Trích Xuất Văn Bản Từ File PDF Bằng JavaScript & Node.js

Extracting structured text from PDFs is a core component of search indexing, invoice parsing, LLM retrieval-augmented generation (RAG), and data analysis. This guide shows how to extract raw text and layout data in JavaScript and Node.js.

Using pdf-parse in Node.js

pdf-parse is a lightweight Node.js library built on top of Mozilla's PDF.js:

Terminal

npm install pdf-parse

TypeScript

import fs from 'fs';
import pdf from 'pdf-parse';

export async function extractPdfText(filePath: string) {
  const dataBuffer = fs.readFileSync(filePath);

  const data = await pdf(dataBuffer);

  return {
    pageCount: data.numpages,
    text: data.text,
    info: data.info,
    metadata: data.metadata,
  };
}

Verified Test Asset.pdf

Download our structured 5MB PDF test file to test text extraction, page counts, and metadata inspection.

Tải File Mẫu 5MB PDF →

Extracting Text in the Browser with PDF.js

Mozilla's pdfjs-dist allows client-side parsing without server roundtrips:

TypeScript

import * as pdfjsLib from 'pdfjs-dist';

// Set worker source
pdfjsLib.GlobalWorkerOptions.workerSrc = '//cdnjs.cloudflare.com/ajax/libs/pdf.js/3.11.174/pdf.worker.min.js';

export async function extractTextInBrowser(file: File): Promise<string> {
  const arrayBuffer = await file.arrayBuffer();
  const pdf = await pdfjsLib.getDocument({ data: arrayBuffer }).promise;
  let fullText = '';

  for (let i = 1; i <= pdf.numPages; i++) {
    const page = await pdf.getPage(i);
    const content = await page.getTextContent();
    const pageText = content.items.map((item: any) => item.str).join(' ');
    fullText += `--- Page ${i} ---\n${pageText}\n\n`;
  }

  return fullText;
}

Handling Text Extraction Gotchas

Two-column layouts: Text extractors reading stream order may read across columns instead of down.
Font encoding: Non-standard embedded font encodings can output garbled Unicode characters without appropriate character map tables.

Câu Hỏi Thường Gặp (FAQ)

What is the difference between pdf-parse and pdfjs-dist?▾

pdf-parse is a lightweight wrapper returning all text as a single string. pdfjs-dist provides lower-level access to individual text items with their coordinates, enabling layout-aware extraction for columns and tables. Use pdf-parse for simple extraction, pdfjs-dist when you need positional data.

Why is extracted text garbled or in wrong order?▾

PDFs store text as glyphs at absolute positions with no inherent reading order. Correct extraction requires sorting by Y position (lines) then X position (left-to-right). For multi-column PDFs, cluster glyphs into columns first using pdfjs-dist's TextContent API.

How do I extract text from a scanned PDF?▾

Scanned PDFs contain images, not text, so OCR is required. Use Tesseract.js in Node.js: render each PDF page to canvas with pdfjs-dist, then pass the canvas image data to Tesseract.

Trích Xuất Văn Bản Từ File PDF Bằng JavaScript & Node.js

Using pdf-parse in Node.js

Extracting Text in the Browser with PDF.js

Handling Text Extraction Gotchas

Nguyễn Đại Long

Bài viết này có hữu ích không?

Thảo Luận Kỹ Thuật & Đóng Góp Ý Kiến (0)

Câu Hỏi Thường Gặp (FAQ)

Bài Viết Liên Quan