office-text-extractor - npm Package Compare versions

Comparing version 3.0.3 to 3.4.0-beta.0

build/lib.d.ts

@@ -1,3 +0,2 @@
-/// <reference types="node" resolution-mode="require"/>
-import { Buffer } from 'node:buffer';
+import { Buffer } from 'buffer/index.js';
 /**
@@ -4,0 +3,0 @@ * A method of text extraction.

build/lib.js

 // source/lib.ts
 // The source code for the library.
-import { Buffer } from 'node:buffer';
+import { Buffer } from 'buffer/index.js';
 import { fileTypeFromBuffer as getFileType } from 'file-type';
@@ -5,0 +5,0 @@ import { readFile, fetchUrl } from './util.js';

build/parsers/doc.d.ts

@@ -1,2 +0,2 @@
-import { type Buffer } from 'node:buffer';
+import { type Buffer } from 'buffer/';
 import type { TextExtractionMethod } from '../lib.js';
@@ -3,0 +3,0 @@ export declare class DocExtractor implements TextExtractionMethod {

build/parsers/doc.js

@@ -20,2 +20,3 @@ // source/parsers/docx.ts
             // Convert the DOCX to text and return the text.
+            // @ts-expect-error: see feross/buffer#353, the types are incomplete.
             const parsedDocx = await parseWordFile({ buffer: input });
@@ -22,0 +23,0 @@ return parsedDocx.value;

build/parsers/excel.d.ts

@@ -1,2 +0,2 @@
-import { type Buffer } from 'node:buffer';
+import { type Buffer } from 'buffer/';
 import type { TextExtractionMethod } from '../lib.js';
@@ -3,0 +3,0 @@ export declare class ExcelExtractor implements TextExtractionMethod {

build/parsers/pdf.d.ts

@@ -1,2 +0,2 @@
-import { type Buffer } from 'node:buffer';
+import { type Buffer } from 'buffer/';
 import type { TextExtractionMethod } from '../lib.js';
@@ -3,0 +3,0 @@ export declare class PdfExtractor implements TextExtractionMethod {

build/parsers/ppt.d.ts

@@ -1,2 +0,2 @@
-import { type Buffer } from 'node:buffer';
+import { type Buffer } from 'buffer/';
 import type { TextExtractionMethod } from '../lib.js';
@@ -3,0 +3,0 @@ export declare class PptExtractor implements TextExtractionMethod {

build/util.d.ts

@@ -1,3 +0,3 @@
-import { type Buffer } from 'node:buffer';
+import { type Buffer } from 'buffer/';
 export declare const readFile: (filePath: string) => Promise<Buffer>;
 export declare const fetchUrl: (url: string) => Promise<Buffer>;

build/util.js

@@ -5,3 +5,3 @@ // source/util.ts
 import { got as fetch } from 'got';
-export const readFile = async (filePath) => read(filePath);
-export const fetchUrl = async (url) => fetch(url).buffer();
+export const readFile = async (filePath) => (await read(filePath));
+export const fetchUrl = async (url) => (await fetch(url).buffer());

package.json

 {
   "name": "office-text-extractor",
-  "version": "3.0.3",
+  "version": "3.4.0-beta.0",
   "description": "Yet another library to extract text from MS Office and PDF files",
@@ -43,24 +43,25 @@ "keywords": [
   "dependencies": {
-    "fflate": "0.8.1",
-    "file-type": "18.5.0",
-    "got": "13.0.0",
+    "buffer": "6.0.3",
+    "fflate": "0.8.2",
+    "file-type": "19.3.0",
+    "got": "14.4.1",
     "js-yaml": "4.1.0",
-    "mammoth": "1.6.0",
+    "mammoth": "1.8.0",
     "pdf-parse": "1.1.1",
     "text-encoding": "0.7.0",
-    "xlsx": "https://cdn.sheetjs.com/xlsx-0.20.2/xlsx-0.20.2.tgz",
+    "xlsx": "https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz",
     "xml2js": "0.6.2"
   },
   "devDependencies": {
-    "@types/js-yaml": "4.0.6",
-    "@types/node": "20.8.3",
-    "@types/text-encoding": "0.0.37",
-    "@types/xml2js": "0.4.12",
-    "ava": "5.3.1",
-    "np": "8.0.4",
+    "@types/js-yaml": "4.0.9",
+    "@types/node": "20.14.11",
+    "@types/text-encoding": "0.0.39",
+    "@types/xml2js": "0.4.14",
+    "ava": "6.1.3",
+    "np": "10.0.7",
     "npm-run-all": "4.1.5",
-    "prettier": "3.0.3",
-    "tsx": "3.13.0",
-    "typescript": "5.2.2",
-    "xo": "0.56.0"
+    "prettier": "3.3.3",
+    "tsimp": "2.0.11",
+    "typescript": "5.5.4",
+    "xo": "0.59.0"
   },
@@ -90,3 +91,3 @@ "prettier": {
     "nodeArguments": [
-      "--no-warnings"
+      "--import=tsimp"
     ]
@@ -99,4 +100,4 @@ },
     "test:quality": "xo source/ test/",
-    "test:integration": "NODE_OPTIONS='--loader=tsx' ava"
+    "test:integration": "TSIMP_DIAG=ignore ava"
   }
 }

source/lib.ts

 // source/lib.ts
 // The source code for the library.
 
-import { Buffer } from 'node:buffer'
+import { Buffer } from 'buffer/index.js'
 import { fileTypeFromBuffer as getFileType } from 'file-type'
@@ -6,0 +6,0 @@ import { readFile, fetchUrl } from './util.js'

source/parsers/doc.ts

 // source/parsers/docx.ts
 // The text extracter for DOCX files.
 
-import { type Buffer } from 'node:buffer'
+import { type Buffer } from 'buffer/'
 import { extractRawText as parseWordFile } from 'mammoth'
@@ -25,2 +25,3 @@
 		// Convert the DOCX to text and return the text.
+		// @ts-expect-error: see feross/buffer#353, the types are incomplete.
 		const parsedDocx = await parseWordFile({ buffer: input })
@@ -27,0 +28,0 @@ return parsedDocx.value

source/parsers/excel.ts

 // source/parsers/excel.ts
 // The text extracter for Excel files.
 
-import { type Buffer } from 'node:buffer'
+import { type Buffer } from 'buffer/'
 import Xlsx, { utils as sheetUtils } from 'xlsx'
@@ -6,0 +6,0 @@ import { dump as convertToYaml } from 'js-yaml'

source/parsers/pdf.ts

 // source/parsers/pdf.ts
 // The text extracter for PDF files.
 
-import { type Buffer } from 'node:buffer'
+import { type Buffer } from 'buffer/'
 // @ts-expect-error There are no types for this package.
@@ -6,0 +6,0 @@ import parsePdf from 'pdf-parse/lib/pdf-parse.js'

source/parsers/ppt.ts

@@ -6,3 +6,3 @@ // source/parsers/ppt.ts
 
-import { type Buffer } from 'node:buffer'
+import { type Buffer } from 'buffer/'
 import { unzip } from 'fflate'
@@ -9,0 +9,0 @@ import { parseStringPromise as xmlToJson } from 'xml2js'

source/util.ts

 // source/util.ts
 // Utility functions to help with the handling of input.
 
-import { type Buffer } from 'node:buffer'
 import { readFile as read } from 'node:fs/promises'
 import { got as fetch } from 'got'
+import { type Buffer } from 'buffer/'
 
 export const readFile = async (filePath: string): Promise<Buffer> =>
-	read(filePath)
+	(await read(filePath)) as unknown as Buffer
 export const fetchUrl = async (url: string): Promise<Buffer> =>
-	fetch(url).buffer()
+	(await fetch(url).buffer()) as unknown as Buffer

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

HTTP dependency

Supply chain risk

Contains a dependency which resolves to a remote HTTP URL which could be used to inject untrusted code and reduce overall package reliability.

Found 1 instance in 1 package

No v1

Quality

Package is not semver >=1. This means it is not stable and does not support ^ ranges.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

HTTP dependency

Supply chain risk

Contains a dependency which resolves to a remote HTTP URL which could be used to inject untrusted code and reduce overall package reliability.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

35001

increased by2.43%

Number of package files

26

increased by4%

Lines of code

781

increased by0.13%

Worsened metrics

Dependency count

10

increased by11.11%

Number of low quality alerts

1

increased byInfinity%

Dependency changes

• + Addedbuffer@6.0.3

• + Added@sec-ant/readable-stream@0.4.1(transitive)

• + Added@sindresorhus/is@6.3.1(transitive)

• + Addedbuffer@6.0.3(transitive)

• + Addedcacheable-request@12.0.1(transitive)

• + Addedfflate@0.8.2(transitive)

• + Addedfile-type@19.3.0(transitive)

• + Addedform-data-encoder@4.0.2(transitive)

• + Addedget-stream@8.0.19.0.1(transitive)

• + Addedgot@14.4.1(transitive)

• + Addedis-stream@4.0.1(transitive)

• + Addedmammoth@1.8.0(transitive)

• + Addedp-cancelable@4.0.1(transitive)

• + Addedstrtok3@8.1.0(transitive)

• + Addedtoken-types@6.0.0(transitive)

• + Addedtype-fest@4.26.1(transitive)

• + Addeduint8array-extras@1.4.0(transitive)

• - Removed@sindresorhus/is@5.6.0(transitive)

• - Removedcacheable-request@10.2.14(transitive)

• - Removedfflate@0.8.1(transitive)

• - Removedfile-type@18.5.0(transitive)

• - Removedform-data-encoder@2.1.4(transitive)

• - Removedget-stream@6.0.1(transitive)

• - Removedgot@13.0.0(transitive)

• - Removedmammoth@1.6.0(transitive)

• - Removedp-cancelable@3.0.0(transitive)

• - Removedreadable-stream@3.6.2(transitive)

• - Removedreadable-web-to-node-stream@3.0.2(transitive)

• - Removedstrtok3@7.1.1(transitive)

• - Removedtoken-types@5.0.1(transitive)

• Updatedfflate@0.8.2

• Updatedfile-type@19.3.0

• Updatedgot@14.4.1

• Updatedmammoth@1.8.0

• Updatedxlsx@https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz