feat: add support of eslint and typescript, plus more tests and excep…

…tion handlings (#346) * feat: add eslint, fix issue #321, add test PDF for #321:image exception will only be warned in log, won't stop the text and form parsing * fix: issue #318, fixed together with #321 * fix: issue #343, resolve pkInfo at build time * fix: issue #255 and #277, keep parsing if img data is not resolved * maint: add ts dependency and config, remove root pdf2json.js, use ./bin/pdf2json.js instead * maint: replace lib/p2jcom*.js with src/cli/p2jcli*.ts, compied cli is moved to /bin/cli * maint: add compiled bundle to bin/cli and dist * maint: remove unnecessary build artifacts * build: add prepare scripts * test: add more page content verification in test * test: add more cli tests after jest * fix: auto create output dir in CLI to make more tests run without setting up directory structure * fix: #262: correct the condition check, also add i262_4hXzVq.pdf to test/misc
modesty · May 24, 2024 · 3e60cde · 3e60cde
1 parent 4c85d0b
commit 3e60cde
Show file tree

Hide file tree

Showing 38 changed files with 2,243 additions and 698 deletions.
diff --git a/.eslintrc.json b/.eslintrc.json
@@ -0,0 +1,134 @@
+{
+	"root": true,
+	"extends": [
+		"eslint:recommended",
+		"plugin:@typescript-eslint/eslint-recommended",
+		"plugin:@typescript-eslint/recommended"],
+	"parser": "@typescript-eslint/parser",
+	"parserOptions": {
+		"ecmaVersion": 2022,
+		"sourceType": "module"
+	},
+  	"ignorePatterns": ["/dist/", "/bin/", "/test/", "/node_modules/", "/base/", "/lib/pdfjs-code.js"],
+	"overrides": [{
+      "files": ["*.js", "*.ts"],
+      "rules": {
+		"no-mixed-spaces-and-tabs": ["error", "smart-tabs"],
+		"@typescript-eslint/naming-convention": [
+			"error",
+			{
+				"selector": "variable",
+				"format": ["camelCase", "PascalCase", "UPPER_CASE"],
+				"filter": {
+					"regex": "^_",
+					"match": false
+				}
+			},
+			{
+				"selector": "function",
+				"format": ["camelCase", "PascalCase"]
+			},
+			{
+				"selector": "parameter",
+				"format": ["camelCase", "PascalCase"],
+				"filter": {
+					"regex": "^_",
+					"match": false
+				}
+			},
+			{
+				"selector": "property",
+				"format": ["camelCase", "PascalCase", "UPPER_CASE"],
+				"filter": {
+					"regex": "(x-ide-git-auth|^_|^sys_*|app_id)",
+					"match": false
+				}
+			},
+			{
+				"selector": "method",
+				"format": ["camelCase", "PascalCase"],
+				"filter": {
+					"regex": "^_",
+					"match": false
+				}
+			},
+			{
+				"selector": "accessor",
+				"format": ["camelCase", "PascalCase"]
+			},
+			{
+				"selector": "enumMember",
+				"format": ["camelCase", "PascalCase", "UPPER_CASE"]
+			},
+			{
+				"selector": "typeLike",
+				"format": ["PascalCase"]
+			}
+		],
+		"@typescript-eslint/no-unused-vars": [
+			"error",
+			{
+				"vars": "all",
+				"args": "after-used",
+				"argsIgnorePattern": "(^_?|fs|uri|options|opts|source|signal|destination|Uri$|args)",
+				"ignoreRestSiblings": true,
+				"destructuredArrayIgnorePattern": "^_?"
+			}
+		],
+		"@typescript-eslint/no-extra-semi": "off",
+		"@typescript-eslint/no-var-requires": "off",
+		"arrow-body-style": ["error", "as-needed"],
+		"dot-notation": ["error"],
+		"eqeqeq": ["error", "always"],
+		"no-case-declarations": "error",
+		"no-duplicate-imports": ["error"],
+		"no-else-return": [
+			"error",
+			{
+				"allowElseIf": true
+			}
+		],
+		"no-eval": [
+			"error",
+			{
+				"allowIndirect": false
+			}
+		],
+		"no-iterator": ["error"],
+		"no-multi-assign": ["error"],
+		"no-new-func": ["error"],
+		"no-new-wrappers": ["error"],
+		"no-object-constructor": ["error"],
+		"no-param-reassign": "off",
+		"no-restricted-imports": ["error", "lodash", "moment"],
+		"no-throw-literal": "warn",
+		"no-useless-call": ["error"],
+		"object-curly-spacing": ["error", "always"],
+		"object-shorthand": [
+			"error",
+			"always",
+			{
+				"avoidExplicitReturnArrows": true
+			}
+		],
+		"prefer-arrow-callback": [
+			"error",
+			{
+				"allowNamedFunctions": true
+			}
+		],
+		"prefer-const": ["error"],
+		"prefer-destructuring": [
+			"error",
+			{
+				"object": true,
+				"array": false
+			}
+		],
+		"prefer-rest-params": ["error"],
+		"prefer-spread": ["error"],
+		"prefer-template": ["error"]
+      }
+    }
+  ]
+}
diff --git a/.gitignore b/.gitignore
@@ -8,6 +8,8 @@ node_modules/
 target/
 .idea
 .npmrc
-vscode/
-dist/
+.vscode/
 lib/pdfjs-code.js
+src/**/*.js
+/dist/
+/bin/cli/
diff --git a/.prettierrc.cjs b/.prettierrc.cjs
@@ -0,0 +1,4 @@
+module.exports = {
+  ...require('prettier-config-standard'),
+  trailingComma: 'es5',
+}
diff --git a/base/core/core.js b/base/core/core.js
@@ -126,7 +126,7 @@ var Page = (function PageClosure() {
       var promise = new Promise();
       if (!this.resources) //empty page
         promise.resolve();
-      else    
+      else
         this.resourcesPromise.then(function resourceSuccess() {
             var objectLoader = new ObjectLoader(this.resources.map,
                                                 keys,
@@ -331,8 +331,8 @@ var PDFDocument = (function PDFDocumentClosure() {
 
   PDFDocument.prototype = {
     parse: function PDFDocument_parse(recoveryMode) {
-      this.setup(recoveryMode);
       try {
+	    this.setup(recoveryMode); // setup may throw
         // checking if AcroForm is present
         this.acroForm = this.catalog.catDict.get('AcroForm');
         if (this.acroForm) {
@@ -345,7 +345,7 @@ var PDFDocument = (function PDFDocumentClosure() {
           }
         }
       } catch (ex) {
-        info('Something wrong with AcroForm entry');
+        error(ex);
         this.acroForm = null;
       }
     },

diff --git a/base/core/image.js b/base/core/image.js
@@ -485,8 +485,14 @@ var PDFImage = (function PDFImageClosure() {
         height: drawHeight,
         data: new Uint8Array(drawWidth * drawHeight * 4)
       };
-      var pixels = imgData.data;
-      this.fillRgbaBuffer(pixels, drawWidth, drawHeight);
+
+		var pixels = imgData.data;
+		try {
+			this.fillRgbaBuffer(pixels, drawWidth, drawHeight);
+		} catch (e) {
+			warn('Unable to decode image: ' + e); // keep moving on
+		}
+
       return imgData;
     },
     getImageBytes: function PDFImage_getImageBytes(length) {

diff --git a/base/core/stream.js b/base/core/stream.js
@@ -1648,7 +1648,9 @@ var CCITTFaxStream = (function CCITTFaxStreamClosure() {
     this.str = str;
     this.dict = str.dict;
 
-    params = params || new Dict();
+	if (!(params instanceof Dict)) {
+		params = new Dict();
+	}
 
     this.encoding = params.get('K') || 0;
     this.eoline = params.get('EndOfLine') || false;
@@ -1707,7 +1709,7 @@ var CCITTFaxStream = (function CCITTFaxStreamClosure() {
 
     if (a1 > codingLine[codingPos]) {
       if (a1 > this.columns) {
-        info('row is wrong length');
+        info('row is wrong length: ' + a1 + ' > ' + this.columns);
         this.err = true;
         a1 = this.columns;
       }
@@ -1727,7 +1729,7 @@ var CCITTFaxStream = (function CCITTFaxStreamClosure() {
 
     if (a1 > codingLine[codingPos]) {
       if (a1 > this.columns) {
-        info('row is wrong length');
+        info('row is wrong length: ' + a1 + ' > ' + this.columns);
         this.err = true;
         a1 = this.columns;
       }

diff --git a/base/display/api.js b/base/display/api.js
@@ -108,7 +108,7 @@ PDFJS.postMessageTransfers = PDFJS.postMessageTransfers === undefined ?
  *  - password - For decrypting password-protected PDFs.
  *  - initialData - A typed array with the first portion or all of the pdf data.
  *                  Used by the extension since some data is already loaded
- *                  before the switch to range requests. 
+ *                  before the switch to range requests.
  *
  * @param {object} pdfDataRangeTransport is optional. It is used if you want
  * to manually serve range requests for data in the PDF. See viewer.js for
@@ -417,7 +417,7 @@ var PDFPageProxy = (function PDFPageProxyClosure() {
             internalRenderTask.operatorListChanged();
           }
           catch(err) {
-            complete(err); 
+            complete(err);
           }
         },
         function pageDisplayReadPromiseError(reason) {
@@ -1005,9 +1005,11 @@ var PDFObjects = (function PDFObjectsClosure() {
 
       // If there isn't an object yet or the object isn't resolved, then the
       // data isn't ready yet!
-      if (!obj || !obj.resolved)
-        error('Requesting object that isn\'t resolved yet ' + objId);
-
+	  if (!obj || !obj.resolved) { // fix issue #255, keep moving on, test PDF is i255_emojis.pdf
+		// error('Requesting object that isn\'t resolved yet ' + objId)
+	    warn('Requesting object that isn\'t resolved yet ' + objId);
+		return null;
+	  }
       return obj.data;
     },
 

diff --git a/bin/pdf2json.js b/bin/pdf2json.js
@@ -1,4 +1,4 @@
 #!/usr/bin/env node
 
-import PDFCLI from "../lib/p2jcmd.js";
+import PDFCLI from "./cli/pdfparser_cli.js";
 new PDFCLI().start();
diff --git a/lib/parserstream.js b/lib/parserstream.js
@@ -1,9 +1,10 @@
 import { Transform, Readable } from "stream";
 import fs from "fs";
+import { Buffer } from "buffer";
 
 export class ParserStream extends Transform {
     static createContentStream(jsonObj) {
-		const rStream = new Readable({objectMode: true});
+		const rStream = new Readable({ objectMode: true });
 		rStream.push(jsonObj);
 		rStream.push(null);
 		return rStream;
@@ -18,8 +19,8 @@ export class ParserStream extends Transform {
 
     #pdfParser = null;
     #chunks = [];
-    #parsedData = {Pages:[]};
-    #_flush_callback = null; 
+    #parsedData = { Pages:[] };
+    #_flush_callback = null;
 
     constructor(pdfParser, options) {
         super(options);
@@ -32,14 +33,14 @@ export class ParserStream extends Transform {
         //     this.#_flush_callback();
         //     this.emit('end', null);
         // });
-        this.#pdfParser.on("readable", meta => this.#parsedData = {...meta, Pages:[]});
+        this.#pdfParser.on("readable", meta => this.#parsedData = { ...meta, Pages:[] });
         this.#pdfParser.on("data", page => {
-            if (!page) {
-                this.push(this.#parsedData);
-                this.#_flush_callback();
-            }
-            else 
-                this.#parsedData.Pages.push(page);
+			if (!page) {
+				this.push(this.#parsedData);
+				this.#_flush_callback();
+			} else {
+				this.#parsedData.Pages.push(page);
+			}
         });
     }
 
@@ -57,17 +58,16 @@ export class ParserStream extends Transform {
     _destroy() {
         super.removeAllListeners();
         this.#pdfParser = null;
-        this.#chunks = [];         
+        this.#chunks = [];
     }
-} 
-
+}
 
 export class StringifyStream extends Transform {
     constructor(options) {
         super(options);
 
         this._readableState.objectMode = false;
-        this._writableState.objectMode = true;    
+        this._writableState.objectMode = true;
     }
 
     _transform(obj, encoding, callback){