Install

npm install @siliph/pdf-engine-core

Quick Start

import { mergePdfs, splitPdf, rotate, fillForm, search } from '@siliph/pdf-engine-core';

// Merge PDFs
const merged = await mergePdfs([doc1, doc2]);

// Split by page ranges
const parts = await splitPdf(doc, [{ start: 0, end: 4 }, { start: 5, end: 9 }]);

// Rotate page 90°
const cmd = await rotate(doc, 0, 90);
if (cmd.ok) doc.applyCommand(cmd.value);

// Fill form
const formCmd = await fillForm(doc, { name: 'John', agree: true });

// Full-text search
for await (const hit of search(doc, 'invoice')) {
  console.log(`Page ${hit.pageIndex}, offset ${hit.offset}`);
}

Tree Shaking

Each task is independently importable — no bundle bloat:

import { mergePdfs } from '@siliph/pdf-engine-core/tasks/mergePdfs'; // ~15KB gzipped

20 Task Functions

| Function | Description | |----------|-------------| | mergePdfs | Merge multiple documents into one | | splitPdf | Split by page ranges | | splitByBookmarks | Split at bookmark boundaries | | splitBySize | Split into max-size chunks | | extractPages | Extract specific pages | | rotate | Rotate pages (90°/180°/270°) | | crop | Crop (metadata or content clip) | | compress | Remove unused resources (GC) | | addWatermark | Add text watermark to all pages | | setHeaderFooter | Add headers/footers with page numbers | | setMetadata | Set title, author, keywords | | redact | Visual or permanent redaction | | fillForm | Fill AcroForm fields | | search | Full-text search (Unicode NFC, async iterator) | | mixPages | Alternate pages from multiple docs | | addBackPages | Insert back pages for duplex printing | | setBookmarks | Set document outline | | getBookmarks | Read current bookmarks | | pdfToImages | Convert pages to PNG/JPEG |

14 Engines

| Engine | Capability | |--------|-----------| | Parser | XRef streams, object streams, lazy loading, brute-force recovery | | Rendering | Canvas2D, OffscreenCanvas Workers, blend modes, transparency groups | | Font | TrueType, Type1, CID, CMap, ToUnicode, glyph LRU cache, Standard 14 | | Text | Run reconstruction, in-place editing, kerning, RTL/BiDi, selection | | Image | JPEG, JPEG2000, JBIG2, CCITTFax, Flate, SMask compositing | | Annotation | 9 types, appearance streams, visual + permanent redaction | | Forms | AcroForms, all field types, validation, radio groups | | Security | RC4, AES-128, AES-256, digital signatures (PKCS#7) | | Page | Merge, split, extract, rotate, crop, reorder, resource dedup | | Export | Incremental save, full save, streaming, Worker offload | | Search | Unicode NFC, ligature decomposition, hyphenation, async iterator | | Selection | Hit-testing (4px tolerance), multi-select, marquee | | Graphics | Shape primitives, ink paths | | Validation | PDF/A-1b conformance checking |

Benchmarks

| Operation | Time | Memory | |-----------|------|--------| | Parse XRef (500pg, 12MB) | ~85ms | 4MB | | First page render | ~180ms | 28MB | | Merge 100pg + 50pg | ~45ms | 18MB | | Full-text search (200pg) | ~120ms | 6MB | | AES-256 decrypt (5MB) | ~95ms | 8MB | | Incremental save | ~35ms | 12MB |

Error Handling

All functions return Result<T, TaskError> — never throws:

const result = await mergePdfs([doc1, doc2]);
if (result.ok) {
  console.log(result.value.getPageCount());
} else {
  console.error(result.error.code);    // 'INVALID_INPUT' | 'ENGINE_FAILURE'
  console.error(result.error.message);
}

Part of @siliph/pdf-engine

| Package | Description | |---------|-------------| | @siliph/pdf-engine | Umbrella (re-exports all) | | @siliph/pdf-engine-core | Parser, engines, task API (this package) | | @siliph/pdf-engine-ui | Viewport, toolbar, accessibility | | @siliph/pdf-engine-plugin-sdk | Plugin extension system |

Links

• 🌐 siliph.com — Documentation
• 📦 GitHub — Source code
• 🐛 Issues — Bug reports

License

MIT