afpp 1.1.0 → 1.2.0

package/README.md

@@ -1,12 +1,7 @@
 # afpp
 
-Another f*cking pdf parser. (alpha)
+Another f\*cking pdf parser. (alpha)
 
-## todo
-
-- [x] pdf2string
-- [ ] pdf2img
-  
 ## Why?
 
 If you are parsing pdf files in nodejs and you are satisfied with your actual solution, good for you, you don't need this.
@@ -35,10 +30,10 @@ then you might find this package useful.
 **commonjs**:
 
 ```js
-const { pdf2string } = require("afpp");
-const path = require("node:path");
+const { pdf2string } = require('afpp');
+const path = require('node:path');
 
-const pathToFile = path.join("example.pdf");
+const pathToFile = path.join('example.pdf');
 
 (async function start() {
   const pdfString = await pdf2string(pathToFile);
@@ -49,10 +44,10 @@ const pathToFile = path.join("example.pdf");
 **esm**:
 
 ```js
-import { pdf2string } from "afpp";
-import path from "node:path";
+import { pdf2string } from 'afpp';
+import path from 'node:path';
 
-const pathToFile = path.join("example.pdf");
+const pathToFile = path.join('example.pdf');
 
 (async function start() {
   const pdfString = await pdf2string(pathToFile);

package/dist/index.js

@@ -2,52 +2,67 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.pdf2string = void 0;
 const promises_1 = require("node:fs/promises");
-const parsePdfFileBuffer = async (data, options) => {
-    return import("pdfjs-dist/legacy/build/pdf.mjs").then(async (pdfjsLib) => {
-        const loadingTask = pdfjsLib.getDocument({
-            data,
-            password: options?.password,
-        });
-        const pdfDocument = await loadingTask.promise;
-        const { numPages } = pdfDocument;
-        const pageContents = new Array(numPages);
-        const pagePromises = [];
-        for (let pageNum = 1; pageNum <= numPages; pageNum += 1) {
-            pagePromises.push(pdfDocument.getPage(pageNum).then(async (page) => {
-                const textContent = await page.getTextContent({
-                    includeMarkedContent: false,
-                });
-                // ? Type assertion of items to TextItem[] should be safe because {includeMarkedContent: false}
-                const items = textContent.items;
-                if (items.length === 0) {
-                    console.log("empty page");
-                    pageContents[pageNum - 1] = "";
-                }
-                else {
-                    const pageText = items.map((item) => item.str || "").join(" ");
-                    pageContents[pageNum - 1] = pageText;
-                }
-            }));
-        }
-        await Promise.all(pagePromises);
-        return pageContents;
+const parsePdfFileBuffer = async (options) => import('pdfjs-dist/legacy/build/pdf.mjs').then(async (pdfjsLib) => {
+    const loadingTask = pdfjsLib.getDocument({
+        ...options,
+        verbosity: 0, // TODO enable for debug
     });
-};
-const pdf2string = async (source, options) => {
-    if (typeof source === "string") {
-        const fileBase64 = await (0, promises_1.readFile)(source, {});
-        const data = new Uint8Array(fileBase64);
-        return parsePdfFileBuffer(data, options);
+    const pdfDocument = await loadingTask.promise;
+    const { numPages } = pdfDocument;
+    const pageContents = new Array(numPages).fill('');
+    const pagePromises = [];
+    for (let pageNum = 1; pageNum <= numPages; pageNum += 1) {
+        pagePromises.push(pdfDocument.getPage(pageNum).then(async (page) => {
+            const textContent = await page.getTextContent({
+                includeMarkedContent: false,
+            });
+            // ? Type assertion of items to TextItem[] should be safe because {includeMarkedContent: false}
+            const items = textContent.items;
+            if (items.length === 0) {
+                pageContents[pageNum - 1] = '';
+            }
+            else {
+                const pageText = items.map((item) => item.str || '').join(' ');
+                pageContents[pageNum - 1] = pageText;
+            }
+        }));
+    }
+    await Promise.all(pagePromises);
+    return pageContents;
+});
+/**
+ * Converts a PDF file from various input formats (Buffer, Uint8Array, string path, or URL) to a string.
+ *
+ * @async
+ * @function pdf2string
+ *
+ * @param {Buffer|Uint8Array|string|URL} input - The PDF source, which can be a file path, URL, Buffer, or Uint8Array.
+ * @param {Object} [options] - Optional parsing options for customizing the PDF parsing process.
+ * @param {string} [options.password] - The password for encrypted PDF files, if required.
+ *
+ * @since — v1.0.0
+ *
+ * @returns {Promise<string>} - A promise that resolves to the string representation of the PDF content.
+ *
+ * @throws {Error} Throws an error if the input type is invalid.
+ */
+const pdf2string = async (input, options) => {
+    if (typeof input === 'string') {
+        const fileBuffer = await (0, promises_1.readFile)(input, {});
+        const data = new Uint8Array(fileBuffer);
+        return parsePdfFileBuffer({ data, ...options });
+    }
+    if (Buffer.isBuffer(input)) {
+        const data = new Uint8Array(input);
+        return parsePdfFileBuffer({ data, ...options });
     }
-    if (Buffer.isBuffer(source)) {
-        const fileBase64 = await (0, promises_1.readFile)(source, {});
-        const data = new Uint8Array(fileBase64);
-        return parsePdfFileBuffer(data, options);
+    if (input instanceof Uint8Array) {
+        return parsePdfFileBuffer({ data: input, ...options });
     }
-    if (source instanceof Uint8Array) {
-        return parsePdfFileBuffer(source, options);
+    if (input instanceof URL) {
+        return parsePdfFileBuffer({ url: input, ...options });
     }
-    throw new Error(`Invalid source type: ${typeof source}`);
+    throw new Error(`Invalid source type: ${typeof input}`);
 };
 exports.pdf2string = pdf2string;
 //# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../index.ts"],"names":[],"mappings":";;;AAAA,+CAA4C;AAY5C,MAAM,kBAAkB,GAAG,KAAK,EAAE,IAAgB,EAAE,OAAsB,EAAE,EAAE;IAC5E,OAAO,MAAM,CAAC,iCAAiC,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,EAAE;QACvE,MAAM,WAAW,GAAG,QAAQ,CAAC,WAAW,CAAC;YACvC,IAAI;YACJ,QAAQ,EAAE,OAAO,EAAE,QAAQ;SAC5B,CAAC,CAAC;QAEH,MAAM,WAAW,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC;QAE9C,MAAM,EAAE,QAAQ,EAAE,GAAG,WAAW,CAAC;QACjC,MAAM,YAAY,GAAa,IAAI,KAAK,CAAC,QAAQ,CAAC,CAAC;QACnD,MAAM,YAAY,GAAmC,EAAE,CAAC;QAExD,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,QAAQ,EAAE,OAAO,IAAI,CAAC,EAAE,CAAC;YACxD,YAAY,CAAC,IAAI,CACf,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;gBAC/C,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,CAAC;oBAC5C,oBAAoB,EAAE,KAAK;iBAC5B,CAAC,CAAC;gBACH,+FAA+F;gBAC/F,MAAM,KAAK,GAAG,WAAW,CAAC,KAAmB,CAAC;gBAC9C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;oBACvB,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;oBAC1B,YAAY,CAAC,OAAO,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC;gBACjC,CAAC;qBAAM,CAAC;oBACN,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;oBAC/D,YAAY,CAAC,OAAO,GAAG,CAAC,CAAC,GAAG,QAAQ,CAAC;gBACvC,CAAC;YACH,CAAC,CAAC,CACH,CAAC;QACJ,CAAC;QACD,MAAM,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;QAChC,OAAO,YAAY,CAAC;IACtB,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAMF,MAAM,UAAU,GAAG,KAAK,EACtB,MAAoC,EACpC,OAAsB,EACtB,EAAE;IACF,IAAI,OAAO,MAAM,KAAK,QAAQ,EAAE,CAAC;QAC/B,MAAM,UAAU,GAAG,MAAM,IAAA,mBAAQ,EAAC,MAAM,EAAE,EAAE,CAAC,CAAC;QAC9C,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,UAAU,CAAC,CAAC;QACxC,OAAO,kBAAkB,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;IAC3C,CAAC;IACD,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;QAC5B,MAAM,UAAU,GAAG,MAAM,IAAA,mBAAQ,EAAC,MAAM,EAAE,EAAE,CAAC,CAAC;QAC9C,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,UAAU,CAAC,CAAC;QACxC,OAAO,kBAAkB,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;IAC3C,CAAC;IACD,IAAI,MAAM,YAAY,UAAU,EAAE,CAAC;QACjC,OAAO,kBAAkB,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC7C,CAAC;IACD,MAAM,IAAI,KAAK,CAAC,wBAAwB,OAAO,MAAM,EAAE,CAAC,CAAC;AAC3D,CAAC,CAAC;AAEO,gCAAU"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";;;AAAA,+CAA4C;AAQ5C,MAAM,kBAAkB,GAAG,KAAK,EAAE,OAA+B,EAAE,EAAE,CACnE,MAAM,CAAC,iCAAiC,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,EAAE;IAChE,MAAM,WAAW,GAAG,QAAQ,CAAC,WAAW,CAAC;QACvC,GAAG,OAAO;QACV,SAAS,EAAE,CAAC,EAAE,wBAAwB;KACvC,CAAC,CAAC;IAEH,MAAM,WAAW,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC;IAE9C,MAAM,EAAE,QAAQ,EAAE,GAAG,WAAW,CAAC;IACjC,MAAM,YAAY,GAAa,IAAI,KAAK,CAAS,QAAQ,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACpE,MAAM,YAAY,GAAmC,EAAE,CAAC;IAExD,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,QAAQ,EAAE,OAAO,IAAI,CAAC,EAAE,CAAC;QACxD,YAAY,CAAC,IAAI,CACf,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;YAC/C,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,CAAC;gBAC5C,oBAAoB,EAAE,KAAK;aAC5B,CAAC,CAAC;YACH,+FAA+F;YAC/F,MAAM,KAAK,GAAG,WAAW,CAAC,KAAmB,CAAC;YAC9C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACvB,YAAY,CAAC,OAAO,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC;YACjC,CAAC;iBAAM,CAAC;gBACN,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBAC/D,YAAY,CAAC,OAAO,GAAG,CAAC,CAAC,GAAG,QAAQ,CAAC;YACvC,CAAC;QACH,CAAC,CAAC,CACH,CAAC;IACJ,CAAC;IACD,MAAM,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;IAChC,OAAO,YAAY,CAAC;AACtB,CAAC,CAAC,CAAC;AAML;;;;;;;;;;;;;;;GAeG;AACH,MAAM,UAAU,GAAG,KAAK,EACtB,KAAyC,EACzC,OAAsB,EACtB,EAAE;IACF,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC9B,MAAM,UAAU,GAAG,MAAM,IAAA,mBAAQ,EAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7C,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,UAAU,CAAC,CAAC;QACxC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IAClD,CAAC;IACD,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,KAAK,CAAC,CAAC;QACnC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IAClD,CAAC;IACD,IAAI,KAAK,YAAY,UAAU,EAAE,CAAC;QAChC,OAAO,kBAAkB,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IACzD,CAAC;IACD,IAAI,KAAK,YAAY,GAAG,EAAE,CAAC;QACzB,OAAO,kBAAkB,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;IACxD,CAAC;IACD,MAAM,IAAI,KAAK,CAAC,wBAAwB,OAAO,KAAK,EAAE,CAAC,CAAC;AAC1D,CAAC,CAAC;AAEO,gCAAU"}

package/dist/types/index.d.ts

@@ -1,5 +1,21 @@
 type ParseOptions = {
     password?: string;
 };
-declare const pdf2string: (source: string | Buffer | Uint8Array, options?: ParseOptions) => Promise<string[]>;
+/**
+ * Converts a PDF file from various input formats (Buffer, Uint8Array, string path, or URL) to a string.
+ *
+ * @async
+ * @function pdf2string
+ *
+ * @param {Buffer|Uint8Array|string|URL} input - The PDF source, which can be a file path, URL, Buffer, or Uint8Array.
+ * @param {Object} [options] - Optional parsing options for customizing the PDF parsing process.
+ * @param {string} [options.password] - The password for encrypted PDF files, if required.
+ *
+ * @since — v1.0.0
+ *
+ * @returns {Promise<string>} - A promise that resolves to the string representation of the PDF content.
+ *
+ * @throws {Error} Throws an error if the input type is invalid.
+ */
+declare const pdf2string: (input: Buffer | URL | Uint8Array | string, options?: ParseOptions) => Promise<string[]>;
 export { pdf2string };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "afpp",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "description": "another f*cking pdf parser",
   "types": "./dist/types/index.d.ts",
   "main": "./dist/index.js",
@@ -8,9 +8,14 @@
     "dist/**"
   ],
   "scripts": {
-    "test": "test",
+    "build": "tsc -p tsconfig.build.json",
+    "commit": "git-cz",
+    "lint": "eslint .",
     "prebuild": "rm -rf dist",
-    "build": "tsc -p tsconfig.json"
+    "prepare": "husky",
+    "test": "NODE_ENV=test npx tsx --test --test-reporter=@voxpelli/node-test-pretty-reporter test/*.test.ts",
+    "test:coverage": "NODE_ENV=test npx tsx --test --experimental-test-coverage test/*.test.ts",
+    "typecheck": "tsc -p tsconfig.json --noEmit"
   },
   "repository": {
     "type": "git",
@@ -34,12 +39,31 @@
   },
   "homepage": "https://github.com/l2ysho/afpp#readme",
   "dependencies": {
-    "pdfjs-dist": "4.6.82"
+    "pdfjs-dist": "4.6.82",
+    "typescript": "5.6.2"
   },
   "devDependencies": {
+    "@commitlint/cli": "19.5.0",
+    "@commitlint/config-conventional": "19.5.0",
     "@types/node": "22.5.5",
+    "@typescript-eslint/eslint-plugin": "7.18.0",
+    "@typescript-eslint/parser": "7.18.0",
+    "@voxpelli/node-test-pretty-reporter": "1.1.2",
+    "commitizen": "4.3.0",
+    "cz-conventional-changelog": "3.3.0",
+    "eslint": "8.56.0",
+    "eslint-config-airbnb": "19.0.4",
+    "eslint-config-airbnb-base": "15.0.0",
+    "eslint-config-airbnb-typescript": "18.0.0",
+    "eslint-config-prettier": "9.1.0",
+    "eslint-import-resolver-typescript": "3.6.1",
+    "eslint-plugin-import": "2.29.1",
+    "eslint-plugin-no-relative-import-paths": "1.5.5",
+    "eslint-plugin-perfectionist": "2.11.0",
+    "eslint-plugin-prettier": "5.2.1",
+    "husky": "9.1.6",
+    "lint-staged": "15.2.10",
     "semantic-release": "24.1.1",
-    "tsx": "4.19.1",
-    "typescript": "5.6.2"
+    "tsx": "4.19.1"
   }
 }