feat(docx-to-vfile): make more isomorphic

libs/reoff/docx-to-vfile/package.json

@@ -28,6 +28,11 @@
     }
   },
   "main": "./index.js",
+  "browser": {
+    "fs": false,
+    "stream": false,
+    "buffer": false
+  },
   "types": "./index.d.ts",
   "files": [
     "index.d.ts",

libs/reoff/docx-to-vfile/src/index.ts

@@ -1 +1,42 @@
+// import type { ReadStream } from 'fs'
+// import { VFile } from 'vfile'
+// import { docxToVFile as docxToVFileBrowser, Options } from './lib/docx-to-vfile-unzipit.js'
+
+// /**
+//  * Takes a docx file as a Blob, Buffer, ArrayBuffer, ReadStream or string
+//  *  and returns a VFile with the contents of the document.xml file as the root, and the contents of the other xml files as data.
+//  *
+//  * Should work in both Node and the browser, but to be sure it works in the browser, import it like
+//  * ```ts
+//  * import { docxToVFile } from 'docx-to-vfile/browser'
+//  * ```
+//  *
+//  * @param file The docx file as a File, Blob, Buffer, ArrayBuffer, ReadStream or string
+//  * @param options Options
+//  * @returns A VFile with the contents of the document.xml file as the root, and the contents of the other xml files as data.
+//  */
+// export async function docxToVFile(
+//   file: ArrayBuffer | File | Blob | Buffer | ReadStream | string,
+//   options?: Options,
+// ): Promise<VFile> {
+//   if (typeof window !== 'undefined') {
+//     return await docxToVFileBrowser(file as Blob, options)
+//   }
+
+//   const { createReadStream, ReadStream } = await import('fs')
+//   const { blob } = await import('stream/consumers')
+//   const { Blob: NodeBlob } = await import('buffer')
+
+//   const inp = typeof file === 'string' ? createReadStream(file) : file
+
+//   const input =
+//     inp instanceof ReadStream
+//       ? await blob(inp)
+//       : inp instanceof ArrayBuffer
+//       ? new NodeBlob([Buffer.from(inp)])
+//       : inp
+
+//   return await docxToVFileBrowser(input as Blob | File, options)
+// }
+
 export * from './lib/docx-to-vfile-unzipit.js'

libs/reoff/docx-to-vfile/src/lib/docx-to-vfile-unzipit.spec.ts

@@ -1,6 +1,5 @@
 import { docxToVFile } from './docx-to-vfile-unzipit.js'
 import fs from 'fs'
-import path from 'path'
 
 describe('reoffDocxToVfile', () => {
   const doc = fs.readFileSync(

libs/reoff/docx-to-vfile/src/lib/docx-to-vfile-unzipit.ts

@@ -1,6 +1,6 @@
 import { Data, VFile } from 'vfile'
 import { unzip } from 'unzipit'
-import { ReadStream } from 'fs'
+import type { ReadStream } from 'fs'
 
 const removeHeader = (text: string | undefined) => (text ? text.replace(/<\?xml.*?\?>/, '') : '')
 
@@ -10,7 +10,7 @@ export interface Options {
   /**
    * Whether or not to include media in the VFile.
    *
-   * By default, images are included on the `data.media` attribute of the VFile as an object of ArrayBuffers, which are accessible both client and serverside.
+   * By default, images are included on the `data.media` attribute of the VFile as an object of Blobs, which are accessible both client and serverside.
    *
    * @default false
    */
@@ -31,6 +31,13 @@ export interface Options {
   include?: string[] | RegExp[] | ((key: string) => boolean) | 'all' | 'allWithDocumentXML'
 }
 
+export interface OptionsWithFetchConfig extends Options {
+  /**
+   * The config to pass to fetch, for e.g. authorization headers.
+   */
+  fetchConfig?: RequestInit
+}
+
 /**
  * The data attribute of a VFile
  * Is set to the DataMap interface in the vfile module
@@ -43,7 +50,7 @@ export interface DocxData extends Data {
   /**
    * The media files in the .docx file
    */
-  media: { [key: string]: ArrayBuffer }
+  media: { [key: string]: Blob }
   /**
    * The relations between the .xml files in the .docx file
    */
@@ -60,7 +67,7 @@ declare module 'vfile' {
      * The media files in the .docx file
      * Possibly undefined only to be compatible with the VFile interface
      */
-    media: { [key: string]: ArrayBuffer }
+    media: { [key: string]: Blob }
     /**
      * The relations between the .xml files in the .docx file
      * Possibly undefined only to be compatible with the VFile interface
@@ -82,25 +89,42 @@ export interface DocxVFile extends VFile {
 }
 
 /**
- * Takes a docx file as an ArrayBuffer and returns a VFile with the contents of the document.xml file as the root, and the contents of the other xml files as data.
+ * Takes a docx file as a Blob or File and returns a VFile with the contents of the document.xml file as the root, and the contents of the other xml files as data.
  *
- * @param file The docx file as an ArrayBuffer or Blob
+ * @param file The docx file as a Blob or File
  * @param options Options
  * @returns A VFile with the contents of the document.xml file as the root, and the contents of the other xml files as data.
  */
 export async function docxToVFile(
-  file: ArrayBuffer | Blob | Buffer,
-  userOptions: Options = {},
+  file: ArrayBuffer | File | Blob | Buffer | ReadStream | string,
+  userOptions?: Options,
 ): Promise<VFile> {
+  let input = file
+
+  // node code
+  if (typeof window === 'undefined') {
+    const { createReadStream, ReadStream } = await import('fs')
+    const { blob } = await import('stream/consumers')
+    const { Blob: NodeBlob } = await import('buffer')
+
+    const inp = typeof file === 'string' ? createReadStream(file) : file
+
+    input = (
+      inp instanceof ReadStream
+        ? await blob(inp)
+        : inp instanceof ArrayBuffer
+        ? new NodeBlob([Buffer.from(inp)])
+        : inp
+    ) as Blob
+  }
+
   const options: Options = {
     withoutMedia: false,
     include: 'all',
     ...userOptions,
   }
 
-  const blb = file instanceof Blob ? file : new Blob([file])
-
-  const { entries } = await unzip(blb)
+  const { entries } = await unzip(input as Blob)
   const rels = await entries['word/_rels/document.xml.rels'].text()
   const relations = Object.fromEntries(
     // eslint-disable-next-line regexp/no-super-linear-backtracking
@@ -144,7 +168,7 @@ export async function docxToVFile(
 
   const vfileData: DocxData = textEntriesObject
   vfileData.relations = relations
-  vfileData.media = {} as { [key: string]: ArrayBuffer }
+  vfileData.media = {} as { [key: string]: Blob }
 
   // vfile.data = vfileData
 
@@ -153,9 +177,9 @@ export async function docxToVFile(
   }
 
   const mediaUrls = Object.values(relations).filter((rel: string) => rel.includes('media/'))
-  const media = {} as { [key: string]: ArrayBuffer }
+  const media = {} as { [key: string]: Blob }
   for (const url of mediaUrls) {
-    media[url] = await entries[`word/${url}`].arrayBuffer()
+    media[url] = await entries[`word/${url}`].blob()
   }
   vfileData.media = media
   return new VFile({ value: removeCarriage(doc), data: vfileData })

nx.json

@@ -58,6 +58,9 @@
     },
     "e2e": {
       "inputs": ["default", "^production"]
+    },
+    "test": {
+      "inputs": ["default", "^production"]
     }
   },
   "tasksRunnerOptions": {

package.json

@@ -178,6 +178,7 @@
     "@nrwl/nest": "15.8.5",
     "@nrwl/node": "15.8.5",
     "@nrwl/react": "15.8.5",
+    "@nrwl/vite": "^15.8.8",
     "@nrwl/web": "15.8.5",
     "@nrwl/workspace": "15.8.5",
     "@nxext/sveltekit": "14.0.0-next.2",
@@ -213,6 +214,8 @@
     "@typescript/analyze-trace": "^0.9.0",
     "@unified-latex/unified-latex-types": "^1.3.0",
     "@vercel/remote-nx": "^1.0.1",
+    "@vitest/coverage-c8": "~0.25.8",
+    "@vitest/ui": "^0.25.8",
     "autoprefixer": "10.4.13",
     "babel-jest": "29.4.3",
     "commander": "^9.0.0",
@@ -238,6 +241,7 @@
     "jest-esm-transformer": "^1.0.0",
     "jest-transform-nearley": "^2.0.0",
     "jsdoc": "^4.0.2",
+    "jsdom": "~20.0.3",
     "jsonc-eslint-parser": "^2.1.0",
     "nx": "15.8.7",
     "postcss": "8.4.19",
@@ -262,7 +266,10 @@
     "typescript": "4.9.5",
     "unist-util-remove-position": "4.0.2",
     "unist-util-visit-parents": "5.1.3",
-    "vercel": "^27.0.1"
+    "vercel": "^27.0.1",
+    "vite-plugin-eslint": "^1.8.1",
+    "vite-tsconfig-paths": "^4.0.2",
+    "vitest": "^0.25.8"
   },
   "packageManager": "yarn@1.22.1",
   "repository": "https://github.com/TrialAndErrorOrg/parsers",
@@ -274,5 +281,8 @@
     "syntax",
     "tree",
     "ast"
-  ]
+  ],
+  "wallaby": {
+    "trace": true
+  }
 }