feat: reading multiple pdf files with a single PDFParser object (#371)

* feat: reading multiple pdf files with a single PDFParser object

* Removed useless if statement

* Disposed the object by calling destroy method

* Added a new method for resetting the PDFJS object

* Removed removal of listeners when reading multiple pdf files with a single object

* Restoring minimum NodeJS version to 18.12.1 for semver compliance #381

* [EVO] Added parameter for using PDFParser as singleton in CLI #381

* [ROLL] Restored node and npm version changes

* [FIX] pdfParser always a singleton, even if not provided by parameter

* [FIX] Possible race condition

---------

Co-authored-by: Modesty Zhang <modestyz@hotmail.com>

Author: nicolabaesso

jest.config.json

@@ -1,5 +1,5 @@
 {
-	"testMatch": ["**/test/_test_.cjs"],
+	"testMatch": ["**/test/_test_*"],
 	"testEnvironment": "node",
 	"bail": false,
 	"testFailureExitCode": 1,

lib/pdf.js

@@ -189,8 +189,7 @@ export default class PDFJSClass extends EventEmitter {
 	}
 
 	parsePDFData(arrayBuffer, password) {
-		this.pdfDocument = null;
-
+		this.resetCurrentObject();
 		const parameters = { password, data: arrayBuffer };
 		PDFJS.getDocument(parameters).then(
 			(pdfDocument) => this.load(pdfDocument, 1),
@@ -299,6 +298,7 @@ export default class PDFJSClass extends EventEmitter {
 				};
 
 				this.pages.push(page);
+
 				this.emit("data", page);
 
 				if (this.needRawText) {
@@ -405,6 +405,14 @@ export default class PDFJSClass extends EventEmitter {
 		return { Pages: this.pages };
 	}
 
+	resetCurrentObject(){
+		if (this.pdfDocument) this.pdfDocument.destroy();
+		this.pdfDocument = null;
+
+		this.pages = [];
+		this.rawTextContents = [];
+	}
+
 	destroy() {
 		this.removeAllListeners();
 

pdfparser.js

@@ -304,6 +304,15 @@ export default class PDFParser extends EventEmitter {
 		return ParserStream.createContentStream(this.getMergedTextBlocksIfNeeded());
 	}
 
+	/**
+	 * Destroys the current instance of PDFJS and sets a new one
+	 * @param {boolean} needRawText - Whether raw text is needed or not
+	 */
+	resetPDFJS(needRawText){
+		this.#PDFJS.destroy();
+		this.#PDFJS=new PDFJS(needRawText);
+	}
+
 	/**
 	 * Destroy the PDFParser instance.
 	 */

src/cli/p2jcli.ts

@@ -18,6 +18,7 @@ const PROCESS_RAW_TEXT_CONTENT = "c" in argv;
 const PROCESS_FIELDS_CONTENT = "t" in argv;
 const PROCESS_MERGE_BROKEN_TEXT_BLOCKS = "m" in argv;
 const PROCESS_WITH_STREAM = "r" in argv;
+const SINGLETON_PDF_PARSER= "si" in argv;
 
 const INPUT_DIR_OR_FILE = argv.f;
 
@@ -123,10 +124,13 @@ class PDFProcessor {
 
 	private parseOnePDFStream() {
 		return new Promise((resolve, reject) => {
-			this.pdfParser = new PDFParser(null, PROCESS_RAW_TEXT_CONTENT);
-			this.pdfParser.on("pdfParser_dataError", (evtData: any) =>
-				this.onPrimaryError(evtData.parserError, reject)
-			);
+			if((SINGLETON_PDF_PARSER && !this.pdfParser) || !SINGLETON_PDF_PARSER){
+				//initialize the parser if the singleton parameter was not provided, or if the singleton parameter was provided and the parser is not initialized
+				this.pdfParser = new PDFParser(null, PROCESS_RAW_TEXT_CONTENT);
+				this.pdfParser.on("pdfParser_dataError", (evtData: any) =>
+					this.onPrimaryError(evtData.parserError, reject)
+				);
+			}
 
 			const outputStream = fs.createWriteStream(this.outputPath);
 			outputStream.on("finish", () => this.onPrimarySuccess(resolve, reject));
@@ -145,10 +149,13 @@ class PDFProcessor {
 
 	private parseOnePDF() {
 		return new Promise((resolve, reject) => {
-			this.pdfParser = new PDFParser(null, PROCESS_RAW_TEXT_CONTENT);
-			this.pdfParser.on("pdfParser_dataError", (evtData: any) => {
-				this.onPrimaryError(evtData.parserError, reject);
-			});
+			if((SINGLETON_PDF_PARSER && !this.pdfParser) || !SINGLETON_PDF_PARSER){
+				//initialize the parser if the singleton parameter was not provided, or if the singleton parameter was provided and the parser is not initialized
+				this.pdfParser = new PDFParser(null, PROCESS_RAW_TEXT_CONTENT);
+				this.pdfParser.on("pdfParser_dataError", (evtData: any) =>
+					this.onPrimaryError(evtData.parserError, reject)
+				);
+			}
 
 			this.pdfParser.on("pdfParser_dataReady", (evtData: any) => {
 				fs.writeFile(this.outputPath, JSON.stringify(evtData), (err) => {

src/cli/p2jcliarg.ts

@@ -21,6 +21,7 @@ export type Argv = {
 	c?: string,
 	m?: string,
 	r?: string,
+	si?: string,
 };
 
 export class CLIArgParser {
@@ -199,4 +200,8 @@ export const yargs = new CLIArgParser(process.argv.slice(2))
 		"r",
 		"stream",
 		"(optional) when specified, will process and parse with buffer/object transform stream rather than file system."
+	).alias(
+		"si",
+		"singleton",
+		"(optional) when specified, only an instance of PDFParser will be initialized."
 	);

test/_test_testMultipleDataPDF.cjs

@@ -0,0 +1,52 @@
+const assert = require("assert");
+const fs = require("fs");
+
+const PDFParser = require("../dist/pdfparser.cjs");
+// we want to read two (or more) PDF files without recreating a reference to PDFParser
+describe("Multiple PDFs with same structure",()=>{
+    test("Read different values",async ()=>{
+        // the target PDFs for this test have only 3 values: Name, Surname and BirthDate.
+        // you can find the PDFs in test/pdf/mpf
+        let parser=new PDFParser();
+        const firstPDFLocation=__dirname+"/pdf/mpf/testPDF.pdf";
+        const secondPDFLocation=__dirname+"/pdf/mpf/testPDF2.pdf";
+        const firstPDFBuffer=fs.readFileSync(firstPDFLocation);
+        const secondPDFBuffer=fs.readFileSync(secondPDFLocation);
+        //we need to check if buffers are indeed different, otherwise it's useless!
+        expect(firstPDFBuffer).not.toBe(secondPDFBuffer);
+        const firstData=await new Promise((resolve,reject)=>{
+		    parser.parseBuffer(firstPDFBuffer,5);
+            parser.on("pdfParser_dataReady", (evtData) => {
+                resolve(evtData);
+            });
+    
+            parser.on("pdfParser_dataError", (evtData) => {
+                reject(evtData);
+            });
+        });
+        const secondData=await new Promise((resolve,reject)=>{
+		    parser.parseBuffer(secondPDFBuffer,5);
+            parser.on("pdfParser_dataReady", (evtData) => {
+                resolve(evtData);
+            });
+    
+            parser.on("pdfParser_dataError", (evtData) => {
+                reject(evtData);
+            });
+        });
+        //first, make sure the files are read
+        expect(firstData).toBeDefined();
+        expect(firstData.Pages[0]).toBeDefined();
+        expect(firstData.Pages[0].Fields).toBeDefined();
+        expect(secondData).toBeDefined();
+        expect(secondData.Pages[0]).toBeDefined();
+        expect(secondData.Pages[0].Fields).toBeDefined();
+        //then, we check if the files have the correct values
+        expect(firstData.Pages[0].Fields[0].V).toBe("Mario");
+        expect(firstData.Pages[0].Fields[1].V).toBe("Rossi");
+        expect(firstData.Pages[0].Fields[2].V).toBe("01/01/1990");
+        expect(secondData.Pages[0].Fields[0].V).toBe("Luigi");
+        expect(secondData.Pages[0].Fields[1].V).toBe("Verdi");
+        expect(secondData.Pages[0].Fields[2].V).toBe("01/01/1991");
+    });
+});