assertZipUncompressedSize(bytes, maxBytes) {
+	const size = zipUncompressedSize(bytes);
+	if (size > maxBytes) throw new Error(`Archive uncompressed size ${size} bytes exceeds the limit of ${maxBytes} bytes`);
+}
+//#endregion
+export { assertZipUncompressedSize, zipUncompressedSize };
+
+//# sourceMappingURL=zip-size.js.map