test-pdf-parsing.jsโข6.67 kB
#!/usr/bin/env node
/**
* Test script for PDF parsing functionality using @opendocsg/pdf2md (v3)
* Verifies parsing of sample PDFs and URL
*/
import { parsePdfToMarkdown } from '../dist/tools/pdf/index.js';
import path from 'path';
import { fileURLToPath } from 'url';
import fs from 'fs/promises';
const __filename = fileURLToPath(import.meta.url);
const __dirname = path.dirname(__filename);
const SAMPLES_DIR = path.join(__dirname, 'samples');
const SAMPLES = [
'01_sample_simple.pdf',
'02_sample_invoice.pdf',
'03_sample_compex.pdf',
// 'gpc-genai-ocsummaryv2-content.pdf',
// '2025-Wharton-GBK-AI-Adoption-Report_Full-Report.pdf'
// 'statement.pdf'
];
const URL_SAMPLE = 'https://pdfobject.com/pdf/sample.pdf';
async function testSample(name, source) {
console.log(`\n================================================================================`);
console.log(`Processing: ${name}`);
console.log(`Source: ${source}`);
console.log(`--------------------------------------------------------------------------------`);
try {
const startTime = Date.now();
// Content (Markdown)
console.log('\n๐ CONTENT PREVIEW (Markdown):');
const result = await parsePdfToMarkdown(source);
// Verify new structure
if (result.pages) {
console.log(`\n๐ Pages Found: ${result.pages.length}`);
result.pages.forEach((p, i) => {
console.log(` Page ${p.pageNumber}: ${p.text.length} chars, ${p.images.length} images`);
});
}
const markdown = result.pages.map(p => p.text).join('');
const images = result.pages.flatMap(p => p.images);
const processingTime = Date.now() - startTime;
console.log('\n--- Full Text Preview ---');
console.log(markdown.substring(0, 200) + '...');
// Save extracted images to disk
if (images && images.length > 0) {
console.log(`\n๐ผ๏ธ EXTRACTED IMAGES (${images.length}):`);
// Create images directory for this PDF
const imagesDir = path.join(SAMPLES_DIR, `${name}_images`);
await fs.mkdir(imagesDir, { recursive: true });
for (let i = 0; i < images.length; i++) {
const img = images[i];
// Determine file extension from MIME type
const ext = img.mimeType.split('/')[1] || 'png';
const filename = `page_${img.page}_img_${i + 1}.${ext}`;
const filepath = path.join(imagesDir, filename);
// Decode base64 and save to file
const buffer = Buffer.from(img.data, 'base64');
await fs.writeFile(filepath, buffer);
console.log(` - Saved: ${filename} (${img.width}x${img.height}, ${img.mimeType})`);
}
console.log(`\n Images saved to: ${imagesDir}`);
}
// save to markdown file
const markdownPath = path.join(SAMPLES_DIR, `${name}.md`);
await fs.writeFile(markdownPath, markdown);
const preview = markdown.substring(0, 500).replace(/\n/g, '\n ');
console.log(` ${preview}...`);
console.log(`\n [Total Length: ${markdown.length} chars]`);
console.log(` Processing Time: ${processingTime}ms`);
} catch (error) {
console.error(`โ Error processing ${name}:`, error);
}
}
async function testPageFiltering() {
console.log(`\n================================================================================`);
console.log(`๐งช Testing Page Filtering`);
console.log(`--------------------------------------------------------------------------------`);
const sampleName = '03_sample_compex.pdf';
const samplePath = path.join(SAMPLES_DIR, sampleName);
try {
// Case 1: Specific Pages (Array)
console.log('\nCase 1: Specific Pages [1]');
const result1 = await parsePdfToMarkdown(samplePath, [1]);
console.log(`Pages returned: ${result1.pages.length}`);
console.log(`Page numbers: ${result1.pages.map(p => p.pageNumber).join(', ')}`);
// Case 2: Page Range (Positive Offset) - First Page
console.log('\nCase 2: Page Range { offset: 0, length: 1 } (First Page)');
const result2 = await parsePdfToMarkdown(samplePath, { offset: 0, length: 1 });
console.log(`Pages returned: ${result2.pages.length}`);
console.log(`Page numbers: ${result2.pages.map(p => p.pageNumber).join(', ')}`);
// Case 3: Page Range (Negative Offset) - Last Page
console.log('\nCase 3: Page Range { offset: -2, length: 2 } (Last Page)');
const result3 = await parsePdfToMarkdown(samplePath, { offset: -2, length: 2 });
console.log(`Pages returned: ${result3.pages.length}`);
console.log(`Page numbers: ${result3.pages.map(p => p.pageNumber).join(', ')}`);
// Case 4: Page Range (Large Length) - All Pages from start
console.log('\nCase 4: Page Range { offset: 0, length: 100 } (All Pages)');
const result4 = await parsePdfToMarkdown(samplePath, { offset: 0, length: 100 });
console.log(`Pages returned: ${result4.pages.length}`);
console.log(`Page numbers: ${result4.pages.map(p => p.pageNumber).join(', ')}`);
// Case 5: Page Range (Large Negative Offset) - All Pages from end
console.log('\nCase 5: Page Range { offset: -100, length: 100 } (All Pages from end)');
const result5 = await parsePdfToMarkdown(samplePath, { offset: -100, length: 100 });
console.log(`Pages returned: ${result5.pages.length}`);
console.log(`Page numbers: ${result5.pages.map(p => p.pageNumber).join(', ')}`);
// Case 6: Page Range (Large Length) - All Pages
console.log('\nCase 6: Page Range [1,5,14] (All Pages)');
const result6 = await parsePdfToMarkdown(samplePath, [1, 5, 14]);
console.log(`Pages returned: ${result6.pages.length}`);
console.log(`Page numbers: ${result6.pages.map(p => p.pageNumber).join(', ')}`);
} catch (error) {
console.error('โ Error in page filtering test:', error);
}
}
async function main() {
console.log('๐งช PDF v3 Sample Test Suite (@opendocsg/pdf2md)');
// Test Page Filtering
await testPageFiltering();
// Test Local Samples
for (const sample of SAMPLES) {
const samplePath = path.join(SAMPLES_DIR, sample);
await testSample(sample, samplePath);
}
// Test URL
await testSample('URL Sample', URL_SAMPLE);
}
if (import.meta.url === `file://${process.argv[1]}`) {
main().catch(console.error);
}