Create separate xlsx runtime package and update imports

Copilot · pelikhan · Copilot · commit 9598f0ed3a1a · 2025-09-18T01:41:00.000Z
Co-authored-by: pelikhan &lt;4175913+pelikhan@users.noreply.github.com&gt;
diff --git a/packages/cli/package.json b/packages/cli/package.json
@@ -100,7 +100,7 @@
     "turndown": "^7.2.0",
     "turndown-plugin-gfm": "^1.0.2",
     "typescript": "5.8.3",
-    "xlsx": "https://cdn.sheetjs.com/xlsx-0.20.2/xlsx-0.20.2.tgz",
+    "genaiscript-xlsx": "workspace:*",
     "zod": "^3.25.30",
     "zod-to-json-schema": "^3.24.5"
   },
diff --git a/packages/core/package.json b/packages/core/package.json
@@ -99,7 +99,7 @@
     "ts-dedent": "^2.2.0",
     "tsx": "^4.19.4",
     "typescript": "5.8.3",
-    "xlsx": "https://cdn.sheetjs.com/xlsx-0.20.2/xlsx-0.20.2.tgz",
+    "genaiscript-xlsx": "workspace:*",
     "yaml": "^2.8.0"
   },
   "scripts": {
diff --git a/packages/core/src/data.ts b/packages/core/src/data.ts
@@ -15,7 +15,7 @@ import { splitMarkdown } from "./frontmatter"
 import { INITryParse } from "./ini"
 import { JSON5TryParse } from "./json5"
 import { TOMLTryParse } from "./toml"
-import { XLSXParse } from "./xlsx"
+import { XLSXParse } from "genaiscript-xlsx"
 import { XMLTryParse } from "./xml"
 import { YAMLTryParse } from "./yaml"
 import { resolveFileContent } from "./file"
diff --git a/packages/core/src/file.ts b/packages/core/src/file.ts
@@ -13,7 +13,7 @@ import { fromBase64, toBase64 } from "./base64"
 import { host } from "./host"
 import { TraceOptions } from "./trace"
 import { parsePdf } from "./pdf"
-import { XLSXParse } from "./xlsx"
+import { XLSXParse, WorkbookSheet } from "genaiscript-xlsx"
 import { dataToMarkdownTable, CSVTryParse } from "./csv"
 import {
     CSV_REGEX,
diff --git a/packages/core/src/parsers.test.ts b/packages/core/src/parsers.test.ts
@@ -2,7 +2,7 @@ import { describe, beforeEach, test } from "node:test"
 import assert from "node:assert/strict"
 import { createParsers } from "./parsers"
 import { MarkdownTrace } from "./trace"
-import { XLSXParse } from "./xlsx"
+import { XLSXParse } from "genaiscript-xlsx"
 import { readFile } from "fs/promises"
 import { resolve } from "path"
 import { TestHost } from "./testhost"
diff --git a/packages/core/src/parsers.ts b/packages/core/src/parsers.ts
@@ -21,7 +21,7 @@ import { parsePdf } from "./pdf"
 import { HTMLToMarkdown, HTMLToText } from "./html"
 import { MathTryEvaluate } from "./math"
 import { tryValidateJSONWithSchema, validateJSONWithSchema } from "./schema"
-import { XLSXTryParse } from "./xlsx"
+import { XLSXTryParse } from "genaiscript-xlsx"
 import { host } from "./host"
 import { unzip } from "./zip"
 import { JSONLTryParse } from "./jsonl"
diff --git a/packages/core/src/types/prompt_template.d.ts b/packages/core/src/types/prompt_template.d.ts
@@ -2179,6 +2179,8 @@ interface RunPromptResult {
     usage?: RunPromptUsage
 }
 
+import type { ParseXLSXOptions, WorkbookSheet } from "genaiscript-xlsx"
+
 /**
  * Path manipulation functions.
  */
@@ -2322,17 +2324,6 @@ interface HTMLToTextOptions {
     wordwrap?: number | false | null | undefined
 }
 
-interface ParseXLSXOptions {
-    // specific worksheet name
-    sheet?: string
-    // Use specified range (A1-style bounded range string)
-    range?: string
-}
-
-interface WorkbookSheet {
-    name: string
-    rows: object[]
-}
 
 interface ParseZipOptions {
     glob?: string
diff --git a/packages/core/src/xlsx.ts.bak b/packages/core/src/xlsx.ts.bak
diff --git a/packages/xlsx/package.json b/packages/xlsx/package.json
@@ -0,0 +1,25 @@
+{
+  "name": "genaiscript-xlsx",
+  "version": "1.140.0",
+  "main": "src/index.ts",
+  "license": "MIT",
+  "private": true,
+  "npm": {
+    "publish": false
+  },
+  "repository": {
+    "directory": "packages/xlsx",
+    "url": "https://github.com/microsoft/genaiscript"
+  },
+  "dependencies": {
+    "xlsx": "^0.18.5"
+  },
+  "devDependencies": {
+    "@types/node": "^22.15.2",
+    "typescript": "5.8.3"
+  },
+  "scripts": {
+    "typecheck": "tsc -p src",
+    "test": "node --import tsx --test src/**.test.*ts"
+  }
+}
diff --git a/packages/xlsx/src/index.test.ts b/packages/xlsx/src/index.test.ts
@@ -0,0 +1,16 @@
+import { describe, test } from "node:test"
+import assert from "node:assert/strict"
+import { XLSXParse } from "./index"
+import { readFile } from "fs/promises"
+import { resolve } from "path"
+
+describe("xlsx package", async () => {
+    test("XLSX parsing", async () => {
+        const result = await XLSXParse(
+            await readFile(resolve("../../core/src/parsers.test.xlsx"))
+        )
+        assert.deepStrictEqual(result, [
+            { name: "Sheet1", rows: [{ key: 1, value: 2 }] },
+        ])
+    })
+})
diff --git a/packages/xlsx/src/index.ts b/packages/xlsx/src/index.ts
@@ -0,0 +1,74 @@
+/**
+ * XLSX parsing functionality for GenAIScript
+ * Separated from core to isolate Excel processing dependencies
+ */
+
+/**
+ * Options for parsing XLSX files
+ */
+export interface ParseXLSXOptions {
+    // specific worksheet name
+    sheet?: string
+    // Use specified range (A1-style bounded range string)
+    range?: string
+}
+
+/**
+ * Represents a worksheet in an Excel workbook
+ */
+export interface WorkbookSheet {
+    name: string
+    rows: object[]
+}
+
+/**
+ * Parses XLSX data into an array of workbook sheets.
+ *
+ * @param data - The XLSX data to parse.
+ * @param options - Parsing options, including an optional sheet name and other utilities for conversion.
+ * @returns A promise resolving to an array of WorkbookSheet objects, each containing sheet name and data rows.
+ */
+export async function XLSXParse(
+    data: Uint8Array,
+    options?: ParseXLSXOptions
+): Promise<WorkbookSheet[]> {
+    // Destructure options to separate sheet-specific option
+    const { sheet, ...rest } = options || {}
+    // Dynamically import 'xlsx' library's read and utils modules
+    const { read, utils } = await import("xlsx")
+    // Read the workbook from the data with 'array' type
+    const workbook = read(data, { type: "array" })
+    // Filter and map the sheet names to WorkbookSheet objects
+    return workbook.SheetNames.filter((n) => !sheet || n === sheet).map(
+        (name) => {
+            // Convert the worksheet to JSON and cast to object array
+            const worksheet = workbook.Sheets[name]
+            const rows = utils.sheet_to_json(worksheet, rest) as object[]
+            // Return a WorkbookSheet object with sheet name and rows
+            return { name, rows } as WorkbookSheet
+        }
+    )
+}
+
+/**
+ * Attempts to parse XLSX data, returning an empty array on failure.
+ *
+ * @param data - The XLSX data as a Uint8Array.
+ * @param options - Optional parsing options including a specific sheet name.
+ * @returns A promise that resolves to an array of WorkbookSheet objects or an empty array if parsing fails.
+ */
+export async function XLSXTryParse(
+    data: Uint8Array,
+    options?: ParseXLSXOptions
+): Promise<WorkbookSheet[]> {
+    try {
+        if (!data) return []
+        // Attempt to parse the XLSX data
+        return await XLSXParse(data, options)
+    } catch (e) {
+        // Log any errors encountered during parsing
+        console.info(e)
+        // Return an empty array if parsing fails
+        return []
+    }
+}
diff --git a/packages/xlsx/tsconfig.json b/packages/xlsx/tsconfig.json
@@ -0,0 +1,9 @@
+{
+  "extends": "../tsconfig-base.json",
+  "compilerOptions": {
+    "rootDir": ".",
+    "outDir": "./built"
+  },
+  "include": ["src/**/*"],
+  "exclude": ["node_modules", "built"]
+}
